Chiebot-Mirror
/
grpc
mirror of https://github.com/grpc/grpc.git
8
0
Fork 0
Code Issues Projects Releases Wiki Activity
The C based gRPC (C++, Python, Ruby, Objective-C, PHP, C#) https://grpc.io/
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
395 Commits
276 Branches
377 Tags
1.2 GiB
Tag: Branch: Tree: 86af8cffc5
Branches Tags
${ item.name }
Create tag ${ searchTerm }
Create branch ${ searchTerm }
from '86af8cffc5'
${ noResults }
grpc/Makefile

9413 lines
476 KiB
Raw Normal View History Unescape

Initial import.
10 years ago
# GRPC global makefile
# This currently builds C and C++ code.
Add msan, tsan, asan targets for Makefile. Expand test runner to be compatible. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83378004
10 years ago
# Configurations
VALID_CONFIG_opt = 1
CC_opt = gcc
CXX_opt = g++
LD_opt = gcc
LDXX_opt = g++
CPPFLAGS_opt = -O2
LDFLAGS_opt =
DEFINES_opt = NDEBUG
VALID_CONFIG_dbg = 1
CC_dbg = gcc
CXX_dbg = g++
LD_dbg = gcc
LDXX_dbg = g++
CPPFLAGS_dbg = -O0
LDFLAGS_dbg =
DEFINES_dbg = _DEBUG DEBUG
Compile OpenSSL differently for each config. MSAN, TSAN, ASAN need to see all the code. Valgrind, MSAN want -DPURIFY.
10 years ago
VALID_CONFIG_valgrind = 1
Force using custom libraries if we are using ?SAN. These libraries need to see all of the source code, so using system installed libraries is out of the question.
10 years ago
REQUIRE_CUSTOM_LIBRARIES_valgrind = 1
Compile OpenSSL differently for each config. MSAN, TSAN, ASAN need to see all the code. Valgrind, MSAN want -DPURIFY.
10 years ago
CC_valgrind = gcc
CXX_valgrind = g++
LD_valgrind = gcc
LDXX_valgrind = g++
CPPFLAGS_valgrind = -O0
OPENSSL_CFLAGS_valgrind = -DPURIFY
LDFLAGS_valgrind =
DEFINES_valgrind = _DEBUG DEBUG
Add msan, tsan, asan targets for Makefile. Expand test runner to be compatible. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83378004
10 years ago
VALID_CONFIG_tsan = 1
Force using custom libraries if we are using ?SAN. These libraries need to see all of the source code, so using system installed libraries is out of the question.
10 years ago
REQUIRE_CUSTOM_LIBRARIES_tsan = 1
Add msan, tsan, asan targets for Makefile. Expand test runner to be compatible. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83378004
10 years ago
CC_tsan = clang
CXX_tsan = clang++
LD_tsan = clang
LDXX_tsan = clang++
CPPFLAGS_tsan = -O1 -fsanitize=thread -fno-omit-frame-pointer
Compile OpenSSL differently for each config. MSAN, TSAN, ASAN need to see all the code. Valgrind, MSAN want -DPURIFY.
10 years ago
OPENSSL_CONFIG_tsan = no-asm
Add msan, tsan, asan targets for Makefile. Expand test runner to be compatible. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83378004
10 years ago
LDFLAGS_tsan = -fsanitize=thread
DEFINES_tsan = NDEBUG
VALID_CONFIG_asan = 1
Force using custom libraries if we are using ?SAN. These libraries need to see all of the source code, so using system installed libraries is out of the question.
10 years ago
REQUIRE_CUSTOM_LIBRARIES_asan = 1
Add msan, tsan, asan targets for Makefile. Expand test runner to be compatible. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83378004
10 years ago
CC_asan = clang
CXX_asan = clang++
LD_asan = clang
LDXX_asan = clang++
CPPFLAGS_asan = -O1 -fsanitize=address -fno-omit-frame-pointer
Compile OpenSSL differently for each config. MSAN, TSAN, ASAN need to see all the code. Valgrind, MSAN want -DPURIFY.
10 years ago
OPENSSL_CONFIG_asan = no-asm
Add msan, tsan, asan targets for Makefile. Expand test runner to be compatible. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83378004
10 years ago
LDFLAGS_asan = -fsanitize=address
DEFINES_asan = NDEBUG
VALID_CONFIG_msan = 1
Force using custom libraries if we are using ?SAN. These libraries need to see all of the source code, so using system installed libraries is out of the question.
10 years ago
REQUIRE_CUSTOM_LIBRARIES_msan = 1
Add msan, tsan, asan targets for Makefile. Expand test runner to be compatible. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83378004
10 years ago
CC_msan = clang
CXX_msan = clang++
LD_msan = clang
LDXX_msan = clang++
CPPFLAGS_msan = -O1 -fsanitize=memory -fno-omit-frame-pointer
Compile OpenSSL differently for each config. MSAN, TSAN, ASAN need to see all the code. Valgrind, MSAN want -DPURIFY.
10 years ago
OPENSSL_CFLAGS_msan = -DPURIFY
OPENSSL_CONFIG_msan = no-asm
Add msan, tsan, asan targets for Makefile. Expand test runner to be compatible. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83378004
10 years ago
LDFLAGS_msan = -fsanitize=memory
DEFINES_msan = NDEBUG
Add tools/run_tests/run_lcov.sh This is a tool for collecting coverage reports from our tests and combining them with the open-source tool 'lcov'.
10 years ago
VALID_CONFIG_gcov = 1
CC_gcov = gcc
CXX_gcov = g++
LD_gcov = gcc
LDXX_gcov = g++
CPPFLAGS_gcov = -O0 -fprofile-arcs -ftest-coverage
LDFLAGS_gcov = -fprofile-arcs -ftest-coverage
DEFINES_gcov = NDEBUG
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
Initial import.
10 years ago
# General settings.
# You may want to change these depending on your system.
prefix ?= /usr/local
PROTOC = protoc
Use CONFIG after defining it. Define CONFIG before using it. Change on 2015/01/06 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83383325
10 years ago
CONFIG ?= opt
Add msan, tsan, asan targets for Makefile. Expand test runner to be compatible. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83378004
10 years ago
CC = $(CC_$(CONFIG))
CXX = $(CXX_$(CONFIG))
LD = $(LD_$(CONFIG))
LDXX = $(LDXX_$(CONFIG))
Initial import.
10 years ago
AR = ar
STRIP = strip --strip-unneeded
INSTALL = install -D
RM = rm -f
Use CONFIG after defining it. Define CONFIG before using it. Change on 2015/01/06 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83383325
10 years ago
ifndef VALID_CONFIG_$(CONFIG)
$(error Invalid CONFIG value '$(CONFIG)')
endif
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# The HOST compiler settings are used to compile the protoc plugins.
# In most cases, you won't have to change anything, but if you are
# cross-compiling, you can override these variables from GNU make's
# command line: make CC=cross-gcc HOST_CC=gcc
Getting the generated Makefile to support .proto files. -) Compiling the proto plugin using the HOSTCC compiler. -) Set up proper rules to invoke the plugin from protoc. -) Few various renaming hacks to fully get out of []. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82000361
10 years ago
HOST_CC = $(CC)
HOST_CXX = $(CXX)
HOST_LD = $(LD)
HOST_LDXX = $(LDXX)
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CPPFLAGS += $(CPPFLAGS_$(CONFIG))
DEFINES += $(DEFINES_$(CONFIG))
Add msan, tsan, asan targets for Makefile. Expand test runner to be compatible. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83378004
10 years ago
LDFLAGS += $(LDFLAGS_$(CONFIG))
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
Initial import.
10 years ago
CFLAGS += -std=c89 -pedantic
CXXFLAGS += -std=c++11
CPPFLAGS += -g -fPIC -Wall -Werror -Wno-long-long
LDFLAGS += -g -pthread -fPIC
INCLUDES = . include gens
Fix opensource build. Change on 2015/01/07 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83469328
10 years ago
LIBS = rt m z pthread
Initial import.
10 years ago
LIBSXX = protobuf
Fixing C++ and ruby code generators for the opensource build. Change on 2014/12/11 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81921764
10 years ago
LIBS_PROTOC = protoc protobuf
Initial import.
10 years ago
ifneq ($(wildcard /usr/src/gtest/src/gtest-all.cc),)
GTEST_LIB = /usr/src/gtest/src/gtest-all.cc -I/usr/src/gtest
else
GTEST_LIB = -lgtest
endif
Build CPP interop client in Git-on-[] as well as in []. To build client in Git-on-[]: # regenerate Makefile net/grpc/tools/buildgen/generate_projects.sh # generate .pb.h and .pb.cc protoc --cpp_out=. --grpc_out=. --plugin=protoc-gen-grpc=[]-bin/net/grpc/compiler/cpp_plugin net/grpc/cpp/test/interop/test.proto net/grpc/cpp/test/interop/empty.proto net/grpc/cpp/test/interop/messages.proto # Complile and link net/grpc/tools/build_grpc_dev.sh bins/interop_client To test against GFE/ESF: # bring up server [] build net/grpc/testing/interop:server_components_env []-bin/net/grpc/testing/interop/server_components_env --manual --rpc_port=25000 # start client /tmp/grpc-codebase/bins/interop_client --enable_ssl=true --server_port="server ssl port listening port" To test [] build against GFE/ESF: [] run net/grpc/cpp:interop_client -- --enable_ssl=true --server_port="server ssl port listening port" Change on 2014/12/10 by chenw <chenw@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81821921
10 years ago
GTEST_LIB += -lgflags
Initial import.
10 years ago
ifeq ($(V),1)
E = @:
Q =
else
E = @echo
Q = @
endif
VERSION = 0.8.0.0
CPPFLAGS_NO_ARCH += $(addprefix -I, $(INCLUDES)) $(addprefix -D, $(DEFINES))
CPPFLAGS += $(CPPFLAGS_NO_ARCH) $(ARCH_FLAGS)
LDFLAGS += $(ARCH_FLAGS)
LDLIBS += $(addprefix -l, $(LIBS))
LDLIBSXX += $(addprefix -l, $(LIBSXX))
Getting the generated Makefile to support .proto files. -) Compiling the proto plugin using the HOSTCC compiler. -) Set up proper rules to invoke the plugin from protoc. -) Few various renaming hacks to fully get out of []. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82000361
10 years ago
HOST_LDLIBS_PROTOC += $(addprefix -l, $(LIBS_PROTOC))
HOST_CPPFLAGS = $(CPPFLAGS)
HOST_CFLAGS = $(CFLAGS)
HOST_CXXFLAGS = $(CXXFLAGS)
HOST_LDFLAGS = $(LDFLAGS)
HOST_LDLIBS = $(LDLIBS)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
# These are automatically computed variables.
# There shouldn't be any need to change anything from now on.
HOST_SYSTEM = $(shell uname | cut -f 1 -d_)
ifeq ($(SYSTEM),)
SYSTEM = $(HOST_SYSTEM)
endif
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
ifeq ($(SYSTEM),MINGW32)
SHARED_EXT = dll
endif
ifeq ($(SYSTEM),Darwin)
SHARED_EXT = dylib
endif
ifeq ($(SHARED_EXT),)
SHARED_EXT = so.$(VERSION)
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(wildcard .git),)
IS_GIT_FOLDER = false
else
IS_GIT_FOLDER = true
endif
Solving chicken-and-egg issue while detecting OpenSSL capabilities. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82690228
10 years ago
OPENSSL_ALPN_CHECK_CMD = $(CC) $(CFLAGS) $(CPPFLAGS) -o /dev/null test/build/openssl-alpn.c -lssl -lcrypto -ldl $(LDFLAGS)
ZLIB_CHECK_CMD = $(CC) $(CFLAGS) $(CPPFLAGS) -o /dev/null test/build/zlib.c -lz $(LDFLAGS)
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
Force using custom libraries if we are using ?SAN. These libraries need to see all of the source code, so using system installed libraries is out of the question.
10 years ago
ifndef REQUIRE_CUSTOM_LIBRARIES_$(CONFIG)
Avoiding bashisms that'll prevent other shells to run our Makefile. Change on 2014/12/15 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82177761
10 years ago
HAS_SYSTEM_OPENSSL_ALPN = $(shell $(OPENSSL_ALPN_CHECK_CMD) 2> /dev/null && echo true || echo false)
HAS_SYSTEM_ZLIB = $(shell $(ZLIB_CHECK_CMD) 2> /dev/null && echo true || echo false)
Force using custom libraries if we are using ?SAN. These libraries need to see all of the source code, so using system installed libraries is out of the question.
10 years ago
else
# override system libraries if the config requires a custom compiled library
HAS_SYSTEM_OPENSSL_ALPN = false
HAS_SYSTEM_ZLIB = false
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(wildcard third_party/openssl/ssl/ssl.h),)
HAS_EMBEDDED_OPENSSL_ALPN = false
else
HAS_EMBEDDED_OPENSSL_ALPN = true
endif
ifeq ($(wildcard third_party/zlib/zlib.h),)
HAS_EMBEDDED_ZLIB = false
else
HAS_EMBEDDED_ZLIB = true
endif
ifeq ($(HAS_SYSTEM_ZLIB),false)
ifeq ($(HAS_EMBEDDED_ZLIB),true)
Support compiling zlib with ?SAN
10 years ago
ZLIB_DEP = libs/$(CONFIG)/zlib/libz.a
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
CPPFLAGS += -Ithird_party/zlib
LDFLAGS += -Lthird_party/zlib
else
DEP_MISSING += zlib
endif
endif
ifeq ($(HAS_SYSTEM_OPENSSL_ALPN),false)
ifeq ($(HAS_EMBEDDED_OPENSSL_ALPN),true)
Compile OpenSSL differently for each config. MSAN, TSAN, ASAN need to see all the code. Valgrind, MSAN want -DPURIFY.
10 years ago
OPENSSL_DEP = libs/$(CONFIG)/openssl/libssl.a
OPENSSL_MERGE_LIBS += libs/$(CONFIG)/openssl/libssl.a libs/$(CONFIG)/openssl/libcrypto.a
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
CPPFLAGS += -Ithird_party/openssl/include
Compile OpenSSL differently for each config. MSAN, TSAN, ASAN need to see all the code. Valgrind, MSAN want -DPURIFY.
10 years ago
LDFLAGS += -Llibs/$(CONFIG)/openssl
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
LIBS_SECURE = dl
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
NO_SECURE = true
endif
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
else
LIBS_SECURE = ssl crypto dl
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
LDLIBS_SECURE += $(addprefix -l, $(LIBS_SECURE))
Generate dependencies when compiling the .o If we don't have a .o, we don't need dependencies, so skip the setup step of building them. Instead, when building the .o, we generate the .dep and output it (to objs/ to make the rules a little easier). The next run of make will include the dep file and any dependencies will be picked up. This change also disables the disabling of dependency checking if we have zlib or openssl being compiled from third_party. Additionally it inverts the logic for including dependencies from ONLY if we are doing a clean to ONLY if we are NOT doing a clean.
10 years ago
ifeq ($(MAKECMDGOALS),clean)
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
NO_DEPS = true
endif
Initial import.
10 years ago
.SECONDARY = %.pb.h %.pb.cc
Remove protoc_plugins, and use a variable. Files that have dependencies on phony targets are always remade, leading to much unnecessary compiling of the C++ tests.
10 years ago
PROTOC_PLUGINS= bins/$(CONFIG)/cpp_plugin bins/$(CONFIG)/ruby_plugin
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(DEP_MISSING),)
Initial import.
10 years ago
all: static shared
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
dep_error:
@echo "You shouldn't see this message - all of your dependencies are correct."
else
all: dep_error git_update stop
dep_error:
@echo
@echo "DEPENDENCY ERROR"
@echo
@echo "You are missing system dependencies that are essential to build grpc,"
@echo "and the third_party directory doesn't have them:"
@echo
@echo " $(DEP_MISSING)"
@echo
@echo "Installing the development packages for your system will solve"
@echo "this issue. Please consult INSTALL to get more information."
@echo
@echo "If you need information about why these tests failed, run:"
@echo
@echo " make run_dep_checks"
@echo
endif
git_update:
ifeq ($(IS_GIT_FOLDER),true)
@echo "Additionally, since you are in a git clone, you can download the"
@echo "missing dependencies in third_party by running the following command:"
@echo
Fix Makefile help text. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82155747
10 years ago
@echo " git submodule update --init"
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
@echo
endif
openssl_dep_error: openssl_dep_message git_update stop
openssl_dep_message:
@echo
@echo "DEPENDENCY ERROR"
@echo
@echo "The target you are trying to run requires OpenSSL with ALPN support."
@echo "Your system doesn't have it, and neither does the third_party directory."
@echo
@echo "Please consult INSTALL to get more information."
@echo
@echo "If you need information about why these tests failed, run:"
@echo
@echo " make run_dep_checks"
@echo
stop:
@false
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
gen_hpack_tables: bins/$(CONFIG)/gen_hpack_tables
cpp_plugin: bins/$(CONFIG)/cpp_plugin
ruby_plugin: bins/$(CONFIG)/ruby_plugin
grpc_byte_buffer_reader_test: bins/$(CONFIG)/grpc_byte_buffer_reader_test
gpr_cancellable_test: bins/$(CONFIG)/gpr_cancellable_test
gpr_log_test: bins/$(CONFIG)/gpr_log_test
gpr_useful_test: bins/$(CONFIG)/gpr_useful_test
gpr_cmdline_test: bins/$(CONFIG)/gpr_cmdline_test
gpr_histogram_test: bins/$(CONFIG)/gpr_histogram_test
gpr_host_port_test: bins/$(CONFIG)/gpr_host_port_test
gpr_slice_buffer_test: bins/$(CONFIG)/gpr_slice_buffer_test
gpr_slice_test: bins/$(CONFIG)/gpr_slice_test
gpr_string_test: bins/$(CONFIG)/gpr_string_test
gpr_sync_test: bins/$(CONFIG)/gpr_sync_test
gpr_thd_test: bins/$(CONFIG)/gpr_thd_test
gpr_time_test: bins/$(CONFIG)/gpr_time_test
murmur_hash_test: bins/$(CONFIG)/murmur_hash_test
grpc_stream_op_test: bins/$(CONFIG)/grpc_stream_op_test
alpn_test: bins/$(CONFIG)/alpn_test
time_averaged_stats_test: bins/$(CONFIG)/time_averaged_stats_test
chttp2_stream_encoder_test: bins/$(CONFIG)/chttp2_stream_encoder_test
hpack_table_test: bins/$(CONFIG)/hpack_table_test
chttp2_stream_map_test: bins/$(CONFIG)/chttp2_stream_map_test
hpack_parser_test: bins/$(CONFIG)/hpack_parser_test
transport_metadata_test: bins/$(CONFIG)/transport_metadata_test
chttp2_status_conversion_test: bins/$(CONFIG)/chttp2_status_conversion_test
chttp2_transport_end2end_test: bins/$(CONFIG)/chttp2_transport_end2end_test
tcp_posix_test: bins/$(CONFIG)/tcp_posix_test
dualstack_socket_test: bins/$(CONFIG)/dualstack_socket_test
no_server_test: bins/$(CONFIG)/no_server_test
resolve_address_test: bins/$(CONFIG)/resolve_address_test
sockaddr_utils_test: bins/$(CONFIG)/sockaddr_utils_test
tcp_server_posix_test: bins/$(CONFIG)/tcp_server_posix_test
tcp_client_posix_test: bins/$(CONFIG)/tcp_client_posix_test
grpc_channel_stack_test: bins/$(CONFIG)/grpc_channel_stack_test
metadata_buffer_test: bins/$(CONFIG)/metadata_buffer_test
grpc_completion_queue_test: bins/$(CONFIG)/grpc_completion_queue_test
grpc_completion_queue_benchmark: bins/$(CONFIG)/grpc_completion_queue_benchmark
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
census_trace_store_test: bins/$(CONFIG)/census_trace_store_test
census_stats_store_test: bins/$(CONFIG)/census_stats_store_test
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
census_window_stats_test: bins/$(CONFIG)/census_window_stats_test
census_statistics_quick_test: bins/$(CONFIG)/census_statistics_quick_test
census_statistics_small_log_test: bins/$(CONFIG)/census_statistics_small_log_test
census_statistics_performance_test: bins/$(CONFIG)/census_statistics_performance_test
census_statistics_multiple_writers_test: bins/$(CONFIG)/census_statistics_multiple_writers_test
census_statistics_multiple_writers_circular_buffer_test: bins/$(CONFIG)/census_statistics_multiple_writers_circular_buffer_test
census_stub_test: bins/$(CONFIG)/census_stub_test
census_hash_table_test: bins/$(CONFIG)/census_hash_table_test
fling_server: bins/$(CONFIG)/fling_server
fling_client: bins/$(CONFIG)/fling_client
fling_test: bins/$(CONFIG)/fling_test
echo_server: bins/$(CONFIG)/echo_server
echo_client: bins/$(CONFIG)/echo_client
echo_test: bins/$(CONFIG)/echo_test
low_level_ping_pong_benchmark: bins/$(CONFIG)/low_level_ping_pong_benchmark
message_compress_test: bins/$(CONFIG)/message_compress_test
bin_encoder_test: bins/$(CONFIG)/bin_encoder_test
secure_endpoint_test: bins/$(CONFIG)/secure_endpoint_test
httpcli_format_request_test: bins/$(CONFIG)/httpcli_format_request_test
httpcli_parser_test: bins/$(CONFIG)/httpcli_parser_test
httpcli_test: bins/$(CONFIG)/httpcli_test
grpc_credentials_test: bins/$(CONFIG)/grpc_credentials_test
grpc_fetch_oauth2: bins/$(CONFIG)/grpc_fetch_oauth2
grpc_base64_test: bins/$(CONFIG)/grpc_base64_test
grpc_json_token_test: bins/$(CONFIG)/grpc_json_token_test
timeout_encoding_test: bins/$(CONFIG)/timeout_encoding_test
fd_posix_test: bins/$(CONFIG)/fd_posix_test
fling_stream_test: bins/$(CONFIG)/fling_stream_test
lame_client_test: bins/$(CONFIG)/lame_client_test
thread_pool_test: bins/$(CONFIG)/thread_pool_test
status_test: bins/$(CONFIG)/status_test
sync_client_async_server_test: bins/$(CONFIG)/sync_client_async_server_test
qps_client: bins/$(CONFIG)/qps_client
qps_server: bins/$(CONFIG)/qps_server
interop_server: bins/$(CONFIG)/interop_server
interop_client: bins/$(CONFIG)/interop_client
TIPS client implementation.
10 years ago
tips_client: bins/$(CONFIG)/tips_client
tips_client_test: bins/$(CONFIG)/tips_client_test
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
end2end_test: bins/$(CONFIG)/end2end_test
channel_arguments_test: bins/$(CONFIG)/channel_arguments_test
Add ServiceAccount Credentials wrapping and handle credentials creation failure. Change on 2015/01/09 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83634736
10 years ago
credentials_test: bins/$(CONFIG)/credentials_test
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
alarm_test: bins/$(CONFIG)/alarm_test
alarm_list_test: bins/$(CONFIG)/alarm_list_test
alarm_heap_test: bins/$(CONFIG)/alarm_heap_test
time_test: bins/$(CONFIG)/time_test
chttp2_fake_security_cancel_after_accept_test: bins/$(CONFIG)/chttp2_fake_security_cancel_after_accept_test
chttp2_fake_security_cancel_after_accept_and_writes_closed_test: bins/$(CONFIG)/chttp2_fake_security_cancel_after_accept_and_writes_closed_test
chttp2_fake_security_cancel_after_invoke_test: bins/$(CONFIG)/chttp2_fake_security_cancel_after_invoke_test
chttp2_fake_security_cancel_before_invoke_test: bins/$(CONFIG)/chttp2_fake_security_cancel_before_invoke_test
chttp2_fake_security_cancel_in_a_vacuum_test: bins/$(CONFIG)/chttp2_fake_security_cancel_in_a_vacuum_test
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
chttp2_fake_security_census_simple_request_test: bins/$(CONFIG)/chttp2_fake_security_census_simple_request_test
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
chttp2_fake_security_disappearing_server_test: bins/$(CONFIG)/chttp2_fake_security_disappearing_server_test
chttp2_fake_security_early_server_shutdown_finishes_inflight_calls_test: bins/$(CONFIG)/chttp2_fake_security_early_server_shutdown_finishes_inflight_calls_test
chttp2_fake_security_early_server_shutdown_finishes_tags_test: bins/$(CONFIG)/chttp2_fake_security_early_server_shutdown_finishes_tags_test
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
chttp2_fake_security_graceful_server_shutdown_test: bins/$(CONFIG)/chttp2_fake_security_graceful_server_shutdown_test
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
chttp2_fake_security_invoke_large_request_test: bins/$(CONFIG)/chttp2_fake_security_invoke_large_request_test
chttp2_fake_security_max_concurrent_streams_test: bins/$(CONFIG)/chttp2_fake_security_max_concurrent_streams_test
chttp2_fake_security_no_op_test: bins/$(CONFIG)/chttp2_fake_security_no_op_test
chttp2_fake_security_ping_pong_streaming_test: bins/$(CONFIG)/chttp2_fake_security_ping_pong_streaming_test
chttp2_fake_security_request_response_with_binary_metadata_and_payload_test: bins/$(CONFIG)/chttp2_fake_security_request_response_with_binary_metadata_and_payload_test
chttp2_fake_security_request_response_with_metadata_and_payload_test: bins/$(CONFIG)/chttp2_fake_security_request_response_with_metadata_and_payload_test
chttp2_fake_security_request_response_with_payload_test: bins/$(CONFIG)/chttp2_fake_security_request_response_with_payload_test
chttp2_fake_security_request_response_with_trailing_metadata_and_payload_test: bins/$(CONFIG)/chttp2_fake_security_request_response_with_trailing_metadata_and_payload_test
chttp2_fake_security_simple_delayed_request_test: bins/$(CONFIG)/chttp2_fake_security_simple_delayed_request_test
chttp2_fake_security_simple_request_test: bins/$(CONFIG)/chttp2_fake_security_simple_request_test
chttp2_fake_security_thread_stress_test: bins/$(CONFIG)/chttp2_fake_security_thread_stress_test
chttp2_fake_security_writes_done_hangs_with_pending_read_test: bins/$(CONFIG)/chttp2_fake_security_writes_done_hangs_with_pending_read_test
chttp2_fullstack_cancel_after_accept_test: bins/$(CONFIG)/chttp2_fullstack_cancel_after_accept_test
chttp2_fullstack_cancel_after_accept_and_writes_closed_test: bins/$(CONFIG)/chttp2_fullstack_cancel_after_accept_and_writes_closed_test
chttp2_fullstack_cancel_after_invoke_test: bins/$(CONFIG)/chttp2_fullstack_cancel_after_invoke_test
chttp2_fullstack_cancel_before_invoke_test: bins/$(CONFIG)/chttp2_fullstack_cancel_before_invoke_test
chttp2_fullstack_cancel_in_a_vacuum_test: bins/$(CONFIG)/chttp2_fullstack_cancel_in_a_vacuum_test
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
chttp2_fullstack_census_simple_request_test: bins/$(CONFIG)/chttp2_fullstack_census_simple_request_test
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
chttp2_fullstack_disappearing_server_test: bins/$(CONFIG)/chttp2_fullstack_disappearing_server_test
chttp2_fullstack_early_server_shutdown_finishes_inflight_calls_test: bins/$(CONFIG)/chttp2_fullstack_early_server_shutdown_finishes_inflight_calls_test
chttp2_fullstack_early_server_shutdown_finishes_tags_test: bins/$(CONFIG)/chttp2_fullstack_early_server_shutdown_finishes_tags_test
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
chttp2_fullstack_graceful_server_shutdown_test: bins/$(CONFIG)/chttp2_fullstack_graceful_server_shutdown_test
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
chttp2_fullstack_invoke_large_request_test: bins/$(CONFIG)/chttp2_fullstack_invoke_large_request_test
chttp2_fullstack_max_concurrent_streams_test: bins/$(CONFIG)/chttp2_fullstack_max_concurrent_streams_test
chttp2_fullstack_no_op_test: bins/$(CONFIG)/chttp2_fullstack_no_op_test
chttp2_fullstack_ping_pong_streaming_test: bins/$(CONFIG)/chttp2_fullstack_ping_pong_streaming_test
chttp2_fullstack_request_response_with_binary_metadata_and_payload_test: bins/$(CONFIG)/chttp2_fullstack_request_response_with_binary_metadata_and_payload_test
chttp2_fullstack_request_response_with_metadata_and_payload_test: bins/$(CONFIG)/chttp2_fullstack_request_response_with_metadata_and_payload_test
chttp2_fullstack_request_response_with_payload_test: bins/$(CONFIG)/chttp2_fullstack_request_response_with_payload_test
chttp2_fullstack_request_response_with_trailing_metadata_and_payload_test: bins/$(CONFIG)/chttp2_fullstack_request_response_with_trailing_metadata_and_payload_test
chttp2_fullstack_simple_delayed_request_test: bins/$(CONFIG)/chttp2_fullstack_simple_delayed_request_test
chttp2_fullstack_simple_request_test: bins/$(CONFIG)/chttp2_fullstack_simple_request_test
chttp2_fullstack_thread_stress_test: bins/$(CONFIG)/chttp2_fullstack_thread_stress_test
chttp2_fullstack_writes_done_hangs_with_pending_read_test: bins/$(CONFIG)/chttp2_fullstack_writes_done_hangs_with_pending_read_test
chttp2_simple_ssl_fullstack_cancel_after_accept_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_after_accept_test
chttp2_simple_ssl_fullstack_cancel_after_accept_and_writes_closed_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_after_accept_and_writes_closed_test
chttp2_simple_ssl_fullstack_cancel_after_invoke_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_after_invoke_test
chttp2_simple_ssl_fullstack_cancel_before_invoke_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_before_invoke_test
chttp2_simple_ssl_fullstack_cancel_in_a_vacuum_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_in_a_vacuum_test
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
chttp2_simple_ssl_fullstack_census_simple_request_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_census_simple_request_test
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
chttp2_simple_ssl_fullstack_disappearing_server_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_disappearing_server_test
chttp2_simple_ssl_fullstack_early_server_shutdown_finishes_inflight_calls_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_early_server_shutdown_finishes_inflight_calls_test
chttp2_simple_ssl_fullstack_early_server_shutdown_finishes_tags_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_early_server_shutdown_finishes_tags_test
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
chttp2_simple_ssl_fullstack_graceful_server_shutdown_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_graceful_server_shutdown_test
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
chttp2_simple_ssl_fullstack_invoke_large_request_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_invoke_large_request_test
chttp2_simple_ssl_fullstack_max_concurrent_streams_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_max_concurrent_streams_test
chttp2_simple_ssl_fullstack_no_op_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_no_op_test
chttp2_simple_ssl_fullstack_ping_pong_streaming_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_ping_pong_streaming_test
chttp2_simple_ssl_fullstack_request_response_with_binary_metadata_and_payload_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_request_response_with_binary_metadata_and_payload_test
chttp2_simple_ssl_fullstack_request_response_with_metadata_and_payload_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_request_response_with_metadata_and_payload_test
chttp2_simple_ssl_fullstack_request_response_with_payload_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_request_response_with_payload_test
chttp2_simple_ssl_fullstack_request_response_with_trailing_metadata_and_payload_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_request_response_with_trailing_metadata_and_payload_test
chttp2_simple_ssl_fullstack_simple_delayed_request_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_simple_delayed_request_test
chttp2_simple_ssl_fullstack_simple_request_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_simple_request_test
chttp2_simple_ssl_fullstack_thread_stress_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_thread_stress_test
chttp2_simple_ssl_fullstack_writes_done_hangs_with_pending_read_test: bins/$(CONFIG)/chttp2_simple_ssl_fullstack_writes_done_hangs_with_pending_read_test
chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_accept_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_accept_test
chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_accept_and_writes_closed_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_accept_and_writes_closed_test
chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_invoke_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_invoke_test
chttp2_simple_ssl_with_oauth2_fullstack_cancel_before_invoke_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_before_invoke_test
chttp2_simple_ssl_with_oauth2_fullstack_cancel_in_a_vacuum_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_in_a_vacuum_test
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
chttp2_simple_ssl_with_oauth2_fullstack_census_simple_request_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_census_simple_request_test
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
chttp2_simple_ssl_with_oauth2_fullstack_disappearing_server_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_disappearing_server_test
chttp2_simple_ssl_with_oauth2_fullstack_early_server_shutdown_finishes_inflight_calls_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_early_server_shutdown_finishes_inflight_calls_test
chttp2_simple_ssl_with_oauth2_fullstack_early_server_shutdown_finishes_tags_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_early_server_shutdown_finishes_tags_test
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
chttp2_simple_ssl_with_oauth2_fullstack_graceful_server_shutdown_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_graceful_server_shutdown_test
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
chttp2_simple_ssl_with_oauth2_fullstack_invoke_large_request_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_invoke_large_request_test
chttp2_simple_ssl_with_oauth2_fullstack_max_concurrent_streams_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_max_concurrent_streams_test
chttp2_simple_ssl_with_oauth2_fullstack_no_op_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_no_op_test
chttp2_simple_ssl_with_oauth2_fullstack_ping_pong_streaming_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_ping_pong_streaming_test
chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_binary_metadata_and_payload_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_binary_metadata_and_payload_test
chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_metadata_and_payload_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_metadata_and_payload_test
chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_payload_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_payload_test
chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_trailing_metadata_and_payload_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_trailing_metadata_and_payload_test
chttp2_simple_ssl_with_oauth2_fullstack_simple_delayed_request_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_simple_delayed_request_test
chttp2_simple_ssl_with_oauth2_fullstack_simple_request_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_simple_request_test
chttp2_simple_ssl_with_oauth2_fullstack_thread_stress_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_thread_stress_test
chttp2_simple_ssl_with_oauth2_fullstack_writes_done_hangs_with_pending_read_test: bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_writes_done_hangs_with_pending_read_test
chttp2_socket_pair_cancel_after_accept_test: bins/$(CONFIG)/chttp2_socket_pair_cancel_after_accept_test
chttp2_socket_pair_cancel_after_accept_and_writes_closed_test: bins/$(CONFIG)/chttp2_socket_pair_cancel_after_accept_and_writes_closed_test
chttp2_socket_pair_cancel_after_invoke_test: bins/$(CONFIG)/chttp2_socket_pair_cancel_after_invoke_test
chttp2_socket_pair_cancel_before_invoke_test: bins/$(CONFIG)/chttp2_socket_pair_cancel_before_invoke_test
chttp2_socket_pair_cancel_in_a_vacuum_test: bins/$(CONFIG)/chttp2_socket_pair_cancel_in_a_vacuum_test
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
chttp2_socket_pair_census_simple_request_test: bins/$(CONFIG)/chttp2_socket_pair_census_simple_request_test
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
chttp2_socket_pair_disappearing_server_test: bins/$(CONFIG)/chttp2_socket_pair_disappearing_server_test
chttp2_socket_pair_early_server_shutdown_finishes_inflight_calls_test: bins/$(CONFIG)/chttp2_socket_pair_early_server_shutdown_finishes_inflight_calls_test
chttp2_socket_pair_early_server_shutdown_finishes_tags_test: bins/$(CONFIG)/chttp2_socket_pair_early_server_shutdown_finishes_tags_test
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
chttp2_socket_pair_graceful_server_shutdown_test: bins/$(CONFIG)/chttp2_socket_pair_graceful_server_shutdown_test
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
chttp2_socket_pair_invoke_large_request_test: bins/$(CONFIG)/chttp2_socket_pair_invoke_large_request_test
chttp2_socket_pair_max_concurrent_streams_test: bins/$(CONFIG)/chttp2_socket_pair_max_concurrent_streams_test
chttp2_socket_pair_no_op_test: bins/$(CONFIG)/chttp2_socket_pair_no_op_test
chttp2_socket_pair_ping_pong_streaming_test: bins/$(CONFIG)/chttp2_socket_pair_ping_pong_streaming_test
chttp2_socket_pair_request_response_with_binary_metadata_and_payload_test: bins/$(CONFIG)/chttp2_socket_pair_request_response_with_binary_metadata_and_payload_test
chttp2_socket_pair_request_response_with_metadata_and_payload_test: bins/$(CONFIG)/chttp2_socket_pair_request_response_with_metadata_and_payload_test
chttp2_socket_pair_request_response_with_payload_test: bins/$(CONFIG)/chttp2_socket_pair_request_response_with_payload_test
chttp2_socket_pair_request_response_with_trailing_metadata_and_payload_test: bins/$(CONFIG)/chttp2_socket_pair_request_response_with_trailing_metadata_and_payload_test
chttp2_socket_pair_simple_delayed_request_test: bins/$(CONFIG)/chttp2_socket_pair_simple_delayed_request_test
chttp2_socket_pair_simple_request_test: bins/$(CONFIG)/chttp2_socket_pair_simple_request_test
chttp2_socket_pair_thread_stress_test: bins/$(CONFIG)/chttp2_socket_pair_thread_stress_test
chttp2_socket_pair_writes_done_hangs_with_pending_read_test: bins/$(CONFIG)/chttp2_socket_pair_writes_done_hangs_with_pending_read_test
chttp2_socket_pair_one_byte_at_a_time_cancel_after_accept_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_after_accept_test
chttp2_socket_pair_one_byte_at_a_time_cancel_after_accept_and_writes_closed_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_after_accept_and_writes_closed_test
chttp2_socket_pair_one_byte_at_a_time_cancel_after_invoke_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_after_invoke_test
chttp2_socket_pair_one_byte_at_a_time_cancel_before_invoke_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_before_invoke_test
chttp2_socket_pair_one_byte_at_a_time_cancel_in_a_vacuum_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_in_a_vacuum_test
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
chttp2_socket_pair_one_byte_at_a_time_census_simple_request_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_census_simple_request_test
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
chttp2_socket_pair_one_byte_at_a_time_disappearing_server_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_disappearing_server_test
chttp2_socket_pair_one_byte_at_a_time_early_server_shutdown_finishes_inflight_calls_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_early_server_shutdown_finishes_inflight_calls_test
chttp2_socket_pair_one_byte_at_a_time_early_server_shutdown_finishes_tags_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_early_server_shutdown_finishes_tags_test
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
chttp2_socket_pair_one_byte_at_a_time_graceful_server_shutdown_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_graceful_server_shutdown_test
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
chttp2_socket_pair_one_byte_at_a_time_invoke_large_request_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_invoke_large_request_test
chttp2_socket_pair_one_byte_at_a_time_max_concurrent_streams_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_max_concurrent_streams_test
chttp2_socket_pair_one_byte_at_a_time_no_op_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_no_op_test
chttp2_socket_pair_one_byte_at_a_time_ping_pong_streaming_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_ping_pong_streaming_test
chttp2_socket_pair_one_byte_at_a_time_request_response_with_binary_metadata_and_payload_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_request_response_with_binary_metadata_and_payload_test
chttp2_socket_pair_one_byte_at_a_time_request_response_with_metadata_and_payload_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_request_response_with_metadata_and_payload_test
chttp2_socket_pair_one_byte_at_a_time_request_response_with_payload_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_request_response_with_payload_test
chttp2_socket_pair_one_byte_at_a_time_request_response_with_trailing_metadata_and_payload_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_request_response_with_trailing_metadata_and_payload_test
chttp2_socket_pair_one_byte_at_a_time_simple_delayed_request_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_simple_delayed_request_test
chttp2_socket_pair_one_byte_at_a_time_simple_request_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_simple_request_test
chttp2_socket_pair_one_byte_at_a_time_thread_stress_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_thread_stress_test
chttp2_socket_pair_one_byte_at_a_time_writes_done_hangs_with_pending_read_test: bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_writes_done_hangs_with_pending_read_test
Build debug/optimized builds into different target directories. This will make auto-running different test configurations a little easier. Change on 2014/12/19 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82555939
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
run_dep_checks:
$(OPENSSL_ALPN_CHECK_CMD) || true
$(ZLIB_CHECK_CMD) || true
Support compiling zlib with ?SAN
10 years ago
libs/$(CONFIG)/zlib/libz.a:
Fixing parallel build. The source code of the libraries depends on openssl being built, as it will setup the headers. Since we don't know exactly which source depends on openssl (we could start flagging it in build.json, but that'd be painful), we just flag all the source for every ssl-enabled library to depend on openssl being built.
10 years ago
$(E) "[MAKE] Building zlib"
$(Q)(cd third_party/zlib ; CC="$(CC)" CFLAGS="-fPIC -fvisibility=hidden $(CPPFLAGS_$(CONFIG))" ./configure --static)
$(Q)$(MAKE) -C third_party/zlib clean
$(Q)$(MAKE) -C third_party/zlib
$(Q)mkdir -p libs/$(CONFIG)/zlib
$(Q)cp third_party/zlib/libz.a libs/$(CONFIG)/zlib
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
Compile OpenSSL differently for each config. MSAN, TSAN, ASAN need to see all the code. Valgrind, MSAN want -DPURIFY.
10 years ago
libs/$(CONFIG)/openssl/libssl.a:
Fixing parallel build. The source code of the libraries depends on openssl being built, as it will setup the headers. Since we don't know exactly which source depends on openssl (we could start flagging it in build.json, but that'd be painful), we just flag all the source for every ssl-enabled library to depend on openssl being built.
10 years ago
$(E) "[MAKE] Building openssl"
$(Q)(cd third_party/openssl ; CC="$(CC) -fPIC -fvisibility=hidden $(CPPFLAGS_$(CONFIG)) $(OPENSSL_CFLAGS_$(CONFIG))" ./config $(OPENSSL_CONFIG_$(CONFIG)))
$(Q)$(MAKE) -C third_party/openssl clean
$(Q)$(MAKE) -C third_party/openssl build_crypto build_ssl
$(Q)mkdir -p libs/$(CONFIG)/openssl
$(Q)cp third_party/openssl/libssl.a third_party/openssl/libcrypto.a libs/$(CONFIG)/openssl
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
Better C/C++ split in the Makefile targets, to better prepare slicing it into two different Makefiles. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81076626
10 years ago
static: static_c static_cxx
Initial import.
10 years ago
Generate dependencies when compiling the .o If we don't have a .o, we don't need dependencies, so skip the setup step of building them. Instead, when building the .o, we generate the .dep and output it (to objs/ to make the rules a little easier). The next run of make will include the dep file and any dependencies will be picked up. This change also disables the disabling of dependency checking if we have zlib or openssl being compiled from third_party. Additionally it inverts the logic for including dependencies from ONLY if we are doing a clean to ONLY if we are NOT doing a clean.
10 years ago
static_c: libs/$(CONFIG)/libgpr.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgrpc_unsecure.a
Initial import.
10 years ago
Generate dependencies when compiling the .o If we don't have a .o, we don't need dependencies, so skip the setup step of building them. Instead, when building the .o, we generate the .dep and output it (to objs/ to make the rules a little easier). The next run of make will include the dep file and any dependencies will be picked up. This change also disables the disabling of dependency checking if we have zlib or openssl being compiled from third_party. Additionally it inverts the logic for including dependencies from ONLY if we are doing a clean to ONLY if we are NOT doing a clean.
10 years ago
static_cxx: libs/$(CONFIG)/libgrpc++.a
Initial import.
10 years ago
Better C/C++ split in the Makefile targets, to better prepare slicing it into two different Makefiles. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81076626
10 years ago
shared: shared_c shared_cxx
Initial import.
10 years ago
Generate dependencies when compiling the .o If we don't have a .o, we don't need dependencies, so skip the setup step of building them. Instead, when building the .o, we generate the .dep and output it (to objs/ to make the rules a little easier). The next run of make will include the dep file and any dependencies will be picked up. This change also disables the disabling of dependency checking if we have zlib or openssl being compiled from third_party. Additionally it inverts the logic for including dependencies from ONLY if we are doing a clean to ONLY if we are NOT doing a clean.
10 years ago
shared_c: libs/$(CONFIG)/libgpr.$(SHARED_EXT) libs/$(CONFIG)/libgrpc.$(SHARED_EXT) libs/$(CONFIG)/libgrpc_unsecure.$(SHARED_EXT)
Initial import.
10 years ago
Generate dependencies when compiling the .o If we don't have a .o, we don't need dependencies, so skip the setup step of building them. Instead, when building the .o, we generate the .dep and output it (to objs/ to make the rules a little easier). The next run of make will include the dep file and any dependencies will be picked up. This change also disables the disabling of dependency checking if we have zlib or openssl being compiled from third_party. Additionally it inverts the logic for including dependencies from ONLY if we are doing a clean to ONLY if we are NOT doing a clean.
10 years ago
shared_cxx: libs/$(CONFIG)/libgrpc++.$(SHARED_EXT)
Initial import.
10 years ago
Better C/C++ split in the Makefile targets, to better prepare slicing it into two different Makefiles. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81076626
10 years ago
privatelibs: privatelibs_c privatelibs_cxx
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
privatelibs_c: libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_cancel_after_accept.a libs/$(CONFIG)/libend2end_test_cancel_after_accept_and_writes_closed.a libs/$(CONFIG)/libend2end_test_cancel_after_invoke.a libs/$(CONFIG)/libend2end_test_cancel_before_invoke.a libs/$(CONFIG)/libend2end_test_cancel_in_a_vacuum.a libs/$(CONFIG)/libend2end_test_census_simple_request.a libs/$(CONFIG)/libend2end_test_disappearing_server.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_inflight_calls.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_tags.a libs/$(CONFIG)/libend2end_test_graceful_server_shutdown.a libs/$(CONFIG)/libend2end_test_invoke_large_request.a libs/$(CONFIG)/libend2end_test_max_concurrent_streams.a libs/$(CONFIG)/libend2end_test_no_op.a libs/$(CONFIG)/libend2end_test_ping_pong_streaming.a libs/$(CONFIG)/libend2end_test_request_response_with_binary_metadata_and_payload.a libs/$(CONFIG)/libend2end_test_request_response_with_metadata_and_payload.a libs/$(CONFIG)/libend2end_test_request_response_with_payload.a libs/$(CONFIG)/libend2end_test_request_response_with_trailing_metadata_and_payload.a libs/$(CONFIG)/libend2end_test_simple_delayed_request.a libs/$(CONFIG)/libend2end_test_simple_request.a libs/$(CONFIG)/libend2end_test_thread_stress.a libs/$(CONFIG)/libend2end_test_writes_done_hangs_with_pending_read.a libs/$(CONFIG)/libend2end_certs.a
Better C/C++ split in the Makefile targets, to better prepare slicing it into two different Makefiles. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81076626
10 years ago
TIPS client implementation.
10 years ago
privatelibs_cxx: libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libtips_client_lib.a
Better C/C++ split in the Makefile targets, to better prepare slicing it into two different Makefiles. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81076626
10 years ago
buildtests: buildtests_c buildtests_cxx
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
buildtests_c: privatelibs_c bins/$(CONFIG)/grpc_byte_buffer_reader_test bins/$(CONFIG)/gpr_cancellable_test bins/$(CONFIG)/gpr_log_test bins/$(CONFIG)/gpr_useful_test bins/$(CONFIG)/gpr_cmdline_test bins/$(CONFIG)/gpr_histogram_test bins/$(CONFIG)/gpr_host_port_test bins/$(CONFIG)/gpr_slice_buffer_test bins/$(CONFIG)/gpr_slice_test bins/$(CONFIG)/gpr_string_test bins/$(CONFIG)/gpr_sync_test bins/$(CONFIG)/gpr_thd_test bins/$(CONFIG)/gpr_time_test bins/$(CONFIG)/murmur_hash_test bins/$(CONFIG)/grpc_stream_op_test bins/$(CONFIG)/alpn_test bins/$(CONFIG)/time_averaged_stats_test bins/$(CONFIG)/chttp2_stream_encoder_test bins/$(CONFIG)/hpack_table_test bins/$(CONFIG)/chttp2_stream_map_test bins/$(CONFIG)/hpack_parser_test bins/$(CONFIG)/transport_metadata_test bins/$(CONFIG)/chttp2_status_conversion_test bins/$(CONFIG)/chttp2_transport_end2end_test bins/$(CONFIG)/tcp_posix_test bins/$(CONFIG)/dualstack_socket_test bins/$(CONFIG)/no_server_test bins/$(CONFIG)/resolve_address_test bins/$(CONFIG)/sockaddr_utils_test bins/$(CONFIG)/tcp_server_posix_test bins/$(CONFIG)/tcp_client_posix_test bins/$(CONFIG)/grpc_channel_stack_test bins/$(CONFIG)/metadata_buffer_test bins/$(CONFIG)/grpc_completion_queue_test bins/$(CONFIG)/census_window_stats_test bins/$(CONFIG)/census_statistics_quick_test bins/$(CONFIG)/census_statistics_small_log_test bins/$(CONFIG)/census_statistics_performance_test bins/$(CONFIG)/census_statistics_multiple_writers_test bins/$(CONFIG)/census_statistics_multiple_writers_circular_buffer_test bins/$(CONFIG)/census_stub_test bins/$(CONFIG)/census_hash_table_test bins/$(CONFIG)/fling_server bins/$(CONFIG)/fling_client bins/$(CONFIG)/fling_test bins/$(CONFIG)/echo_server bins/$(CONFIG)/echo_client bins/$(CONFIG)/echo_test bins/$(CONFIG)/message_compress_test bins/$(CONFIG)/bin_encoder_test bins/$(CONFIG)/secure_endpoint_test bins/$(CONFIG)/httpcli_format_request_test bins/$(CONFIG)/httpcli_parser_test bins/$(CONFIG)/httpcli_test bins/$(CONFIG)/grpc_credentials_test bins/$(CONFIG)/grpc_base64_test bins/$(CONFIG)/grpc_json_token_test bins/$(CONFIG)/timeout_encoding_test bins/$(CONFIG)/fd_posix_test bins/$(CONFIG)/fling_stream_test bins/$(CONFIG)/lame_client_test bins/$(CONFIG)/alarm_test bins/$(CONFIG)/alarm_list_test bins/$(CONFIG)/alarm_heap_test bins/$(CONFIG)/time_test bins/$(CONFIG)/chttp2_fake_security_cancel_after_accept_test bins/$(CONFIG)/chttp2_fake_security_cancel_after_accept_and_writes_closed_test bins/$(CONFIG)/chttp2_fake_security_cancel_after_invoke_test bins/$(CONFIG)/chttp2_fake_security_cancel_before_invoke_test bins/$(CONFIG)/chttp2_fake_security_cancel_in_a_vacuum_test bins/$(CONFIG)/chttp2_fake_security_census_simple_request_test bins/$(CONFIG)/chttp2_fake_security_disappearing_server_test bins/$(CONFIG)/chttp2_fake_security_early_server_shutdown_finishes_inflight_calls_test bins/$(CONFIG)/chttp2_fake_security_early_server_shutdown_finishes_tags_test bins/$(CONFIG)/chttp2_fake_security_graceful_server_shutdown_test bins/$(CONFIG)/chttp2_fake_security_invoke_large_request_test bins/$(CONFIG)/chttp2_fake_security_max_concurrent_streams_test bins/$(CONFIG)/chttp2_fake_security_no_op_test bins/$(CONFIG)/chttp2_fake_security_ping_pong_streaming_test bins/$(CONFIG)/chttp2_fake_security_request_response_with_binary_metadata_and_payload_test bins/$(CONFIG)/chttp2_fake_security_request_response_with_metadata_and_payload_test bins/$(CONFIG)/chttp2_fake_security_request_response_with_payload_test bins/$(CONFIG)/chttp2_fake_security_request_response_with_trailing_metadata_and_payload_test bins/$(CONFIG)/chttp2_fake_security_simple_delayed_request_test bins/$(CONFIG)/chttp2_fake_security_simple_request_test bins/$(CONFIG)/chttp2_fake_security_thread_stress_test bins/$(CONFIG)/chttp2_fake_security_writes_done_hangs_with_pending_read_test bins/$(CONFIG)/chttp2_fullstack_cancel_after_accept_test bins/$(CONFIG)/chttp2_fullstack_cancel_after_accept_and_writes_closed_test bins/$(CONFIG)/chttp2_fullstack_cancel_after_invoke_test bins/$(CONFIG)/chttp2_fullstack_cancel_before_invoke_test bins/$(CONFIG)/chttp2_fullstack
Better C/C++ split in the Makefile targets, to better prepare slicing it into two different Makefiles. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81076626
10 years ago
TIPS client implementation.
10 years ago
buildtests_cxx: privatelibs_cxx bins/$(CONFIG)/thread_pool_test bins/$(CONFIG)/status_test bins/$(CONFIG)/sync_client_async_server_test bins/$(CONFIG)/qps_client bins/$(CONFIG)/qps_server bins/$(CONFIG)/interop_server bins/$(CONFIG)/interop_client bins/$(CONFIG)/tips_client bins/$(CONFIG)/tips_client_test bins/$(CONFIG)/end2end_test bins/$(CONFIG)/channel_arguments_test bins/$(CONFIG)/credentials_test
Better C/C++ split in the Makefile targets, to better prepare slicing it into two different Makefiles. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81076626
10 years ago
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
test: test_c test_cxx
Better C/C++ split in the Makefile targets, to better prepare slicing it into two different Makefiles. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81076626
10 years ago
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
test_c: buildtests_c
Initial import.
10 years ago
$(E) "[RUN] Testing grpc_byte_buffer_reader_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/grpc_byte_buffer_reader_test || ( echo test grpc_byte_buffer_reader_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing gpr_cancellable_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/gpr_cancellable_test || ( echo test gpr_cancellable_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing gpr_log_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/gpr_log_test || ( echo test gpr_log_test failed ; exit 1 )
Add a test for useful.h. Additionally add rotl, rotr which will be needed to implement some hashing functions shortly. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82149559
10 years ago
$(E) "[RUN] Testing gpr_useful_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/gpr_useful_test || ( echo test gpr_useful_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing gpr_cmdline_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/gpr_cmdline_test || ( echo test gpr_cmdline_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing gpr_histogram_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/gpr_histogram_test || ( echo test gpr_histogram_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing gpr_host_port_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/gpr_host_port_test || ( echo test gpr_host_port_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing gpr_slice_buffer_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/gpr_slice_buffer_test || ( echo test gpr_slice_buffer_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing gpr_slice_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/gpr_slice_test || ( echo test gpr_slice_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing gpr_string_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/gpr_string_test || ( echo test gpr_string_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing gpr_sync_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/gpr_sync_test || ( echo test gpr_sync_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing gpr_thd_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/gpr_thd_test || ( echo test gpr_thd_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing gpr_time_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/gpr_time_test || ( echo test gpr_time_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing murmur_hash_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/murmur_hash_test || ( echo test murmur_hash_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing grpc_stream_op_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/grpc_stream_op_test || ( echo test grpc_stream_op_test failed ; exit 1 )
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[RUN] Testing alpn_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/alpn_test || ( echo test alpn_test failed ; exit 1 )
Add time averaged stats for future alarm list work. This is a trivial C89-ification of the []2 implementation of the same idea. Indeed the implementation files and tests have been branched from the [] versions. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82168822
10 years ago
$(E) "[RUN] Testing time_averaged_stats_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/time_averaged_stats_test || ( echo test time_averaged_stats_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_stream_encoder_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_stream_encoder_test || ( echo test chttp2_stream_encoder_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing hpack_table_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/hpack_table_test || ( echo test hpack_table_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_stream_map_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_stream_map_test || ( echo test chttp2_stream_map_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing hpack_parser_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/hpack_parser_test || ( echo test hpack_parser_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing transport_metadata_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/transport_metadata_test || ( echo test transport_metadata_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_status_conversion_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_status_conversion_test || ( echo test chttp2_status_conversion_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_transport_end2end_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_transport_end2end_test || ( echo test chttp2_transport_end2end_test failed ; exit 1 )
Introducing iomgr. Move eventmanager and platform dependent endpoint functionality into a single library called 'iomgr'. This is primarily to prepare for a Windows port - where posix socket semantics lead to poor quality code. Mostly this is a code movement CL, with some small changes to help prepare the way for porting: - em style fd objects can only be held internally in iomgr, and own their memory - added grpc_iomgr_create_endpoint_pair() to accomodate the common pattern of creating a tcp endpoint from the output of socketpair - this will help keep our tests portable - separated em alarm interface into a separate file, as this part of event manager is needed higher up the stack - made the eventmanager bits a true singleton, simplifying API's across the stack as there's no longer a reason to carry a pointer there. Initial design document is here: https://docs.google.com/document/d/1VmafcHvvrP5kwtQkz84R5yXF7u7fW-9Pn0bkSUQHDt8/edit?disco=AAAAARNByxg Change on 2014/12/09 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81716456
10 years ago
$(E) "[RUN] Testing tcp_posix_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/tcp_posix_test || ( echo test tcp_posix_test failed ; exit 1 )
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[RUN] Testing dualstack_socket_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/dualstack_socket_test || ( echo test dualstack_socket_test failed ; exit 1 )
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[RUN] Testing no_server_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/no_server_test || ( echo test no_server_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing resolve_address_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/resolve_address_test || ( echo test resolve_address_test failed ; exit 1 )
Introducing iomgr. Move eventmanager and platform dependent endpoint functionality into a single library called 'iomgr'. This is primarily to prepare for a Windows port - where posix socket semantics lead to poor quality code. Mostly this is a code movement CL, with some small changes to help prepare the way for porting: - em style fd objects can only be held internally in iomgr, and own their memory - added grpc_iomgr_create_endpoint_pair() to accomodate the common pattern of creating a tcp endpoint from the output of socketpair - this will help keep our tests portable - separated em alarm interface into a separate file, as this part of event manager is needed higher up the stack - made the eventmanager bits a true singleton, simplifying API's across the stack as there's no longer a reason to carry a pointer there. Initial design document is here: https://docs.google.com/document/d/1VmafcHvvrP5kwtQkz84R5yXF7u7fW-9Pn0bkSUQHDt8/edit?disco=AAAAARNByxg Change on 2014/12/09 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81716456
10 years ago
$(E) "[RUN] Testing sockaddr_utils_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/sockaddr_utils_test || ( echo test sockaddr_utils_test failed ; exit 1 )
Introducing iomgr. Move eventmanager and platform dependent endpoint functionality into a single library called 'iomgr'. This is primarily to prepare for a Windows port - where posix socket semantics lead to poor quality code. Mostly this is a code movement CL, with some small changes to help prepare the way for porting: - em style fd objects can only be held internally in iomgr, and own their memory - added grpc_iomgr_create_endpoint_pair() to accomodate the common pattern of creating a tcp endpoint from the output of socketpair - this will help keep our tests portable - separated em alarm interface into a separate file, as this part of event manager is needed higher up the stack - made the eventmanager bits a true singleton, simplifying API's across the stack as there's no longer a reason to carry a pointer there. Initial design document is here: https://docs.google.com/document/d/1VmafcHvvrP5kwtQkz84R5yXF7u7fW-9Pn0bkSUQHDt8/edit?disco=AAAAARNByxg Change on 2014/12/09 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81716456
10 years ago
$(E) "[RUN] Testing tcp_server_posix_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/tcp_server_posix_test || ( echo test tcp_server_posix_test failed ; exit 1 )
Introducing iomgr. Move eventmanager and platform dependent endpoint functionality into a single library called 'iomgr'. This is primarily to prepare for a Windows port - where posix socket semantics lead to poor quality code. Mostly this is a code movement CL, with some small changes to help prepare the way for porting: - em style fd objects can only be held internally in iomgr, and own their memory - added grpc_iomgr_create_endpoint_pair() to accomodate the common pattern of creating a tcp endpoint from the output of socketpair - this will help keep our tests portable - separated em alarm interface into a separate file, as this part of event manager is needed higher up the stack - made the eventmanager bits a true singleton, simplifying API's across the stack as there's no longer a reason to carry a pointer there. Initial design document is here: https://docs.google.com/document/d/1VmafcHvvrP5kwtQkz84R5yXF7u7fW-9Pn0bkSUQHDt8/edit?disco=AAAAARNByxg Change on 2014/12/09 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81716456
10 years ago
$(E) "[RUN] Testing tcp_client_posix_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/tcp_client_posix_test || ( echo test tcp_client_posix_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing grpc_channel_stack_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/grpc_channel_stack_test || ( echo test grpc_channel_stack_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing metadata_buffer_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/metadata_buffer_test || ( echo test metadata_buffer_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing grpc_completion_queue_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/grpc_completion_queue_test || ( echo test grpc_completion_queue_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing census_window_stats_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/census_window_stats_test || ( echo test census_window_stats_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing census_statistics_quick_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/census_statistics_quick_test || ( echo test census_statistics_quick_test failed ; exit 1 )
Fix a bug in base-log, and add test. The tests compute usable space in the log, but do so using subtraction on unsigned values and did not correctly account for number of blocks used per-core. This could lead underflow, and an incorrect space calculation. In addition, testing showed that with current defaults, tests would fail on a system with > ~64 cores. This can be alleviated through use of a larger log. An explicit check in the log initialization has been added for this case, and the log size increased if necessary. Change on 2014/12/15 by aveitch <aveitch@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82154432
10 years ago
$(E) "[RUN] Testing census_statistics_small_log_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/census_statistics_small_log_test || ( echo test census_statistics_small_log_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing census_statistics_performance_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/census_statistics_performance_test || ( echo test census_statistics_performance_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing census_statistics_multiple_writers_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/census_statistics_multiple_writers_test || ( echo test census_statistics_multiple_writers_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing census_statistics_multiple_writers_circular_buffer_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/census_statistics_multiple_writers_circular_buffer_test || ( echo test census_statistics_multiple_writers_circular_buffer_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing census_stub_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/census_stub_test || ( echo test census_stub_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing census_hash_table_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/census_hash_table_test || ( echo test census_hash_table_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing fling_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/fling_test || ( echo test fling_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing echo_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/echo_test || ( echo test echo_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing message_compress_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/message_compress_test || ( echo test message_compress_test failed ; exit 1 )
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[RUN] Testing bin_encoder_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/bin_encoder_test || ( echo test bin_encoder_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing secure_endpoint_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/secure_endpoint_test || ( echo test secure_endpoint_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing httpcli_format_request_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/httpcli_format_request_test || ( echo test httpcli_format_request_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing httpcli_parser_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/httpcli_parser_test || ( echo test httpcli_parser_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing httpcli_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/httpcli_test || ( echo test httpcli_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing grpc_credentials_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/grpc_credentials_test || ( echo test grpc_credentials_test failed ; exit 1 )
Adding base utils for JWT service account workflow. OpenSSL base64 decoding is a disaster and does not support url_safe which we need for the JWT. Change on 2014/12/12 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020601
10 years ago
$(E) "[RUN] Testing grpc_base64_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/grpc_base64_test || ( echo test grpc_base64_test failed ; exit 1 )
Adding base utils for JWT service account workflow. OpenSSL base64 decoding is a disaster and does not support url_safe which we need for the JWT. Change on 2014/12/12 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020601
10 years ago
$(E) "[RUN] Testing grpc_json_token_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/grpc_json_token_test || ( echo test grpc_json_token_test failed ; exit 1 )
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
$(E) "[RUN] Testing timeout_encoding_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/timeout_encoding_test || ( echo test timeout_encoding_test failed ; exit 1 )
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
$(E) "[RUN] Testing fd_posix_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/fd_posix_test || ( echo test fd_posix_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing fling_stream_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/fling_stream_test || ( echo test fling_stream_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing lame_client_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/lame_client_test || ( echo test lame_client_test failed ; exit 1 )
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
$(E) "[RUN] Testing alarm_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/alarm_test || ( echo test alarm_test failed ; exit 1 )
Port [] alarm management to GRPC. This change implements a platform independent alarm manager in alarm.c. It's integrated with iomgr, and some tests are cleaned up. The alarm implementation itself is a fairly direct port of LazyAlarmList from eventmanager. SpinLock has been replaced for now with gpr_mu, and other atomic operations have been dropped (again, for now). A majority of tests have been ported. Change on 2014/12/19 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82551363
10 years ago
$(E) "[RUN] Testing alarm_list_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/alarm_list_test || ( echo test alarm_list_test failed ; exit 1 )
Port [] alarm management to GRPC. This change implements a platform independent alarm manager in alarm.c. It's integrated with iomgr, and some tests are cleaned up. The alarm implementation itself is a fairly direct port of LazyAlarmList from eventmanager. SpinLock has been replaced for now with gpr_mu, and other atomic operations have been dropped (again, for now). A majority of tests have been ported. Change on 2014/12/19 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82551363
10 years ago
$(E) "[RUN] Testing alarm_heap_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/alarm_heap_test || ( echo test alarm_heap_test failed ; exit 1 )
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
$(E) "[RUN] Testing time_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/time_test || ( echo test time_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_cancel_after_accept_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_cancel_after_accept_test || ( echo test chttp2_fake_security_cancel_after_accept_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_cancel_after_accept_and_writes_closed_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_cancel_after_accept_and_writes_closed_test || ( echo test chttp2_fake_security_cancel_after_accept_and_writes_closed_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_cancel_after_invoke_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_cancel_after_invoke_test || ( echo test chttp2_fake_security_cancel_after_invoke_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_cancel_before_invoke_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_cancel_before_invoke_test || ( echo test chttp2_fake_security_cancel_before_invoke_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_cancel_in_a_vacuum_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_cancel_in_a_vacuum_test || ( echo test chttp2_fake_security_cancel_in_a_vacuum_test failed ; exit 1 )
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_census_simple_request_test"
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_census_simple_request_test || ( echo test chttp2_fake_security_census_simple_request_test failed ; exit 1 )
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_disappearing_server_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_disappearing_server_test || ( echo test chttp2_fake_security_disappearing_server_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_early_server_shutdown_finishes_inflight_calls_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_early_server_shutdown_finishes_inflight_calls_test || ( echo test chttp2_fake_security_early_server_shutdown_finishes_inflight_calls_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_early_server_shutdown_finishes_tags_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_early_server_shutdown_finishes_tags_test || ( echo test chttp2_fake_security_early_server_shutdown_finishes_tags_test failed ; exit 1 )
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_graceful_server_shutdown_test"
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_graceful_server_shutdown_test || ( echo test chttp2_fake_security_graceful_server_shutdown_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_invoke_large_request_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_invoke_large_request_test || ( echo test chttp2_fake_security_invoke_large_request_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_max_concurrent_streams_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_max_concurrent_streams_test || ( echo test chttp2_fake_security_max_concurrent_streams_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_no_op_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_no_op_test || ( echo test chttp2_fake_security_no_op_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_ping_pong_streaming_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_ping_pong_streaming_test || ( echo test chttp2_fake_security_ping_pong_streaming_test failed ; exit 1 )
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_request_response_with_binary_metadata_and_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_request_response_with_binary_metadata_and_payload_test || ( echo test chttp2_fake_security_request_response_with_binary_metadata_and_payload_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_request_response_with_metadata_and_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_request_response_with_metadata_and_payload_test || ( echo test chttp2_fake_security_request_response_with_metadata_and_payload_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_request_response_with_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_request_response_with_payload_test || ( echo test chttp2_fake_security_request_response_with_payload_test failed ; exit 1 )
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_request_response_with_trailing_metadata_and_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_request_response_with_trailing_metadata_and_payload_test || ( echo test chttp2_fake_security_request_response_with_trailing_metadata_and_payload_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_simple_delayed_request_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_simple_delayed_request_test || ( echo test chttp2_fake_security_simple_delayed_request_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_simple_request_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_simple_request_test || ( echo test chttp2_fake_security_simple_request_test failed ; exit 1 )
Strip "_test" from the name of thread_stress_test.c. Having a BUILD target named chttp2_fullstack_thread_stress_test_test is deceptive as it is a direct test of the system rather than a test of a test of the system. Change on 2014/12/09 by nathaniel <nathaniel@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81688649
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_thread_stress_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_thread_stress_test || ( echo test chttp2_fake_security_thread_stress_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fake_security_writes_done_hangs_with_pending_read_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fake_security_writes_done_hangs_with_pending_read_test || ( echo test chttp2_fake_security_writes_done_hangs_with_pending_read_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_cancel_after_accept_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_cancel_after_accept_test || ( echo test chttp2_fullstack_cancel_after_accept_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_cancel_after_accept_and_writes_closed_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_cancel_after_accept_and_writes_closed_test || ( echo test chttp2_fullstack_cancel_after_accept_and_writes_closed_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_cancel_after_invoke_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_cancel_after_invoke_test || ( echo test chttp2_fullstack_cancel_after_invoke_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_cancel_before_invoke_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_cancel_before_invoke_test || ( echo test chttp2_fullstack_cancel_before_invoke_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_cancel_in_a_vacuum_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_cancel_in_a_vacuum_test || ( echo test chttp2_fullstack_cancel_in_a_vacuum_test failed ; exit 1 )
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_census_simple_request_test"
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_census_simple_request_test || ( echo test chttp2_fullstack_census_simple_request_test failed ; exit 1 )
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_disappearing_server_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_disappearing_server_test || ( echo test chttp2_fullstack_disappearing_server_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_early_server_shutdown_finishes_inflight_calls_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_early_server_shutdown_finishes_inflight_calls_test || ( echo test chttp2_fullstack_early_server_shutdown_finishes_inflight_calls_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_early_server_shutdown_finishes_tags_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_early_server_shutdown_finishes_tags_test || ( echo test chttp2_fullstack_early_server_shutdown_finishes_tags_test failed ; exit 1 )
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_graceful_server_shutdown_test"
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_graceful_server_shutdown_test || ( echo test chttp2_fullstack_graceful_server_shutdown_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_invoke_large_request_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_invoke_large_request_test || ( echo test chttp2_fullstack_invoke_large_request_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_max_concurrent_streams_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_max_concurrent_streams_test || ( echo test chttp2_fullstack_max_concurrent_streams_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_no_op_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_no_op_test || ( echo test chttp2_fullstack_no_op_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_ping_pong_streaming_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_ping_pong_streaming_test || ( echo test chttp2_fullstack_ping_pong_streaming_test failed ; exit 1 )
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_request_response_with_binary_metadata_and_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_request_response_with_binary_metadata_and_payload_test || ( echo test chttp2_fullstack_request_response_with_binary_metadata_and_payload_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_request_response_with_metadata_and_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_request_response_with_metadata_and_payload_test || ( echo test chttp2_fullstack_request_response_with_metadata_and_payload_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_request_response_with_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_request_response_with_payload_test || ( echo test chttp2_fullstack_request_response_with_payload_test failed ; exit 1 )
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_request_response_with_trailing_metadata_and_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_request_response_with_trailing_metadata_and_payload_test || ( echo test chttp2_fullstack_request_response_with_trailing_metadata_and_payload_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_simple_delayed_request_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_simple_delayed_request_test || ( echo test chttp2_fullstack_simple_delayed_request_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_simple_request_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_simple_request_test || ( echo test chttp2_fullstack_simple_request_test failed ; exit 1 )
Strip "_test" from the name of thread_stress_test.c. Having a BUILD target named chttp2_fullstack_thread_stress_test_test is deceptive as it is a direct test of the system rather than a test of a test of the system. Change on 2014/12/09 by nathaniel <nathaniel@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81688649
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_thread_stress_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_thread_stress_test || ( echo test chttp2_fullstack_thread_stress_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_fullstack_writes_done_hangs_with_pending_read_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_fullstack_writes_done_hangs_with_pending_read_test || ( echo test chttp2_fullstack_writes_done_hangs_with_pending_read_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_cancel_after_accept_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_after_accept_test || ( echo test chttp2_simple_ssl_fullstack_cancel_after_accept_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_cancel_after_accept_and_writes_closed_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_after_accept_and_writes_closed_test || ( echo test chttp2_simple_ssl_fullstack_cancel_after_accept_and_writes_closed_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_cancel_after_invoke_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_after_invoke_test || ( echo test chttp2_simple_ssl_fullstack_cancel_after_invoke_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_cancel_before_invoke_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_before_invoke_test || ( echo test chttp2_simple_ssl_fullstack_cancel_before_invoke_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_cancel_in_a_vacuum_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_in_a_vacuum_test || ( echo test chttp2_simple_ssl_fullstack_cancel_in_a_vacuum_test failed ; exit 1 )
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_census_simple_request_test"
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_census_simple_request_test || ( echo test chttp2_simple_ssl_fullstack_census_simple_request_test failed ; exit 1 )
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_disappearing_server_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_disappearing_server_test || ( echo test chttp2_simple_ssl_fullstack_disappearing_server_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_early_server_shutdown_finishes_inflight_calls_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_early_server_shutdown_finishes_inflight_calls_test || ( echo test chttp2_simple_ssl_fullstack_early_server_shutdown_finishes_inflight_calls_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_early_server_shutdown_finishes_tags_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_early_server_shutdown_finishes_tags_test || ( echo test chttp2_simple_ssl_fullstack_early_server_shutdown_finishes_tags_test failed ; exit 1 )
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_graceful_server_shutdown_test"
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_graceful_server_shutdown_test || ( echo test chttp2_simple_ssl_fullstack_graceful_server_shutdown_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_invoke_large_request_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_invoke_large_request_test || ( echo test chttp2_simple_ssl_fullstack_invoke_large_request_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_max_concurrent_streams_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_max_concurrent_streams_test || ( echo test chttp2_simple_ssl_fullstack_max_concurrent_streams_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_no_op_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_no_op_test || ( echo test chttp2_simple_ssl_fullstack_no_op_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_ping_pong_streaming_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_ping_pong_streaming_test || ( echo test chttp2_simple_ssl_fullstack_ping_pong_streaming_test failed ; exit 1 )
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_request_response_with_binary_metadata_and_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_request_response_with_binary_metadata_and_payload_test || ( echo test chttp2_simple_ssl_fullstack_request_response_with_binary_metadata_and_payload_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_request_response_with_metadata_and_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_request_response_with_metadata_and_payload_test || ( echo test chttp2_simple_ssl_fullstack_request_response_with_metadata_and_payload_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_request_response_with_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_request_response_with_payload_test || ( echo test chttp2_simple_ssl_fullstack_request_response_with_payload_test failed ; exit 1 )
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_request_response_with_trailing_metadata_and_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_request_response_with_trailing_metadata_and_payload_test || ( echo test chttp2_simple_ssl_fullstack_request_response_with_trailing_metadata_and_payload_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_simple_delayed_request_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_simple_delayed_request_test || ( echo test chttp2_simple_ssl_fullstack_simple_delayed_request_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_simple_request_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_simple_request_test || ( echo test chttp2_simple_ssl_fullstack_simple_request_test failed ; exit 1 )
Strip "_test" from the name of thread_stress_test.c. Having a BUILD target named chttp2_fullstack_thread_stress_test_test is deceptive as it is a direct test of the system rather than a test of a test of the system. Change on 2014/12/09 by nathaniel <nathaniel@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81688649
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_thread_stress_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_thread_stress_test || ( echo test chttp2_simple_ssl_fullstack_thread_stress_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_fullstack_writes_done_hangs_with_pending_read_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_fullstack_writes_done_hangs_with_pending_read_test || ( echo test chttp2_simple_ssl_fullstack_writes_done_hangs_with_pending_read_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_accept_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_accept_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_accept_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_accept_and_writes_closed_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_accept_and_writes_closed_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_accept_and_writes_closed_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_invoke_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_invoke_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_invoke_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_cancel_before_invoke_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_before_invoke_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_cancel_before_invoke_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_cancel_in_a_vacuum_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_in_a_vacuum_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_cancel_in_a_vacuum_test failed ; exit 1 )
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_census_simple_request_test"
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_census_simple_request_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_census_simple_request_test failed ; exit 1 )
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_disappearing_server_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_disappearing_server_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_disappearing_server_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_early_server_shutdown_finishes_inflight_calls_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_early_server_shutdown_finishes_inflight_calls_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_early_server_shutdown_finishes_inflight_calls_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_early_server_shutdown_finishes_tags_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_early_server_shutdown_finishes_tags_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_early_server_shutdown_finishes_tags_test failed ; exit 1 )
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_graceful_server_shutdown_test"
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_graceful_server_shutdown_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_graceful_server_shutdown_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_invoke_large_request_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_invoke_large_request_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_invoke_large_request_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_max_concurrent_streams_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_max_concurrent_streams_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_max_concurrent_streams_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_no_op_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_no_op_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_no_op_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_ping_pong_streaming_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_ping_pong_streaming_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_ping_pong_streaming_test failed ; exit 1 )
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_binary_metadata_and_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_binary_metadata_and_payload_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_binary_metadata_and_payload_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_metadata_and_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_metadata_and_payload_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_metadata_and_payload_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_payload_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_payload_test failed ; exit 1 )
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_trailing_metadata_and_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_trailing_metadata_and_payload_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_trailing_metadata_and_payload_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_simple_delayed_request_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_simple_delayed_request_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_simple_delayed_request_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_simple_request_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_simple_request_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_simple_request_test failed ; exit 1 )
Strip "_test" from the name of thread_stress_test.c. Having a BUILD target named chttp2_fullstack_thread_stress_test_test is deceptive as it is a direct test of the system rather than a test of a test of the system. Change on 2014/12/09 by nathaniel <nathaniel@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81688649
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_thread_stress_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_thread_stress_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_thread_stress_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_simple_ssl_with_oauth2_fullstack_writes_done_hangs_with_pending_read_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_writes_done_hangs_with_pending_read_test || ( echo test chttp2_simple_ssl_with_oauth2_fullstack_writes_done_hangs_with_pending_read_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_cancel_after_accept_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_cancel_after_accept_test || ( echo test chttp2_socket_pair_cancel_after_accept_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_cancel_after_accept_and_writes_closed_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_cancel_after_accept_and_writes_closed_test || ( echo test chttp2_socket_pair_cancel_after_accept_and_writes_closed_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_cancel_after_invoke_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_cancel_after_invoke_test || ( echo test chttp2_socket_pair_cancel_after_invoke_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_cancel_before_invoke_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_cancel_before_invoke_test || ( echo test chttp2_socket_pair_cancel_before_invoke_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_cancel_in_a_vacuum_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_cancel_in_a_vacuum_test || ( echo test chttp2_socket_pair_cancel_in_a_vacuum_test failed ; exit 1 )
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_census_simple_request_test"
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_census_simple_request_test || ( echo test chttp2_socket_pair_census_simple_request_test failed ; exit 1 )
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_disappearing_server_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_disappearing_server_test || ( echo test chttp2_socket_pair_disappearing_server_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_early_server_shutdown_finishes_inflight_calls_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_early_server_shutdown_finishes_inflight_calls_test || ( echo test chttp2_socket_pair_early_server_shutdown_finishes_inflight_calls_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_early_server_shutdown_finishes_tags_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_early_server_shutdown_finishes_tags_test || ( echo test chttp2_socket_pair_early_server_shutdown_finishes_tags_test failed ; exit 1 )
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_graceful_server_shutdown_test"
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_graceful_server_shutdown_test || ( echo test chttp2_socket_pair_graceful_server_shutdown_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_invoke_large_request_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_invoke_large_request_test || ( echo test chttp2_socket_pair_invoke_large_request_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_max_concurrent_streams_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_max_concurrent_streams_test || ( echo test chttp2_socket_pair_max_concurrent_streams_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_no_op_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_no_op_test || ( echo test chttp2_socket_pair_no_op_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_ping_pong_streaming_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_ping_pong_streaming_test || ( echo test chttp2_socket_pair_ping_pong_streaming_test failed ; exit 1 )
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_request_response_with_binary_metadata_and_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_request_response_with_binary_metadata_and_payload_test || ( echo test chttp2_socket_pair_request_response_with_binary_metadata_and_payload_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_request_response_with_metadata_and_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_request_response_with_metadata_and_payload_test || ( echo test chttp2_socket_pair_request_response_with_metadata_and_payload_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_request_response_with_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_request_response_with_payload_test || ( echo test chttp2_socket_pair_request_response_with_payload_test failed ; exit 1 )
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_request_response_with_trailing_metadata_and_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_request_response_with_trailing_metadata_and_payload_test || ( echo test chttp2_socket_pair_request_response_with_trailing_metadata_and_payload_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_simple_delayed_request_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_simple_delayed_request_test || ( echo test chttp2_socket_pair_simple_delayed_request_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_simple_request_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_simple_request_test || ( echo test chttp2_socket_pair_simple_request_test failed ; exit 1 )
Strip "_test" from the name of thread_stress_test.c. Having a BUILD target named chttp2_fullstack_thread_stress_test_test is deceptive as it is a direct test of the system rather than a test of a test of the system. Change on 2014/12/09 by nathaniel <nathaniel@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81688649
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_thread_stress_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_thread_stress_test || ( echo test chttp2_socket_pair_thread_stress_test failed ; exit 1 )
Initial import.
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_writes_done_hangs_with_pending_read_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_writes_done_hangs_with_pending_read_test || ( echo test chttp2_socket_pair_writes_done_hangs_with_pending_read_test failed ; exit 1 )
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_cancel_after_accept_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_after_accept_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_cancel_after_accept_test failed ; exit 1 )
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_cancel_after_accept_and_writes_closed_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_after_accept_and_writes_closed_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_cancel_after_accept_and_writes_closed_test failed ; exit 1 )
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_cancel_after_invoke_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_after_invoke_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_cancel_after_invoke_test failed ; exit 1 )
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_cancel_before_invoke_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_before_invoke_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_cancel_before_invoke_test failed ; exit 1 )
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_cancel_in_a_vacuum_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_in_a_vacuum_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_cancel_in_a_vacuum_test failed ; exit 1 )
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_census_simple_request_test"
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_census_simple_request_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_census_simple_request_test failed ; exit 1 )
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_disappearing_server_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_disappearing_server_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_disappearing_server_test failed ; exit 1 )
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_early_server_shutdown_finishes_inflight_calls_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_early_server_shutdown_finishes_inflight_calls_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_early_server_shutdown_finishes_inflight_calls_test failed ; exit 1 )
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_early_server_shutdown_finishes_tags_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_early_server_shutdown_finishes_tags_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_early_server_shutdown_finishes_tags_test failed ; exit 1 )
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_graceful_server_shutdown_test"
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_graceful_server_shutdown_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_graceful_server_shutdown_test failed ; exit 1 )
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_invoke_large_request_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_invoke_large_request_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_invoke_large_request_test failed ; exit 1 )
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_max_concurrent_streams_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_max_concurrent_streams_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_max_concurrent_streams_test failed ; exit 1 )
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_no_op_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_no_op_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_no_op_test failed ; exit 1 )
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_ping_pong_streaming_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_ping_pong_streaming_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_ping_pong_streaming_test failed ; exit 1 )
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_request_response_with_binary_metadata_and_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_request_response_with_binary_metadata_and_payload_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_request_response_with_binary_metadata_and_payload_test failed ; exit 1 )
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_request_response_with_metadata_and_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_request_response_with_metadata_and_payload_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_request_response_with_metadata_and_payload_test failed ; exit 1 )
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_request_response_with_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_request_response_with_payload_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_request_response_with_payload_test failed ; exit 1 )
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_request_response_with_trailing_metadata_and_payload_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_request_response_with_trailing_metadata_and_payload_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_request_response_with_trailing_metadata_and_payload_test failed ; exit 1 )
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_simple_delayed_request_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_simple_delayed_request_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_simple_delayed_request_test failed ; exit 1 )
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_simple_request_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_simple_request_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_simple_request_test failed ; exit 1 )
Strip "_test" from the name of thread_stress_test.c. Having a BUILD target named chttp2_fullstack_thread_stress_test_test is deceptive as it is a direct test of the system rather than a test of a test of the system. Change on 2014/12/09 by nathaniel <nathaniel@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81688649
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_thread_stress_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_thread_stress_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_thread_stress_test failed ; exit 1 )
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[RUN] Testing chttp2_socket_pair_one_byte_at_a_time_writes_done_hangs_with_pending_read_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_writes_done_hangs_with_pending_read_test || ( echo test chttp2_socket_pair_one_byte_at_a_time_writes_done_hangs_with_pending_read_test failed ; exit 1 )
Initial import.
10 years ago
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
test_cxx: buildtests_cxx
Better C/C++ split in the Makefile targets, to better prepare slicing it into two different Makefiles. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81076626
10 years ago
$(E) "[RUN] Testing thread_pool_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/thread_pool_test || ( echo test thread_pool_test failed ; exit 1 )
Better C/C++ split in the Makefile targets, to better prepare slicing it into two different Makefiles. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81076626
10 years ago
$(E) "[RUN] Testing status_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/status_test || ( echo test status_test failed ; exit 1 )
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
$(E) "[RUN] Testing sync_client_async_server_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/sync_client_async_server_test || ( echo test sync_client_async_server_test failed ; exit 1 )
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
$(E) "[RUN] Testing qps_client"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/qps_client || ( echo test qps_client failed ; exit 1 )
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
$(E) "[RUN] Testing qps_server"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/qps_server || ( echo test qps_server failed ; exit 1 )
TIPS client implementation.
10 years ago
$(E) "[RUN] Testing tips_client_test"
$(Q) ./bins/$(CONFIG)/tips_client_test || ( echo test tips_client_test failed ; exit 1 )
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
$(E) "[RUN] Testing end2end_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/end2end_test || ( echo test end2end_test failed ; exit 1 )
Client side implementation of creating channel with credentials. The old test_ssl_channel related code is deleted and the new create channel call is used for interop tests. Change on 2014/12/19 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82540921
10 years ago
$(E) "[RUN] Testing channel_arguments_test"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ./bins/$(CONFIG)/channel_arguments_test || ( echo test channel_arguments_test failed ; exit 1 )
Add ServiceAccount Credentials wrapping and handle credentials creation failure. Change on 2015/01/09 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83634736
10 years ago
$(E) "[RUN] Testing credentials_test"
$(Q) ./bins/$(CONFIG)/credentials_test || ( echo test credentials_test failed ; exit 1 )
Better C/C++ split in the Makefile targets, to better prepare slicing it into two different Makefiles. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81076626
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
tools: privatelibs bins/$(CONFIG)/gen_hpack_tables bins/$(CONFIG)/grpc_fetch_oauth2
Initial import.
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
buildbenchmarks: privatelibs bins/$(CONFIG)/grpc_completion_queue_benchmark bins/$(CONFIG)/low_level_ping_pong_benchmark
Initial import.
10 years ago
benchmarks: buildbenchmarks
strip: strip-static strip-shared
The built-in ssl target now properly merges the archives instead of nesting them. Change on 2014/12/16 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82273781
10 years ago
strip-static: strip-static_c strip-static_cxx
strip-shared: strip-shared_c strip-shared_cxx
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# TODO(nnoble): the strip target is stripping in-place, instead
# of copying files in a temporary folder.
# This prevents proper debugging after running make install.
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
strip-static_c: static_c
Initial import.
10 years ago
$(E) "[STRIP] Stripping libgpr.a"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(STRIP) libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[STRIP] Stripping libgrpc.a"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(STRIP) libs/$(CONFIG)/libgrpc.a
Initial import.
10 years ago
$(E) "[STRIP] Stripping libgrpc_unsecure.a"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(STRIP) libs/$(CONFIG)/libgrpc_unsecure.a
Initial import.
10 years ago
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
strip-static_cxx: static_cxx
$(E) "[STRIP] Stripping libgrpc++.a"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(STRIP) libs/$(CONFIG)/libgrpc++.a
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
strip-shared_c: shared_c
Initial import.
10 years ago
$(E) "[STRIP] Stripping libgpr.so"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(STRIP) libs/$(CONFIG)/libgpr.$(SHARED_EXT)
Initial import.
10 years ago
$(E) "[STRIP] Stripping libgrpc.so"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(STRIP) libs/$(CONFIG)/libgrpc.$(SHARED_EXT)
Initial import.
10 years ago
$(E) "[STRIP] Stripping libgrpc_unsecure.so"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(STRIP) libs/$(CONFIG)/libgrpc_unsecure.$(SHARED_EXT)
Initial import.
10 years ago
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
strip-shared_cxx: shared_cxx
$(E) "[STRIP] Stripping libgrpc++.so"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(STRIP) libs/$(CONFIG)/libgrpc++.$(SHARED_EXT)
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
TIPS client implementation.
10 years ago
gens/examples/tips/empty.pb.cc: examples/tips/empty.proto $(PROTOC_PLUGINS)
$(E) "[PROTOC] Generating protobuf CC file from $<"
$(Q) mkdir -p `dirname $@`
$(Q) $(PROTOC) --cpp_out=gens --grpc_out=gens --plugin=protoc-gen-grpc=bins/$(CONFIG)/cpp_plugin $<
gens/examples/tips/label.pb.cc: examples/tips/label.proto $(PROTOC_PLUGINS)
$(E) "[PROTOC] Generating protobuf CC file from $<"
$(Q) mkdir -p `dirname $@`
$(Q) $(PROTOC) --cpp_out=gens --grpc_out=gens --plugin=protoc-gen-grpc=bins/$(CONFIG)/cpp_plugin $<
gens/examples/tips/pubsub.pb.cc: examples/tips/pubsub.proto $(PROTOC_PLUGINS)
$(E) "[PROTOC] Generating protobuf CC file from $<"
$(Q) mkdir -p `dirname $@`
$(Q) $(PROTOC) --cpp_out=gens --grpc_out=gens --plugin=protoc-gen-grpc=bins/$(CONFIG)/cpp_plugin $<
Remove protoc_plugins, and use a variable. Files that have dependencies on phony targets are always remade, leading to much unnecessary compiling of the C++ tests.
10 years ago
gens/test/cpp/interop/empty.pb.cc: test/cpp/interop/empty.proto $(PROTOC_PLUGINS)
Initial import.
10 years ago
$(E) "[PROTOC] Generating protobuf CC file from $<"
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(PROTOC) --cpp_out=gens --grpc_out=gens --plugin=protoc-gen-grpc=bins/$(CONFIG)/cpp_plugin $<
Initial import.
10 years ago
Remove protoc_plugins, and use a variable. Files that have dependencies on phony targets are always remade, leading to much unnecessary compiling of the C++ tests.
10 years ago
gens/test/cpp/interop/messages.pb.cc: test/cpp/interop/messages.proto $(PROTOC_PLUGINS)
Getting the generated Makefile to support .proto files. -) Compiling the proto plugin using the HOSTCC compiler. -) Set up proper rules to invoke the plugin from protoc. -) Few various renaming hacks to fully get out of []. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82000361
10 years ago
$(E) "[PROTOC] Generating protobuf CC file from $<"
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(PROTOC) --cpp_out=gens --grpc_out=gens --plugin=protoc-gen-grpc=bins/$(CONFIG)/cpp_plugin $<
Getting the generated Makefile to support .proto files. -) Compiling the proto plugin using the HOSTCC compiler. -) Set up proper rules to invoke the plugin from protoc. -) Few various renaming hacks to fully get out of []. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82000361
10 years ago
Remove protoc_plugins, and use a variable. Files that have dependencies on phony targets are always remade, leading to much unnecessary compiling of the C++ tests.
10 years ago
gens/test/cpp/interop/test.pb.cc: test/cpp/interop/test.proto $(PROTOC_PLUGINS)
Getting the generated Makefile to support .proto files. -) Compiling the proto plugin using the HOSTCC compiler. -) Set up proper rules to invoke the plugin from protoc. -) Few various renaming hacks to fully get out of []. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82000361
10 years ago
$(E) "[PROTOC] Generating protobuf CC file from $<"
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(PROTOC) --cpp_out=gens --grpc_out=gens --plugin=protoc-gen-grpc=bins/$(CONFIG)/cpp_plugin $<
Getting the generated Makefile to support .proto files. -) Compiling the proto plugin using the HOSTCC compiler. -) Set up proper rules to invoke the plugin from protoc. -) Few various renaming hacks to fully get out of []. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82000361
10 years ago
Remove protoc_plugins, and use a variable. Files that have dependencies on phony targets are always remade, leading to much unnecessary compiling of the C++ tests.
10 years ago
gens/test/cpp/qps/qpstest.pb.cc: test/cpp/qps/qpstest.proto $(PROTOC_PLUGINS)
Fix qps test. This was failing to compile.
10 years ago
$(E) "[PROTOC] Generating protobuf CC file from $<"
$(Q) mkdir -p `dirname $@`
$(Q) $(PROTOC) --cpp_out=gens --grpc_out=gens --plugin=protoc-gen-grpc=bins/$(CONFIG)/cpp_plugin $<
Remove protoc_plugins, and use a variable. Files that have dependencies on phony targets are always remade, leading to much unnecessary compiling of the C++ tests.
10 years ago
gens/test/cpp/util/echo.pb.cc: test/cpp/util/echo.proto $(PROTOC_PLUGINS)
Getting the generated Makefile to support .proto files. -) Compiling the proto plugin using the HOSTCC compiler. -) Set up proper rules to invoke the plugin from protoc. -) Few various renaming hacks to fully get out of []. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82000361
10 years ago
$(E) "[PROTOC] Generating protobuf CC file from $<"
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(PROTOC) --cpp_out=gens --grpc_out=gens --plugin=protoc-gen-grpc=bins/$(CONFIG)/cpp_plugin $<
Getting the generated Makefile to support .proto files. -) Compiling the proto plugin using the HOSTCC compiler. -) Set up proper rules to invoke the plugin from protoc. -) Few various renaming hacks to fully get out of []. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82000361
10 years ago
Remove protoc_plugins, and use a variable. Files that have dependencies on phony targets are always remade, leading to much unnecessary compiling of the C++ tests.
10 years ago
gens/test/cpp/util/echo_duplicate.pb.cc: test/cpp/util/echo_duplicate.proto $(PROTOC_PLUGINS)
Add a test to export two TestService on the same server, only with different package names. Make sure they work as intended. This is based on [] Change on 2015/01/08 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83558743
10 years ago
$(E) "[PROTOC] Generating protobuf CC file from $<"
$(Q) mkdir -p `dirname $@`
$(Q) $(PROTOC) --cpp_out=gens --grpc_out=gens --plugin=protoc-gen-grpc=bins/$(CONFIG)/cpp_plugin $<
Remove protoc_plugins, and use a variable. Files that have dependencies on phony targets are always remade, leading to much unnecessary compiling of the C++ tests.
10 years ago
gens/test/cpp/util/messages.pb.cc: test/cpp/util/messages.proto $(PROTOC_PLUGINS)
Add a test to export two TestService on the same server, only with different package names. Make sure they work as intended. This is based on [] Change on 2015/01/08 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83558743
10 years ago
$(E) "[PROTOC] Generating protobuf CC file from $<"
$(Q) mkdir -p `dirname $@`
$(Q) $(PROTOC) --cpp_out=gens --grpc_out=gens --plugin=protoc-gen-grpc=bins/$(CONFIG)/cpp_plugin $<
Initial import.
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
objs/$(CONFIG)/%.o : %.c
Initial import.
10 years ago
$(E) "[C] Compiling $<"
$(Q) mkdir -p `dirname $@`
Generate dependencies when compiling the .o If we don't have a .o, we don't need dependencies, so skip the setup step of building them. Instead, when building the .o, we generate the .dep and output it (to objs/ to make the rules a little easier). The next run of make will include the dep file and any dependencies will be picked up. This change also disables the disabling of dependency checking if we have zlib or openssl being compiled from third_party. Additionally it inverts the logic for including dependencies from ONLY if we are doing a clean to ONLY if we are NOT doing a clean.
10 years ago
$(Q) $(CC) $(CFLAGS) $(CPPFLAGS) -MMD -MF $(addsuffix .dep, $(basename $@)) -c -o $@ $<
Initial import.
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
objs/$(CONFIG)/%.o : gens/%.pb.cc
Initial import.
10 years ago
$(E) "[CXX] Compiling $<"
$(Q) mkdir -p `dirname $@`
Generate dependencies when compiling the .o If we don't have a .o, we don't need dependencies, so skip the setup step of building them. Instead, when building the .o, we generate the .dep and output it (to objs/ to make the rules a little easier). The next run of make will include the dep file and any dependencies will be picked up. This change also disables the disabling of dependency checking if we have zlib or openssl being compiled from third_party. Additionally it inverts the logic for including dependencies from ONLY if we are doing a clean to ONLY if we are NOT doing a clean.
10 years ago
$(Q) $(CXX) $(CXXFLAGS) $(CPPFLAGS) -MMD -MF $(addsuffix .dep, $(basename $@)) -c -o $@ $<
Initial import.
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
objs/$(CONFIG)/src/compiler/%.o : src/compiler/%.cc
Getting the generated Makefile to support .proto files. -) Compiling the proto plugin using the HOSTCC compiler. -) Set up proper rules to invoke the plugin from protoc. -) Few various renaming hacks to fully get out of []. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82000361
10 years ago
$(E) "[HOSTCXX] Compiling $<"
$(Q) mkdir -p `dirname $@`
Generate dependencies when compiling the .o If we don't have a .o, we don't need dependencies, so skip the setup step of building them. Instead, when building the .o, we generate the .dep and output it (to objs/ to make the rules a little easier). The next run of make will include the dep file and any dependencies will be picked up. This change also disables the disabling of dependency checking if we have zlib or openssl being compiled from third_party. Additionally it inverts the logic for including dependencies from ONLY if we are doing a clean to ONLY if we are NOT doing a clean.
10 years ago
$(Q) $(HOST_CXX) $(HOST_CXXFLAGS) $(HOST_CPPFLAGS) -MMD -MF $(addsuffix .dep, $(basename $@)) -c -o $@ $<
Getting the generated Makefile to support .proto files. -) Compiling the proto plugin using the HOSTCC compiler. -) Set up proper rules to invoke the plugin from protoc. -) Few various renaming hacks to fully get out of []. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82000361
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
objs/$(CONFIG)/%.o : %.cc
Initial import.
10 years ago
$(E) "[CXX] Compiling $<"
$(Q) mkdir -p `dirname $@`
Generate dependencies when compiling the .o If we don't have a .o, we don't need dependencies, so skip the setup step of building them. Instead, when building the .o, we generate the .dep and output it (to objs/ to make the rules a little easier). The next run of make will include the dep file and any dependencies will be picked up. This change also disables the disabling of dependency checking if we have zlib or openssl being compiled from third_party. Additionally it inverts the logic for including dependencies from ONLY if we are doing a clean to ONLY if we are NOT doing a clean.
10 years ago
$(Q) $(CXX) $(CXXFLAGS) $(CPPFLAGS) -MMD -MF $(addsuffix .dep, $(basename $@)) -c -o $@ $<
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Initial import.
10 years ago
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
install: install_c install_cxx
install_c: install-headers_c install-static_c install-shared_c
install_cxx: install-headers_cxx install-static_cxx install-shared_cxx
install-headers: install-headers_c install-headers_cxx
Initial import.
10 years ago
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
install-headers_c:
$(E) "[INSTALL] Installing public C headers"
$(Q) $(foreach h, $(PUBLIC_HEADERS_C), $(INSTALL) $(h) $(prefix)/$(h) && ) exit 0 || exit 1
Initial import.
10 years ago
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
install-headers_cxx:
$(E) "[INSTALL] Installing public C++ headers"
$(Q) $(foreach h, $(PUBLIC_HEADERS_CXX), $(INSTALL) $(h) $(prefix)/$(h) && ) exit 0 || exit 1
install-static: install-static_c install-static_cxx
install-static_c: static_c strip-static_c
Initial import.
10 years ago
$(E) "[INSTALL] Installing libgpr.a"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(INSTALL) libs/$(CONFIG)/libgpr.a $(prefix)/lib/libgpr.a
Initial import.
10 years ago
$(E) "[INSTALL] Installing libgrpc.a"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(INSTALL) libs/$(CONFIG)/libgrpc.a $(prefix)/lib/libgrpc.a
Initial import.
10 years ago
$(E) "[INSTALL] Installing libgrpc_unsecure.a"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(INSTALL) libs/$(CONFIG)/libgrpc_unsecure.a $(prefix)/lib/libgrpc_unsecure.a
Initial import.
10 years ago
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
install-static_cxx: static_cxx strip-static_cxx
$(E) "[INSTALL] Installing libgrpc++.a"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(INSTALL) libs/$(CONFIG)/libgrpc++.a $(prefix)/lib/libgrpc++.a
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
install-shared_c: shared_c strip-shared_c
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
ifeq ($(SYSTEM),MINGW32)
$(E) "[INSTALL] Installing gpr.$(SHARED_EXT)"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(INSTALL) libs/$(CONFIG)/gpr.$(SHARED_EXT) $(prefix)/lib/gpr.$(SHARED_EXT)
$(Q) $(INSTALL) libs/$(CONFIG)/libgpr-imp.a $(prefix)/lib/libgpr-imp.a
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
else
$(E) "[INSTALL] Installing libgpr.$(SHARED_EXT)"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(INSTALL) libs/$(CONFIG)/libgpr.$(SHARED_EXT) $(prefix)/lib/libgpr.$(SHARED_EXT)
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
ifneq ($(SYSTEM),Darwin)
$(Q) ln -sf libgpr.$(SHARED_EXT) $(prefix)/lib/libgpr.so
endif
endif
ifeq ($(SYSTEM),MINGW32)
$(E) "[INSTALL] Installing grpc.$(SHARED_EXT)"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(INSTALL) libs/$(CONFIG)/grpc.$(SHARED_EXT) $(prefix)/lib/grpc.$(SHARED_EXT)
$(Q) $(INSTALL) libs/$(CONFIG)/libgrpc-imp.a $(prefix)/lib/libgrpc-imp.a
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
else
$(E) "[INSTALL] Installing libgrpc.$(SHARED_EXT)"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(INSTALL) libs/$(CONFIG)/libgrpc.$(SHARED_EXT) $(prefix)/lib/libgrpc.$(SHARED_EXT)
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
ifneq ($(SYSTEM),Darwin)
$(Q) ln -sf libgrpc.$(SHARED_EXT) $(prefix)/lib/libgrpc.so
endif
endif
ifeq ($(SYSTEM),MINGW32)
$(E) "[INSTALL] Installing grpc_unsecure.$(SHARED_EXT)"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(INSTALL) libs/$(CONFIG)/grpc_unsecure.$(SHARED_EXT) $(prefix)/lib/grpc_unsecure.$(SHARED_EXT)
$(Q) $(INSTALL) libs/$(CONFIG)/libgrpc_unsecure-imp.a $(prefix)/lib/libgrpc_unsecure-imp.a
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
else
$(E) "[INSTALL] Installing libgrpc_unsecure.$(SHARED_EXT)"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(INSTALL) libs/$(CONFIG)/libgrpc_unsecure.$(SHARED_EXT) $(prefix)/lib/libgrpc_unsecure.$(SHARED_EXT)
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
ifneq ($(SYSTEM),Darwin)
$(Q) ln -sf libgrpc_unsecure.$(SHARED_EXT) $(prefix)/lib/libgrpc_unsecure.so
endif
endif
ifneq ($(SYSTEM),MINGW32)
ifneq ($(SYSTEM),Darwin)
$(Q) ldconfig
endif
endif
Initial import.
10 years ago
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
install-shared_cxx: shared_cxx strip-shared_cxx
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
ifeq ($(SYSTEM),MINGW32)
$(E) "[INSTALL] Installing grpc++.$(SHARED_EXT)"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(INSTALL) libs/$(CONFIG)/grpc++.$(SHARED_EXT) $(prefix)/lib/grpc++.$(SHARED_EXT)
$(Q) $(INSTALL) libs/$(CONFIG)/libgrpc++-imp.a $(prefix)/lib/libgrpc++-imp.a
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
else
$(E) "[INSTALL] Installing libgrpc++.$(SHARED_EXT)"
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(INSTALL) libs/$(CONFIG)/libgrpc++.$(SHARED_EXT) $(prefix)/lib/libgrpc++.$(SHARED_EXT)
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
ifneq ($(SYSTEM),Darwin)
$(Q) ln -sf libgrpc++.$(SHARED_EXT) $(prefix)/lib/libgrpc++.so
endif
endif
ifneq ($(SYSTEM),MINGW32)
ifneq ($(SYSTEM),Darwin)
$(Q) ldconfig
endif
endif
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
Remove clean rules It's safe enough to just delete the build product directories.
10 years ago
clean:
Generate dependencies when compiling the .o If we don't have a .o, we don't need dependencies, so skip the setup step of building them. Instead, when building the .o, we generate the .dep and output it (to objs/ to make the rules a little easier). The next run of make will include the dep file and any dependencies will be picked up. This change also disables the disabling of dependency checking if we have zlib or openssl being compiled from third_party. Additionally it inverts the logic for including dependencies from ONLY if we are doing a clean to ONLY if we are NOT doing a clean.
10 years ago
$(Q) $(RM) -rf objs libs bins gens
Initial import.
10 years ago
# The various libraries
LIBGPR_SRC = \
src/core/support/alloc.c \
src/core/support/cancellable.c \
src/core/support/cmdline.c \
Provide a port of cpu related code to posix, and linux. Properly set this up in our build environment. Change on 2014/12/23 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82725899
10 years ago
src/core/support/cpu_linux.c \
Initial import.
10 years ago
src/core/support/cpu_posix.c \
src/core/support/histogram.c \
src/core/support/host_port.c \
Remove endpoint/ directories. Fold endpoint interface into iomgr, move secure_endpoint into security/. This will make it easier for endpoint to rely on some iomgr defined types (like pollset). Change on 2014/12/17 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82338036
10 years ago
src/core/support/log_android.c \
Initial import.
10 years ago
src/core/support/log.c \
src/core/support/log_linux.c \
Remove endpoint/ directories. Fold endpoint interface into iomgr, move secure_endpoint into security/. This will make it easier for endpoint to rely on some iomgr defined types (like pollset). Change on 2014/12/17 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82338036
10 years ago
src/core/support/log_posix.c \
Initial import.
10 years ago
src/core/support/log_win32.c \
src/core/support/murmur_hash.c \
src/core/support/slice_buffer.c \
Remove endpoint/ directories. Fold endpoint interface into iomgr, move secure_endpoint into security/. This will make it easier for endpoint to rely on some iomgr defined types (like pollset). Change on 2014/12/17 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82338036
10 years ago
src/core/support/slice.c \
Initial import.
10 years ago
src/core/support/string.c \
src/core/support/string_posix.c \
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
src/core/support/string_win32.c \
Initial import.
10 years ago
src/core/support/sync.c \
src/core/support/sync_posix.c \
Fixes for GPR library on Windows (+ getting rid of warnings under MS C++ compiler) Change on 2014/12/09 by jtattermusch <jtattermusch@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81704538
10 years ago
src/core/support/sync_win32.c \
Initial import.
10 years ago
src/core/support/thd_posix.c \
src/core/support/thd_win32.c \
src/core/support/time.c \
src/core/support/time_posix.c \
src/core/support/time_win32.c \
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
PUBLIC_HEADERS_C += \
Initial import.
10 years ago
include/grpc/support/alloc.h \
include/grpc/support/atm_gcc_atomic.h \
include/grpc/support/atm_gcc_sync.h \
include/grpc/support/atm.h \
include/grpc/support/atm_win32.h \
include/grpc/support/cancellable_platform.h \
include/grpc/support/cmdline.h \
include/grpc/support/histogram.h \
include/grpc/support/host_port.h \
include/grpc/support/log.h \
include/grpc/support/port_platform.h \
include/grpc/support/slice_buffer.h \
include/grpc/support/slice.h \
include/grpc/support/string.h \
include/grpc/support/sync_generic.h \
include/grpc/support/sync.h \
include/grpc/support/sync_posix.h \
include/grpc/support/sync_win32.h \
include/grpc/support/thd.h \
include/grpc/support/thd_posix.h \
include/grpc/support/thd_win32.h \
include/grpc/support/time.h \
include/grpc/support/time_posix.h \
include/grpc/support/time_win32.h \
include/grpc/support/useful.h \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBGPR_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBGPR_SRC))))
Initial import.
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libgpr.a: $(ZLIB_DEP) $(LIBGPR_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libgpr.a $(LIBGPR_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
ifeq ($(SYSTEM),MINGW32)
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/gpr.$(SHARED_EXT): $(LIBGPR_OBJS) $(ZLIB_DEP)
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(LD) $(LDFLAGS) -Llibs/$(CONFIG) -shared -Wl,--output-def=libs/$(CONFIG)/gpr.def -Wl,--out-implib=libs/$(CONFIG)/libgpr-imp.a -o libs/$(CONFIG)/gpr.$(SHARED_EXT) $(LIBGPR_OBJS) $(LDLIBS)
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
else
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libgpr.$(SHARED_EXT): $(LIBGPR_OBJS) $(ZLIB_DEP)
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
ifeq ($(SYSTEM),Darwin)
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(LD) $(LDFLAGS) -Llibs/$(CONFIG) -dynamiclib -o libs/$(CONFIG)/libgpr.$(SHARED_EXT) $(LIBGPR_OBJS) $(LDLIBS)
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
else
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(LD) $(LDFLAGS) -Llibs/$(CONFIG) -shared -Wl,-soname,libgpr.so.0 -o libs/$(CONFIG)/libgpr.$(SHARED_EXT) $(LIBGPR_OBJS) $(LDLIBS)
$(Q) ln -sf libgpr.$(SHARED_EXT) libs/$(CONFIG)/libgpr.so
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
endif
endif
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBGPR_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/src/core/support/alloc.o:
objs/$(CONFIG)/src/core/support/cancellable.o:
objs/$(CONFIG)/src/core/support/cmdline.o:
objs/$(CONFIG)/src/core/support/cpu_linux.o:
objs/$(CONFIG)/src/core/support/cpu_posix.o:
objs/$(CONFIG)/src/core/support/histogram.o:
objs/$(CONFIG)/src/core/support/host_port.o:
objs/$(CONFIG)/src/core/support/log_android.o:
objs/$(CONFIG)/src/core/support/log.o:
objs/$(CONFIG)/src/core/support/log_linux.o:
objs/$(CONFIG)/src/core/support/log_posix.o:
objs/$(CONFIG)/src/core/support/log_win32.o:
objs/$(CONFIG)/src/core/support/murmur_hash.o:
objs/$(CONFIG)/src/core/support/slice_buffer.o:
objs/$(CONFIG)/src/core/support/slice.o:
objs/$(CONFIG)/src/core/support/string.o:
objs/$(CONFIG)/src/core/support/string_posix.o:
objs/$(CONFIG)/src/core/support/string_win32.o:
objs/$(CONFIG)/src/core/support/sync.o:
objs/$(CONFIG)/src/core/support/sync_posix.o:
objs/$(CONFIG)/src/core/support/sync_win32.o:
objs/$(CONFIG)/src/core/support/thd_posix.o:
objs/$(CONFIG)/src/core/support/thd_win32.o:
objs/$(CONFIG)/src/core/support/time.o:
objs/$(CONFIG)/src/core/support/time_posix.o:
objs/$(CONFIG)/src/core/support/time_win32.o:
Initial import.
10 years ago
LIBGRPC_SRC = \
Adding plugins to the mako rendering system. Change on 2015/01/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83302581
10 years ago
src/core/security/auth.c \
src/core/security/base64.c \
src/core/security/credentials.c \
Moving the google-internal security code in their own files. Checked that the opensource version still builds. Just addressing core for now. We will do c++ later. Change on 2015/01/12 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83804014
10 years ago
src/core/security/factories.c \
Adding plugins to the mako rendering system. Change on 2015/01/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83302581
10 years ago
src/core/security/google_root_certs.c \
src/core/security/json_token.c \
src/core/security/secure_endpoint.c \
src/core/security/secure_transport_setup.c \
src/core/security/security_context.c \
src/core/security/server_secure_chttp2.c \
src/core/tsi/fake_transport_security.c \
src/core/tsi/ssl_transport_security.c \
src/core/tsi/transport_security.c \
Initial import.
10 years ago
src/core/channel/call_op_string.c \
src/core/channel/census_filter.c \
src/core/channel/channel_args.c \
src/core/channel/channel_stack.c \
Split client_channel into client_channel & child_channel. Safely managing disconnection and goaways needs reference counting at the child channel level, which was near impossible to provide with the previous embedding of the child channel in the client channel. This is a (hopefully) no-op refactoring to provide that split. The next CL in this series will actually get disconnection and goaway somewhat right. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81985590
10 years ago
src/core/channel/child_channel.c \
Initial import.
10 years ago
src/core/channel/client_channel.c \
src/core/channel/client_setup.c \
src/core/channel/connected_channel.c \
src/core/channel/http_client_filter.c \
src/core/channel/http_filter.c \
src/core/channel/http_server_filter.c \
src/core/channel/metadata_buffer.c \
src/core/channel/noop_filter.c \
src/core/compression/algorithm.c \
src/core/compression/message_compress.c \
Introducing iomgr. Move eventmanager and platform dependent endpoint functionality into a single library called 'iomgr'. This is primarily to prepare for a Windows port - where posix socket semantics lead to poor quality code. Mostly this is a code movement CL, with some small changes to help prepare the way for porting: - em style fd objects can only be held internally in iomgr, and own their memory - added grpc_iomgr_create_endpoint_pair() to accomodate the common pattern of creating a tcp endpoint from the output of socketpair - this will help keep our tests portable - separated em alarm interface into a separate file, as this part of event manager is needed higher up the stack - made the eventmanager bits a true singleton, simplifying API's across the stack as there's no longer a reason to carry a pointer there. Initial design document is here: https://docs.google.com/document/d/1VmafcHvvrP5kwtQkz84R5yXF7u7fW-9Pn0bkSUQHDt8/edit?disco=AAAAARNByxg Change on 2014/12/09 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81716456
10 years ago
src/core/httpcli/format_request.c \
src/core/httpcli/httpcli.c \
src/core/httpcli/httpcli_security_context.c \
src/core/httpcli/parser.c \
Update build.json, fix opensource build. Change on 2014/12/20 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82580775
10 years ago
src/core/iomgr/alarm.c \
src/core/iomgr/alarm_heap.c \
Remove endpoint/ directories. Fold endpoint interface into iomgr, move secure_endpoint into security/. This will make it easier for endpoint to rely on some iomgr defined types (like pollset). Change on 2014/12/17 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82338036
10 years ago
src/core/iomgr/endpoint.c \
Introducing iomgr. Move eventmanager and platform dependent endpoint functionality into a single library called 'iomgr'. This is primarily to prepare for a Windows port - where posix socket semantics lead to poor quality code. Mostly this is a code movement CL, with some small changes to help prepare the way for porting: - em style fd objects can only be held internally in iomgr, and own their memory - added grpc_iomgr_create_endpoint_pair() to accomodate the common pattern of creating a tcp endpoint from the output of socketpair - this will help keep our tests portable - separated em alarm interface into a separate file, as this part of event manager is needed higher up the stack - made the eventmanager bits a true singleton, simplifying API's across the stack as there's no longer a reason to carry a pointer there. Initial design document is here: https://docs.google.com/document/d/1VmafcHvvrP5kwtQkz84R5yXF7u7fW-9Pn0bkSUQHDt8/edit?disco=AAAAARNByxg Change on 2014/12/09 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81716456
10 years ago
src/core/iomgr/endpoint_pair_posix.c \
Remove libevent. Fixed any exposed bugs across the stack. Add a poll() based implementation. Heavily leverages pollset infrastructure to allow small polls to be the norm. Exposes a mechanism to plug in epoll/kqueue for platforms where we have them. Simplify iomgr callbacks to return one bit of success or failure (instead of the multi valued result that was mostly unused previously). This will ease the burden on new implementations, and the previous system provided no real value anyway. Removed timeouts on endpoint read/write routines. This simplifies porting burden by providing a more orthogonal interface, and the functionality can always be replicated when desired by using an alarm combined with endpoint_shutdown. I'm fairly certain we ended up with this interface because it was convenient to do from libevent. Things that need attention still: - adding an fd to a pollset is O(n^2) - but this is probably ok given that we'll not use this for multipolling once platform specific implementations are added. - we rely on the backup poller too often - especially for SSL handshakes and for client connection establishment we should have a better mechanism ([] [] - Linux needs to use epoll for multiple fds, FreeBSD variants (including Darwin) need to use kqueue. ([] [] - Linux needs to use eventfd for poll kicking. ([] Change on 2015/01/07 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83461069
10 years ago
src/core/iomgr/fd_posix.c \
src/core/iomgr/iomgr.c \
src/core/iomgr/iomgr_posix.c \
src/core/iomgr/pollset_multipoller_with_poll_posix.c \
src/core/iomgr/pollset_posix.c \
Introducing iomgr. Move eventmanager and platform dependent endpoint functionality into a single library called 'iomgr'. This is primarily to prepare for a Windows port - where posix socket semantics lead to poor quality code. Mostly this is a code movement CL, with some small changes to help prepare the way for porting: - em style fd objects can only be held internally in iomgr, and own their memory - added grpc_iomgr_create_endpoint_pair() to accomodate the common pattern of creating a tcp endpoint from the output of socketpair - this will help keep our tests portable - separated em alarm interface into a separate file, as this part of event manager is needed higher up the stack - made the eventmanager bits a true singleton, simplifying API's across the stack as there's no longer a reason to carry a pointer there. Initial design document is here: https://docs.google.com/document/d/1VmafcHvvrP5kwtQkz84R5yXF7u7fW-9Pn0bkSUQHDt8/edit?disco=AAAAARNByxg Change on 2014/12/09 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81716456
10 years ago
src/core/iomgr/resolve_address_posix.c \
src/core/iomgr/sockaddr_utils.c \
src/core/iomgr/socket_utils_common_posix.c \
src/core/iomgr/socket_utils_linux.c \
src/core/iomgr/socket_utils_posix.c \
src/core/iomgr/tcp_client_posix.c \
src/core/iomgr/tcp_posix.c \
src/core/iomgr/tcp_server_posix.c \
Add time averaged stats for future alarm list work. This is a trivial C89-ification of the []2 implementation of the same idea. Indeed the implementation files and tests have been branched from the [] versions. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82168822
10 years ago
src/core/iomgr/time_averaged_stats.c \
Introducing iomgr. Move eventmanager and platform dependent endpoint functionality into a single library called 'iomgr'. This is primarily to prepare for a Windows port - where posix socket semantics lead to poor quality code. Mostly this is a code movement CL, with some small changes to help prepare the way for porting: - em style fd objects can only be held internally in iomgr, and own their memory - added grpc_iomgr_create_endpoint_pair() to accomodate the common pattern of creating a tcp endpoint from the output of socketpair - this will help keep our tests portable - separated em alarm interface into a separate file, as this part of event manager is needed higher up the stack - made the eventmanager bits a true singleton, simplifying API's across the stack as there's no longer a reason to carry a pointer there. Initial design document is here: https://docs.google.com/document/d/1VmafcHvvrP5kwtQkz84R5yXF7u7fW-9Pn0bkSUQHDt8/edit?disco=AAAAARNByxg Change on 2014/12/09 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81716456
10 years ago
src/core/statistics/census_init.c \
Remove endpoint/ directories. Fold endpoint interface into iomgr, move secure_endpoint into security/. This will make it easier for endpoint to rely on some iomgr defined types (like pollset). Change on 2014/12/17 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82338036
10 years ago
src/core/statistics/census_log.c \
Introducing iomgr. Move eventmanager and platform dependent endpoint functionality into a single library called 'iomgr'. This is primarily to prepare for a Windows port - where posix socket semantics lead to poor quality code. Mostly this is a code movement CL, with some small changes to help prepare the way for porting: - em style fd objects can only be held internally in iomgr, and own their memory - added grpc_iomgr_create_endpoint_pair() to accomodate the common pattern of creating a tcp endpoint from the output of socketpair - this will help keep our tests portable - separated em alarm interface into a separate file, as this part of event manager is needed higher up the stack - made the eventmanager bits a true singleton, simplifying API's across the stack as there's no longer a reason to carry a pointer there. Initial design document is here: https://docs.google.com/document/d/1VmafcHvvrP5kwtQkz84R5yXF7u7fW-9Pn0bkSUQHDt8/edit?disco=AAAAARNByxg Change on 2014/12/09 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81716456
10 years ago
src/core/statistics/census_rpc_stats.c \
src/core/statistics/census_tracing.c \
src/core/statistics/hash_table.c \
src/core/statistics/window_stats.c \
Initial import.
10 years ago
src/core/surface/byte_buffer.c \
src/core/surface/byte_buffer_reader.c \
src/core/surface/call.c \
src/core/surface/channel.c \
src/core/surface/channel_create.c \
src/core/surface/client.c \
src/core/surface/completion_queue.c \
src/core/surface/event_string.c \
src/core/surface/init.c \
Introducing iomgr. Move eventmanager and platform dependent endpoint functionality into a single library called 'iomgr'. This is primarily to prepare for a Windows port - where posix socket semantics lead to poor quality code. Mostly this is a code movement CL, with some small changes to help prepare the way for porting: - em style fd objects can only be held internally in iomgr, and own their memory - added grpc_iomgr_create_endpoint_pair() to accomodate the common pattern of creating a tcp endpoint from the output of socketpair - this will help keep our tests portable - separated em alarm interface into a separate file, as this part of event manager is needed higher up the stack - made the eventmanager bits a true singleton, simplifying API's across the stack as there's no longer a reason to carry a pointer there. Initial design document is here: https://docs.google.com/document/d/1VmafcHvvrP5kwtQkz84R5yXF7u7fW-9Pn0bkSUQHDt8/edit?disco=AAAAARNByxg Change on 2014/12/09 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81716456
10 years ago
src/core/surface/lame_client.c \
src/core/surface/secure_channel_create.c \
src/core/surface/secure_server_create.c \
Initial import.
10 years ago
src/core/surface/server.c \
src/core/surface/server_chttp2.c \
src/core/surface/server_create.c \
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
src/core/transport/chttp2/alpn.c \
src/core/transport/chttp2/bin_encoder.c \
Initial import.
10 years ago
src/core/transport/chttp2/frame_data.c \
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
src/core/transport/chttp2/frame_goaway.c \
Initial import.
10 years ago
src/core/transport/chttp2/frame_ping.c \
src/core/transport/chttp2/frame_rst_stream.c \
src/core/transport/chttp2/frame_settings.c \
src/core/transport/chttp2/frame_window_update.c \
src/core/transport/chttp2/hpack_parser.c \
src/core/transport/chttp2/hpack_table.c \
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
src/core/transport/chttp2/huffsyms.c \
Initial import.
10 years ago
src/core/transport/chttp2/status_conversion.c \
src/core/transport/chttp2/stream_encoder.c \
src/core/transport/chttp2/stream_map.c \
src/core/transport/chttp2/timeout_encoding.c \
Add a --forever flag, to continuously run tests as things change. Change on 2015/01/07 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83451760
10 years ago
src/core/transport/chttp2/varint.c \
Remove libevent. Fixed any exposed bugs across the stack. Add a poll() based implementation. Heavily leverages pollset infrastructure to allow small polls to be the norm. Exposes a mechanism to plug in epoll/kqueue for platforms where we have them. Simplify iomgr callbacks to return one bit of success or failure (instead of the multi valued result that was mostly unused previously). This will ease the burden on new implementations, and the previous system provided no real value anyway. Removed timeouts on endpoint read/write routines. This simplifies porting burden by providing a more orthogonal interface, and the functionality can always be replicated when desired by using an alarm combined with endpoint_shutdown. I'm fairly certain we ended up with this interface because it was convenient to do from libevent. Things that need attention still: - adding an fd to a pollset is O(n^2) - but this is probably ok given that we'll not use this for multipolling once platform specific implementations are added. - we rely on the backup poller too often - especially for SSL handshakes and for client connection establishment we should have a better mechanism ([] [] - Linux needs to use epoll for multiple fds, FreeBSD variants (including Darwin) need to use kqueue. ([] [] - Linux needs to use eventfd for poll kicking. ([] Change on 2015/01/07 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83461069
10 years ago
src/core/transport/chttp2_transport.c \
Initial import.
10 years ago
src/core/transport/metadata.c \
src/core/transport/stream_op.c \
src/core/transport/transport.c \
third_party/cJSON/cJSON.c \
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
PUBLIC_HEADERS_C += \
Adding plugins to the mako rendering system. Change on 2015/01/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83302581
10 years ago
include/grpc/grpc_security.h \
Initial import.
10 years ago
include/grpc/byte_buffer.h \
include/grpc/byte_buffer_reader.h \
include/grpc/grpc.h \
include/grpc/status.h \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBGRPC_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBGRPC_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure libraries if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
libs/$(CONFIG)/libgrpc.a: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
ifeq ($(SYSTEM),MINGW32)
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
libs/$(CONFIG)/grpc.$(SHARED_EXT): openssl_dep_error
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
else
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
libs/$(CONFIG)/libgrpc.$(SHARED_EXT): openssl_dep_error
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Fixing parallel build. The source code of the libraries depends on openssl being built, as it will setup the headers. Since we don't know exactly which source depends on openssl (we could start flagging it in build.json, but that'd be painful), we just flag all the source for every ssl-enabled library to depend on openssl being built.
10 years ago
ifneq ($(OPENSSL_DEP),)
src/core/security/auth.c: $(OPENSSL_DEP)
src/core/security/base64.c: $(OPENSSL_DEP)
src/core/security/credentials.c: $(OPENSSL_DEP)
src/core/security/factories.c: $(OPENSSL_DEP)
src/core/security/google_root_certs.c: $(OPENSSL_DEP)
src/core/security/json_token.c: $(OPENSSL_DEP)
src/core/security/secure_endpoint.c: $(OPENSSL_DEP)
src/core/security/secure_transport_setup.c: $(OPENSSL_DEP)
src/core/security/security_context.c: $(OPENSSL_DEP)
src/core/security/server_secure_chttp2.c: $(OPENSSL_DEP)
src/core/tsi/fake_transport_security.c: $(OPENSSL_DEP)
src/core/tsi/ssl_transport_security.c: $(OPENSSL_DEP)
src/core/tsi/transport_security.c: $(OPENSSL_DEP)
src/core/channel/call_op_string.c: $(OPENSSL_DEP)
src/core/channel/census_filter.c: $(OPENSSL_DEP)
src/core/channel/channel_args.c: $(OPENSSL_DEP)
src/core/channel/channel_stack.c: $(OPENSSL_DEP)
src/core/channel/child_channel.c: $(OPENSSL_DEP)
src/core/channel/client_channel.c: $(OPENSSL_DEP)
src/core/channel/client_setup.c: $(OPENSSL_DEP)
src/core/channel/connected_channel.c: $(OPENSSL_DEP)
src/core/channel/http_client_filter.c: $(OPENSSL_DEP)
src/core/channel/http_filter.c: $(OPENSSL_DEP)
src/core/channel/http_server_filter.c: $(OPENSSL_DEP)
src/core/channel/metadata_buffer.c: $(OPENSSL_DEP)
src/core/channel/noop_filter.c: $(OPENSSL_DEP)
src/core/compression/algorithm.c: $(OPENSSL_DEP)
src/core/compression/message_compress.c: $(OPENSSL_DEP)
src/core/httpcli/format_request.c: $(OPENSSL_DEP)
src/core/httpcli/httpcli.c: $(OPENSSL_DEP)
src/core/httpcli/httpcli_security_context.c: $(OPENSSL_DEP)
src/core/httpcli/parser.c: $(OPENSSL_DEP)
src/core/iomgr/alarm.c: $(OPENSSL_DEP)
src/core/iomgr/alarm_heap.c: $(OPENSSL_DEP)
src/core/iomgr/endpoint.c: $(OPENSSL_DEP)
src/core/iomgr/endpoint_pair_posix.c: $(OPENSSL_DEP)
src/core/iomgr/fd_posix.c: $(OPENSSL_DEP)
src/core/iomgr/iomgr.c: $(OPENSSL_DEP)
src/core/iomgr/iomgr_posix.c: $(OPENSSL_DEP)
src/core/iomgr/pollset_multipoller_with_poll_posix.c: $(OPENSSL_DEP)
src/core/iomgr/pollset_posix.c: $(OPENSSL_DEP)
src/core/iomgr/resolve_address_posix.c: $(OPENSSL_DEP)
src/core/iomgr/sockaddr_utils.c: $(OPENSSL_DEP)
src/core/iomgr/socket_utils_common_posix.c: $(OPENSSL_DEP)
src/core/iomgr/socket_utils_linux.c: $(OPENSSL_DEP)
src/core/iomgr/socket_utils_posix.c: $(OPENSSL_DEP)
src/core/iomgr/tcp_client_posix.c: $(OPENSSL_DEP)
src/core/iomgr/tcp_posix.c: $(OPENSSL_DEP)
src/core/iomgr/tcp_server_posix.c: $(OPENSSL_DEP)
src/core/iomgr/time_averaged_stats.c: $(OPENSSL_DEP)
src/core/statistics/census_init.c: $(OPENSSL_DEP)
src/core/statistics/census_log.c: $(OPENSSL_DEP)
src/core/statistics/census_rpc_stats.c: $(OPENSSL_DEP)
src/core/statistics/census_tracing.c: $(OPENSSL_DEP)
src/core/statistics/hash_table.c: $(OPENSSL_DEP)
src/core/statistics/window_stats.c: $(OPENSSL_DEP)
src/core/surface/byte_buffer.c: $(OPENSSL_DEP)
src/core/surface/byte_buffer_reader.c: $(OPENSSL_DEP)
src/core/surface/call.c: $(OPENSSL_DEP)
src/core/surface/channel.c: $(OPENSSL_DEP)
src/core/surface/channel_create.c: $(OPENSSL_DEP)
src/core/surface/client.c: $(OPENSSL_DEP)
src/core/surface/completion_queue.c: $(OPENSSL_DEP)
src/core/surface/event_string.c: $(OPENSSL_DEP)
src/core/surface/init.c: $(OPENSSL_DEP)
src/core/surface/lame_client.c: $(OPENSSL_DEP)
src/core/surface/secure_channel_create.c: $(OPENSSL_DEP)
src/core/surface/secure_server_create.c: $(OPENSSL_DEP)
src/core/surface/server.c: $(OPENSSL_DEP)
src/core/surface/server_chttp2.c: $(OPENSSL_DEP)
src/core/surface/server_create.c: $(OPENSSL_DEP)
src/core/transport/chttp2/alpn.c: $(OPENSSL_DEP)
src/core/transport/chttp2/bin_encoder.c: $(OPENSSL_DEP)
src/core/transport/chttp2/frame_data.c: $(OPENSSL_DEP)
src/core/transport/chttp2/frame_goaway.c: $(OPENSSL_DEP)
src/core/transport/chttp2/frame_ping.c: $(OPENSSL_DEP)
src/core/transport/chttp2/frame_rst_stream.c: $(OPENSSL_DEP)
src/core/transport/chttp2/frame_settings.c: $(OPENSSL_DEP)
src/core/transport/chttp2/frame_window_update.c: $(OPENSSL_DEP)
src/core/transport/chttp2/hpack_parser.c: $(OPENSSL_DEP)
src/core/transport/chttp2/hpack_table.c: $(OPENSSL_DEP)
src/core/transport/chttp2/huffsyms.c: $(OPENSSL_DEP)
src/core/transport/chttp2/status_conversion.c: $(OPENSSL_DEP)
src/core/transport/chttp2/stream_encoder.c: $(OPENSSL_DEP)
src/core/transport/chttp2/stream_map.c: $(OPENSSL_DEP)
src/core/transport/chttp2/timeout_encoding.c: $(OPENSSL_DEP)
src/core/transport/chttp2/varint.c: $(OPENSSL_DEP)
src/core/transport/chttp2_transport.c: $(OPENSSL_DEP)
src/core/transport/metadata.c: $(OPENSSL_DEP)
src/core/transport/stream_op.c: $(OPENSSL_DEP)
src/core/transport/transport.c: $(OPENSSL_DEP)
third_party/cJSON/cJSON.c: $(OPENSSL_DEP)
endif
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libgrpc.a: $(ZLIB_DEP) $(OPENSSL_DEP) $(LIBGRPC_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libgrpc.a $(LIBGRPC_OBJS)
Fix some C++ test dependencies
10 years ago
$(Q) rm -rf tmp-merge
The built-in ssl target now properly merges the archives instead of nesting them. Change on 2014/12/16 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82273781
10 years ago
$(Q) mkdir tmp-merge
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) ( cd tmp-merge ; $(AR) x ../libs/$(CONFIG)/libgrpc.a )
The built-in ssl target now properly merges the archives instead of nesting them. Change on 2014/12/16 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82273781
10 years ago
$(Q) for l in $(OPENSSL_MERGE_LIBS) ; do ( cd tmp-merge ; ar x ../$${l} ) ; done
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) rm -f libs/$(CONFIG)/libgrpc.a tmp-merge/__.SYMDEF*
$(Q) ar rcs libs/$(CONFIG)/libgrpc.a tmp-merge/*
The built-in ssl target now properly merges the archives instead of nesting them. Change on 2014/12/16 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82273781
10 years ago
$(Q) rm -rf tmp-merge
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
ifeq ($(SYSTEM),MINGW32)
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/grpc.$(SHARED_EXT): $(LIBGRPC_OBJS) $(ZLIB_DEP)libs/$(CONFIG)/gpr.$(SHARED_EXT) $(OPENSSL_DEP)
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(LD) $(LDFLAGS) -Llibs/$(CONFIG) -shared -Wl,--output-def=libs/$(CONFIG)/grpc.def -Wl,--out-implib=libs/$(CONFIG)/libgrpc-imp.a -o libs/$(CONFIG)/grpc.$(SHARED_EXT) $(LIBGRPC_OBJS) $(LDLIBS) $(LDLIBS_SECURE) $(OPENSSL_MERGE_LIBS) -lgpr-imp
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
else
Add missing space
10 years ago
libs/$(CONFIG)/libgrpc.$(SHARED_EXT): $(LIBGRPC_OBJS) $(ZLIB_DEP) libs/$(CONFIG)/libgpr.$(SHARED_EXT) $(OPENSSL_DEP)
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
ifeq ($(SYSTEM),Darwin)
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(LD) $(LDFLAGS) -Llibs/$(CONFIG) -dynamiclib -o libs/$(CONFIG)/libgrpc.$(SHARED_EXT) $(LIBGRPC_OBJS) $(LDLIBS) $(LDLIBS_SECURE) $(OPENSSL_MERGE_LIBS) -lgpr
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
else
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(LD) $(LDFLAGS) -Llibs/$(CONFIG) -shared -Wl,-soname,libgrpc.so.0 -o libs/$(CONFIG)/libgrpc.$(SHARED_EXT) $(LIBGRPC_OBJS) $(LDLIBS) $(LDLIBS_SECURE) $(OPENSSL_MERGE_LIBS) -lgpr
$(Q) ln -sf libgrpc.$(SHARED_EXT) libs/$(CONFIG)/libgrpc.so
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
endif
endif
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBGRPC_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/src/core/security/auth.o:
objs/$(CONFIG)/src/core/security/base64.o:
objs/$(CONFIG)/src/core/security/credentials.o:
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/src/core/security/factories.o:
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/src/core/security/google_root_certs.o:
objs/$(CONFIG)/src/core/security/json_token.o:
objs/$(CONFIG)/src/core/security/secure_endpoint.o:
objs/$(CONFIG)/src/core/security/secure_transport_setup.o:
objs/$(CONFIG)/src/core/security/security_context.o:
objs/$(CONFIG)/src/core/security/server_secure_chttp2.o:
objs/$(CONFIG)/src/core/tsi/fake_transport_security.o:
objs/$(CONFIG)/src/core/tsi/ssl_transport_security.o:
objs/$(CONFIG)/src/core/tsi/transport_security.o:
objs/$(CONFIG)/src/core/channel/call_op_string.o:
objs/$(CONFIG)/src/core/channel/census_filter.o:
objs/$(CONFIG)/src/core/channel/channel_args.o:
objs/$(CONFIG)/src/core/channel/channel_stack.o:
objs/$(CONFIG)/src/core/channel/child_channel.o:
objs/$(CONFIG)/src/core/channel/client_channel.o:
objs/$(CONFIG)/src/core/channel/client_setup.o:
objs/$(CONFIG)/src/core/channel/connected_channel.o:
objs/$(CONFIG)/src/core/channel/http_client_filter.o:
objs/$(CONFIG)/src/core/channel/http_filter.o:
objs/$(CONFIG)/src/core/channel/http_server_filter.o:
objs/$(CONFIG)/src/core/channel/metadata_buffer.o:
objs/$(CONFIG)/src/core/channel/noop_filter.o:
objs/$(CONFIG)/src/core/compression/algorithm.o:
objs/$(CONFIG)/src/core/compression/message_compress.o:
objs/$(CONFIG)/src/core/httpcli/format_request.o:
objs/$(CONFIG)/src/core/httpcli/httpcli.o:
objs/$(CONFIG)/src/core/httpcli/httpcli_security_context.o:
objs/$(CONFIG)/src/core/httpcli/parser.o:
objs/$(CONFIG)/src/core/iomgr/alarm.o:
objs/$(CONFIG)/src/core/iomgr/alarm_heap.o:
objs/$(CONFIG)/src/core/iomgr/endpoint.o:
objs/$(CONFIG)/src/core/iomgr/endpoint_pair_posix.o:
objs/$(CONFIG)/src/core/iomgr/fd_posix.o:
objs/$(CONFIG)/src/core/iomgr/iomgr.o:
objs/$(CONFIG)/src/core/iomgr/iomgr_posix.o:
objs/$(CONFIG)/src/core/iomgr/pollset_multipoller_with_poll_posix.o:
objs/$(CONFIG)/src/core/iomgr/pollset_posix.o:
objs/$(CONFIG)/src/core/iomgr/resolve_address_posix.o:
objs/$(CONFIG)/src/core/iomgr/sockaddr_utils.o:
objs/$(CONFIG)/src/core/iomgr/socket_utils_common_posix.o:
objs/$(CONFIG)/src/core/iomgr/socket_utils_linux.o:
objs/$(CONFIG)/src/core/iomgr/socket_utils_posix.o:
objs/$(CONFIG)/src/core/iomgr/tcp_client_posix.o:
objs/$(CONFIG)/src/core/iomgr/tcp_posix.o:
objs/$(CONFIG)/src/core/iomgr/tcp_server_posix.o:
objs/$(CONFIG)/src/core/iomgr/time_averaged_stats.o:
objs/$(CONFIG)/src/core/statistics/census_init.o:
objs/$(CONFIG)/src/core/statistics/census_log.o:
objs/$(CONFIG)/src/core/statistics/census_rpc_stats.o:
objs/$(CONFIG)/src/core/statistics/census_tracing.o:
objs/$(CONFIG)/src/core/statistics/hash_table.o:
objs/$(CONFIG)/src/core/statistics/window_stats.o:
objs/$(CONFIG)/src/core/surface/byte_buffer.o:
objs/$(CONFIG)/src/core/surface/byte_buffer_reader.o:
objs/$(CONFIG)/src/core/surface/call.o:
objs/$(CONFIG)/src/core/surface/channel.o:
objs/$(CONFIG)/src/core/surface/channel_create.o:
objs/$(CONFIG)/src/core/surface/client.o:
objs/$(CONFIG)/src/core/surface/completion_queue.o:
objs/$(CONFIG)/src/core/surface/event_string.o:
objs/$(CONFIG)/src/core/surface/init.o:
objs/$(CONFIG)/src/core/surface/lame_client.o:
objs/$(CONFIG)/src/core/surface/secure_channel_create.o:
objs/$(CONFIG)/src/core/surface/secure_server_create.o:
objs/$(CONFIG)/src/core/surface/server.o:
objs/$(CONFIG)/src/core/surface/server_chttp2.o:
objs/$(CONFIG)/src/core/surface/server_create.o:
objs/$(CONFIG)/src/core/transport/chttp2/alpn.o:
objs/$(CONFIG)/src/core/transport/chttp2/bin_encoder.o:
objs/$(CONFIG)/src/core/transport/chttp2/frame_data.o:
objs/$(CONFIG)/src/core/transport/chttp2/frame_goaway.o:
objs/$(CONFIG)/src/core/transport/chttp2/frame_ping.o:
objs/$(CONFIG)/src/core/transport/chttp2/frame_rst_stream.o:
objs/$(CONFIG)/src/core/transport/chttp2/frame_settings.o:
objs/$(CONFIG)/src/core/transport/chttp2/frame_window_update.o:
objs/$(CONFIG)/src/core/transport/chttp2/hpack_parser.o:
objs/$(CONFIG)/src/core/transport/chttp2/hpack_table.o:
objs/$(CONFIG)/src/core/transport/chttp2/huffsyms.o:
objs/$(CONFIG)/src/core/transport/chttp2/status_conversion.o:
objs/$(CONFIG)/src/core/transport/chttp2/stream_encoder.o:
objs/$(CONFIG)/src/core/transport/chttp2/stream_map.o:
objs/$(CONFIG)/src/core/transport/chttp2/timeout_encoding.o:
objs/$(CONFIG)/src/core/transport/chttp2/varint.o:
objs/$(CONFIG)/src/core/transport/chttp2_transport.o:
objs/$(CONFIG)/src/core/transport/metadata.o:
objs/$(CONFIG)/src/core/transport/stream_op.o:
objs/$(CONFIG)/src/core/transport/transport.o:
objs/$(CONFIG)/third_party/cJSON/cJSON.o:
Initial import.
10 years ago
Adding plugins to the mako rendering system. Change on 2015/01/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83302581
10 years ago
LIBGRPC_UNSECURE_SRC = \
src/core/channel/call_op_string.c \
src/core/channel/census_filter.c \
src/core/channel/channel_args.c \
src/core/channel/channel_stack.c \
src/core/channel/child_channel.c \
src/core/channel/client_channel.c \
src/core/channel/client_setup.c \
src/core/channel/connected_channel.c \
src/core/channel/http_client_filter.c \
src/core/channel/http_filter.c \
src/core/channel/http_server_filter.c \
src/core/channel/metadata_buffer.c \
src/core/channel/noop_filter.c \
src/core/compression/algorithm.c \
src/core/compression/message_compress.c \
src/core/httpcli/format_request.c \
src/core/httpcli/httpcli.c \
src/core/httpcli/httpcli_security_context.c \
src/core/httpcli/parser.c \
src/core/iomgr/alarm.c \
src/core/iomgr/alarm_heap.c \
src/core/iomgr/endpoint.c \
src/core/iomgr/endpoint_pair_posix.c \
Remove libevent. Fixed any exposed bugs across the stack. Add a poll() based implementation. Heavily leverages pollset infrastructure to allow small polls to be the norm. Exposes a mechanism to plug in epoll/kqueue for platforms where we have them. Simplify iomgr callbacks to return one bit of success or failure (instead of the multi valued result that was mostly unused previously). This will ease the burden on new implementations, and the previous system provided no real value anyway. Removed timeouts on endpoint read/write routines. This simplifies porting burden by providing a more orthogonal interface, and the functionality can always be replicated when desired by using an alarm combined with endpoint_shutdown. I'm fairly certain we ended up with this interface because it was convenient to do from libevent. Things that need attention still: - adding an fd to a pollset is O(n^2) - but this is probably ok given that we'll not use this for multipolling once platform specific implementations are added. - we rely on the backup poller too often - especially for SSL handshakes and for client connection establishment we should have a better mechanism ([] [] - Linux needs to use epoll for multiple fds, FreeBSD variants (including Darwin) need to use kqueue. ([] [] - Linux needs to use eventfd for poll kicking. ([] Change on 2015/01/07 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83461069
10 years ago
src/core/iomgr/fd_posix.c \
src/core/iomgr/iomgr.c \
src/core/iomgr/iomgr_posix.c \
src/core/iomgr/pollset_multipoller_with_poll_posix.c \
src/core/iomgr/pollset_posix.c \
Adding plugins to the mako rendering system. Change on 2015/01/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83302581
10 years ago
src/core/iomgr/resolve_address_posix.c \
src/core/iomgr/sockaddr_utils.c \
src/core/iomgr/socket_utils_common_posix.c \
src/core/iomgr/socket_utils_linux.c \
src/core/iomgr/socket_utils_posix.c \
src/core/iomgr/tcp_client_posix.c \
src/core/iomgr/tcp_posix.c \
src/core/iomgr/tcp_server_posix.c \
src/core/iomgr/time_averaged_stats.c \
src/core/statistics/census_init.c \
src/core/statistics/census_log.c \
src/core/statistics/census_rpc_stats.c \
src/core/statistics/census_tracing.c \
src/core/statistics/hash_table.c \
src/core/statistics/window_stats.c \
src/core/surface/byte_buffer.c \
src/core/surface/byte_buffer_reader.c \
src/core/surface/call.c \
src/core/surface/channel.c \
src/core/surface/channel_create.c \
src/core/surface/client.c \
src/core/surface/completion_queue.c \
src/core/surface/event_string.c \
src/core/surface/init.c \
src/core/surface/lame_client.c \
src/core/surface/secure_channel_create.c \
src/core/surface/secure_server_create.c \
src/core/surface/server.c \
src/core/surface/server_chttp2.c \
src/core/surface/server_create.c \
src/core/transport/chttp2/alpn.c \
src/core/transport/chttp2/bin_encoder.c \
src/core/transport/chttp2/frame_data.c \
src/core/transport/chttp2/frame_goaway.c \
src/core/transport/chttp2/frame_ping.c \
src/core/transport/chttp2/frame_rst_stream.c \
src/core/transport/chttp2/frame_settings.c \
src/core/transport/chttp2/frame_window_update.c \
src/core/transport/chttp2/hpack_parser.c \
src/core/transport/chttp2/hpack_table.c \
src/core/transport/chttp2/huffsyms.c \
src/core/transport/chttp2/status_conversion.c \
src/core/transport/chttp2/stream_encoder.c \
src/core/transport/chttp2/stream_map.c \
src/core/transport/chttp2/timeout_encoding.c \
Add a --forever flag, to continuously run tests as things change. Change on 2015/01/07 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83451760
10 years ago
src/core/transport/chttp2/varint.c \
Remove libevent. Fixed any exposed bugs across the stack. Add a poll() based implementation. Heavily leverages pollset infrastructure to allow small polls to be the norm. Exposes a mechanism to plug in epoll/kqueue for platforms where we have them. Simplify iomgr callbacks to return one bit of success or failure (instead of the multi valued result that was mostly unused previously). This will ease the burden on new implementations, and the previous system provided no real value anyway. Removed timeouts on endpoint read/write routines. This simplifies porting burden by providing a more orthogonal interface, and the functionality can always be replicated when desired by using an alarm combined with endpoint_shutdown. I'm fairly certain we ended up with this interface because it was convenient to do from libevent. Things that need attention still: - adding an fd to a pollset is O(n^2) - but this is probably ok given that we'll not use this for multipolling once platform specific implementations are added. - we rely on the backup poller too often - especially for SSL handshakes and for client connection establishment we should have a better mechanism ([] [] - Linux needs to use epoll for multiple fds, FreeBSD variants (including Darwin) need to use kqueue. ([] [] - Linux needs to use eventfd for poll kicking. ([] Change on 2015/01/07 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83461069
10 years ago
src/core/transport/chttp2_transport.c \
Adding plugins to the mako rendering system. Change on 2015/01/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83302581
10 years ago
src/core/transport/metadata.c \
src/core/transport/stream_op.c \
src/core/transport/transport.c \
third_party/cJSON/cJSON.c \
PUBLIC_HEADERS_C += \
include/grpc/byte_buffer.h \
include/grpc/byte_buffer_reader.h \
include/grpc/grpc.h \
include/grpc/status.h \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBGRPC_UNSECURE_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBGRPC_UNSECURE_SRC))))
Adding plugins to the mako rendering system. Change on 2015/01/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83302581
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libgrpc_unsecure.a: $(ZLIB_DEP) $(LIBGRPC_UNSECURE_OBJS)
Adding plugins to the mako rendering system. Change on 2015/01/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83302581
10 years ago
$(E) "[AR] Creating $@"
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libgrpc_unsecure.a $(LIBGRPC_UNSECURE_OBJS)
Adding plugins to the mako rendering system. Change on 2015/01/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83302581
10 years ago
ifeq ($(SYSTEM),MINGW32)
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/grpc_unsecure.$(SHARED_EXT): $(LIBGRPC_UNSECURE_OBJS) $(ZLIB_DEP)libs/$(CONFIG)/gpr.$(SHARED_EXT)
Adding plugins to the mako rendering system. Change on 2015/01/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83302581
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(LD) $(LDFLAGS) -Llibs/$(CONFIG) -shared -Wl,--output-def=libs/$(CONFIG)/grpc_unsecure.def -Wl,--out-implib=libs/$(CONFIG)/libgrpc_unsecure-imp.a -o libs/$(CONFIG)/grpc_unsecure.$(SHARED_EXT) $(LIBGRPC_UNSECURE_OBJS) $(LDLIBS) -lgpr-imp
Adding plugins to the mako rendering system. Change on 2015/01/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83302581
10 years ago
else
Add missing space
10 years ago
libs/$(CONFIG)/libgrpc_unsecure.$(SHARED_EXT): $(LIBGRPC_UNSECURE_OBJS) $(ZLIB_DEP) libs/$(CONFIG)/libgpr.$(SHARED_EXT)
Adding plugins to the mako rendering system. Change on 2015/01/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83302581
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
ifeq ($(SYSTEM),Darwin)
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(LD) $(LDFLAGS) -Llibs/$(CONFIG) -dynamiclib -o libs/$(CONFIG)/libgrpc_unsecure.$(SHARED_EXT) $(LIBGRPC_UNSECURE_OBJS) $(LDLIBS) -lgpr
Adding plugins to the mako rendering system. Change on 2015/01/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83302581
10 years ago
else
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(LD) $(LDFLAGS) -Llibs/$(CONFIG) -shared -Wl,-soname,libgrpc_unsecure.so.0 -o libs/$(CONFIG)/libgrpc_unsecure.$(SHARED_EXT) $(LIBGRPC_UNSECURE_OBJS) $(LDLIBS) -lgpr
$(Q) ln -sf libgrpc_unsecure.$(SHARED_EXT) libs/$(CONFIG)/libgrpc_unsecure.so
Adding plugins to the mako rendering system. Change on 2015/01/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83302581
10 years ago
endif
endif
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBGRPC_UNSECURE_OBJS:.o=.dep)
Adding plugins to the mako rendering system. Change on 2015/01/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83302581
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/src/core/channel/call_op_string.o:
objs/$(CONFIG)/src/core/channel/census_filter.o:
objs/$(CONFIG)/src/core/channel/channel_args.o:
objs/$(CONFIG)/src/core/channel/channel_stack.o:
objs/$(CONFIG)/src/core/channel/child_channel.o:
objs/$(CONFIG)/src/core/channel/client_channel.o:
objs/$(CONFIG)/src/core/channel/client_setup.o:
objs/$(CONFIG)/src/core/channel/connected_channel.o:
objs/$(CONFIG)/src/core/channel/http_client_filter.o:
objs/$(CONFIG)/src/core/channel/http_filter.o:
objs/$(CONFIG)/src/core/channel/http_server_filter.o:
objs/$(CONFIG)/src/core/channel/metadata_buffer.o:
objs/$(CONFIG)/src/core/channel/noop_filter.o:
objs/$(CONFIG)/src/core/compression/algorithm.o:
objs/$(CONFIG)/src/core/compression/message_compress.o:
objs/$(CONFIG)/src/core/httpcli/format_request.o:
objs/$(CONFIG)/src/core/httpcli/httpcli.o:
objs/$(CONFIG)/src/core/httpcli/httpcli_security_context.o:
objs/$(CONFIG)/src/core/httpcli/parser.o:
objs/$(CONFIG)/src/core/iomgr/alarm.o:
objs/$(CONFIG)/src/core/iomgr/alarm_heap.o:
objs/$(CONFIG)/src/core/iomgr/endpoint.o:
objs/$(CONFIG)/src/core/iomgr/endpoint_pair_posix.o:
objs/$(CONFIG)/src/core/iomgr/fd_posix.o:
objs/$(CONFIG)/src/core/iomgr/iomgr.o:
objs/$(CONFIG)/src/core/iomgr/iomgr_posix.o:
objs/$(CONFIG)/src/core/iomgr/pollset_multipoller_with_poll_posix.o:
objs/$(CONFIG)/src/core/iomgr/pollset_posix.o:
objs/$(CONFIG)/src/core/iomgr/resolve_address_posix.o:
objs/$(CONFIG)/src/core/iomgr/sockaddr_utils.o:
objs/$(CONFIG)/src/core/iomgr/socket_utils_common_posix.o:
objs/$(CONFIG)/src/core/iomgr/socket_utils_linux.o:
objs/$(CONFIG)/src/core/iomgr/socket_utils_posix.o:
objs/$(CONFIG)/src/core/iomgr/tcp_client_posix.o:
objs/$(CONFIG)/src/core/iomgr/tcp_posix.o:
objs/$(CONFIG)/src/core/iomgr/tcp_server_posix.o:
objs/$(CONFIG)/src/core/iomgr/time_averaged_stats.o:
objs/$(CONFIG)/src/core/statistics/census_init.o:
objs/$(CONFIG)/src/core/statistics/census_log.o:
objs/$(CONFIG)/src/core/statistics/census_rpc_stats.o:
objs/$(CONFIG)/src/core/statistics/census_tracing.o:
objs/$(CONFIG)/src/core/statistics/hash_table.o:
objs/$(CONFIG)/src/core/statistics/window_stats.o:
objs/$(CONFIG)/src/core/surface/byte_buffer.o:
objs/$(CONFIG)/src/core/surface/byte_buffer_reader.o:
objs/$(CONFIG)/src/core/surface/call.o:
objs/$(CONFIG)/src/core/surface/channel.o:
objs/$(CONFIG)/src/core/surface/channel_create.o:
objs/$(CONFIG)/src/core/surface/client.o:
objs/$(CONFIG)/src/core/surface/completion_queue.o:
objs/$(CONFIG)/src/core/surface/event_string.o:
objs/$(CONFIG)/src/core/surface/init.o:
objs/$(CONFIG)/src/core/surface/lame_client.o:
objs/$(CONFIG)/src/core/surface/secure_channel_create.o:
objs/$(CONFIG)/src/core/surface/secure_server_create.o:
objs/$(CONFIG)/src/core/surface/server.o:
objs/$(CONFIG)/src/core/surface/server_chttp2.o:
objs/$(CONFIG)/src/core/surface/server_create.o:
objs/$(CONFIG)/src/core/transport/chttp2/alpn.o:
objs/$(CONFIG)/src/core/transport/chttp2/bin_encoder.o:
objs/$(CONFIG)/src/core/transport/chttp2/frame_data.o:
objs/$(CONFIG)/src/core/transport/chttp2/frame_goaway.o:
objs/$(CONFIG)/src/core/transport/chttp2/frame_ping.o:
objs/$(CONFIG)/src/core/transport/chttp2/frame_rst_stream.o:
objs/$(CONFIG)/src/core/transport/chttp2/frame_settings.o:
objs/$(CONFIG)/src/core/transport/chttp2/frame_window_update.o:
objs/$(CONFIG)/src/core/transport/chttp2/hpack_parser.o:
objs/$(CONFIG)/src/core/transport/chttp2/hpack_table.o:
objs/$(CONFIG)/src/core/transport/chttp2/huffsyms.o:
objs/$(CONFIG)/src/core/transport/chttp2/status_conversion.o:
objs/$(CONFIG)/src/core/transport/chttp2/stream_encoder.o:
objs/$(CONFIG)/src/core/transport/chttp2/stream_map.o:
objs/$(CONFIG)/src/core/transport/chttp2/timeout_encoding.o:
objs/$(CONFIG)/src/core/transport/chttp2/varint.o:
objs/$(CONFIG)/src/core/transport/chttp2_transport.o:
objs/$(CONFIG)/src/core/transport/metadata.o:
objs/$(CONFIG)/src/core/transport/stream_op.o:
objs/$(CONFIG)/src/core/transport/transport.o:
objs/$(CONFIG)/third_party/cJSON/cJSON.o:
Adding plugins to the mako rendering system. Change on 2015/01/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83302581
10 years ago
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
LIBGPR_TEST_UTIL_SRC = \
test/core/util/test_config.c \
LIBGPR_TEST_UTIL_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBGPR_TEST_UTIL_SRC))))
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure libraries if you don't have OpenSSL with ALPN.
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
libs/$(CONFIG)/libgpr_test_util.a: openssl_dep_error
else
Fixing parallel build. The source code of the libraries depends on openssl being built, as it will setup the headers. Since we don't know exactly which source depends on openssl (we could start flagging it in build.json, but that'd be painful), we just flag all the source for every ssl-enabled library to depend on openssl being built.
10 years ago
ifneq ($(OPENSSL_DEP),)
test/core/util/test_config.c: $(OPENSSL_DEP)
endif
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libgpr_test_util.a: $(ZLIB_DEP) $(OPENSSL_DEP) $(LIBGPR_TEST_UTIL_OBJS)
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(E) "[AR] Creating $@"
$(Q) mkdir -p `dirname $@`
$(Q) $(AR) rcs libs/$(CONFIG)/libgpr_test_util.a $(LIBGPR_TEST_UTIL_OBJS)
endif
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBGPR_TEST_UTIL_OBJS:.o=.dep)
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
endif
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/util/test_config.o:
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
Initial import.
10 years ago
LIBGRPC_TEST_UTIL_SRC = \
Remove endpoint/ directories. Fold endpoint interface into iomgr, move secure_endpoint into security/. This will make it easier for endpoint to rely on some iomgr defined types (like pollset). Change on 2014/12/17 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82338036
10 years ago
test/core/end2end/cq_verifier.c \
Add SSL root for production GFE. Change on 2014/12/19 by chenw <chenw@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82554526
10 years ago
test/core/end2end/data/test_root_cert.c \
test/core/end2end/data/prod_roots_certs.c \
Remove endpoint/ directories. Fold endpoint interface into iomgr, move secure_endpoint into security/. This will make it easier for endpoint to rely on some iomgr defined types (like pollset). Change on 2014/12/17 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82338036
10 years ago
test/core/end2end/data/server1_cert.c \
test/core/end2end/data/server1_key.c \
test/core/iomgr/endpoint_tests.c \
test/core/statistics/census_log_tests.c \
test/core/transport/transport_end2end_tests.c \
Initial import.
10 years ago
test/core/util/grpc_profiler.c \
Add mako templates to generate MS Visual Studio projects. Current state for Windows platform: -- gpr and grpc_test_util can be compiled -- gpr_log_test and gpr_cmdline_test pass Change on 2014/12/08 by jtattermusch <jtattermusch@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81615574
10 years ago
test/core/util/port_posix.c \
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
test/core/util/parse_hexstring.c \
Initial import.
10 years ago
test/core/util/slice_splitter.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBGRPC_TEST_UTIL_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBGRPC_TEST_UTIL_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure libraries if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
libs/$(CONFIG)/libgrpc_test_util.a: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Fixing parallel build. The source code of the libraries depends on openssl being built, as it will setup the headers. Since we don't know exactly which source depends on openssl (we could start flagging it in build.json, but that'd be painful), we just flag all the source for every ssl-enabled library to depend on openssl being built.
10 years ago
ifneq ($(OPENSSL_DEP),)
test/core/end2end/cq_verifier.c: $(OPENSSL_DEP)
test/core/end2end/data/test_root_cert.c: $(OPENSSL_DEP)
test/core/end2end/data/prod_roots_certs.c: $(OPENSSL_DEP)
test/core/end2end/data/server1_cert.c: $(OPENSSL_DEP)
test/core/end2end/data/server1_key.c: $(OPENSSL_DEP)
test/core/iomgr/endpoint_tests.c: $(OPENSSL_DEP)
test/core/statistics/census_log_tests.c: $(OPENSSL_DEP)
test/core/transport/transport_end2end_tests.c: $(OPENSSL_DEP)
test/core/util/grpc_profiler.c: $(OPENSSL_DEP)
test/core/util/port_posix.c: $(OPENSSL_DEP)
test/core/util/parse_hexstring.c: $(OPENSSL_DEP)
test/core/util/slice_splitter.c: $(OPENSSL_DEP)
endif
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libgrpc_test_util.a: $(ZLIB_DEP) $(OPENSSL_DEP) $(LIBGRPC_TEST_UTIL_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libgrpc_test_util.a $(LIBGRPC_TEST_UTIL_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBGRPC_TEST_UTIL_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/cq_verifier.o:
objs/$(CONFIG)/test/core/end2end/data/test_root_cert.o:
objs/$(CONFIG)/test/core/end2end/data/prod_roots_certs.o:
objs/$(CONFIG)/test/core/end2end/data/server1_cert.o:
objs/$(CONFIG)/test/core/end2end/data/server1_key.o:
objs/$(CONFIG)/test/core/iomgr/endpoint_tests.o:
objs/$(CONFIG)/test/core/statistics/census_log_tests.o:
objs/$(CONFIG)/test/core/transport/transport_end2end_tests.o:
objs/$(CONFIG)/test/core/util/grpc_profiler.o:
objs/$(CONFIG)/test/core/util/port_posix.o:
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/util/parse_hexstring.o:
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/util/slice_splitter.o:
Initial import.
10 years ago
LIBGRPC++_SRC = \
src/cpp/client/channel.cc \
Client side implementation of creating channel with credentials. The old test_ssl_channel related code is deleted and the new create channel call is used for interop tests. Change on 2014/12/19 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82540921
10 years ago
src/cpp/client/channel_arguments.cc \
Initial import.
10 years ago
src/cpp/client/client_context.cc \
Remove endpoint/ directories. Fold endpoint interface into iomgr, move secure_endpoint into security/. This will make it easier for endpoint to rely on some iomgr defined types (like pollset). Change on 2014/12/17 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82338036
10 years ago
src/cpp/client/create_channel.cc \
Prepare QPS tests for running in GCE Edit build.json to include qps tests and credentials Perform required MOE edits for running qps tests in VM Minor changes to underlying source files to support issues like more restrictive warning flags in cloud. Change on 2014/12/17 by vpai <vpai@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82349829
10 years ago
src/cpp/client/credentials.cc \
Initial import.
10 years ago
src/cpp/client/internal_stub.cc \
src/cpp/proto/proto_utils.cc \
Fix gRPC compilation failure Update location of rpc_method.cc on the Makefile Change on 2015/01/12 by rsilvera <rsilvera@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83786942
10 years ago
src/cpp/common/rpc_method.cc \
Remove endpoint/ directories. Fold endpoint interface into iomgr, move secure_endpoint into security/. This will make it easier for endpoint to rely on some iomgr defined types (like pollset). Change on 2014/12/17 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82338036
10 years ago
src/cpp/server/async_server.cc \
src/cpp/server/async_server_context.cc \
src/cpp/server/completion_queue.cc \
src/cpp/server/server_builder.cc \
Sync up open source build with [] Change on 2014/12/17 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82377745
10 years ago
src/cpp/server/server_context_impl.cc \
Remove endpoint/ directories. Fold endpoint interface into iomgr, move secure_endpoint into security/. This will make it easier for endpoint to rely on some iomgr defined types (like pollset). Change on 2014/12/17 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82338036
10 years ago
src/cpp/server/server.cc \
src/cpp/server/server_rpc_handler.cc \
Prepare QPS tests for running in GCE Edit build.json to include qps tests and credentials Perform required MOE edits for running qps tests in VM Minor changes to underlying source files to support issues like more restrictive warning flags in cloud. Change on 2014/12/17 by vpai <vpai@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82349829
10 years ago
src/cpp/server/server_credentials.cc \
Remove endpoint/ directories. Fold endpoint interface into iomgr, move secure_endpoint into security/. This will make it easier for endpoint to rely on some iomgr defined types (like pollset). Change on 2014/12/17 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82338036
10 years ago
src/cpp/server/thread_pool.cc \
src/cpp/stream/stream_context.cc \
src/cpp/util/status.cc \
src/cpp/util/time.cc \
Initial import.
10 years ago
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
PUBLIC_HEADERS_CXX += \
Remove endpoint/ directories. Fold endpoint interface into iomgr, move secure_endpoint into security/. This will make it easier for endpoint to rely on some iomgr defined types (like pollset). Change on 2014/12/17 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82338036
10 years ago
include/grpc++/async_server_context.h \
Initial import.
10 years ago
include/grpc++/async_server.h \
Client side implementation of creating channel with credentials. The old test_ssl_channel related code is deleted and the new create channel call is used for interop tests. Change on 2014/12/19 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82540921
10 years ago
include/grpc++/channel_arguments.h \
Remove endpoint/ directories. Fold endpoint interface into iomgr, move secure_endpoint into security/. This will make it easier for endpoint to rely on some iomgr defined types (like pollset). Change on 2014/12/17 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82338036
10 years ago
include/grpc++/channel_interface.h \
include/grpc++/client_context.h \
include/grpc++/completion_queue.h \
include/grpc++/config.h \
Initial import.
10 years ago
include/grpc++/create_channel.h \
Prepare QPS tests for running in GCE Edit build.json to include qps tests and credentials Perform required MOE edits for running qps tests in VM Minor changes to underlying source files to support issues like more restrictive warning flags in cloud. Change on 2014/12/17 by vpai <vpai@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82349829
10 years ago
include/grpc++/credentials.h \
Add impl subdirectory under public and move headers that need to be installed to it. Change on 2015/01/09 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83640373
10 years ago
include/grpc++/impl/internal_stub.h \
include/grpc++/impl/rpc_method.h \
include/grpc++/impl/rpc_service_method.h \
Initial import.
10 years ago
include/grpc++/server_builder.h \
Sync up open source build with [] Change on 2014/12/17 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82377745
10 years ago
include/grpc++/server_context.h \
Prepare QPS tests for running in GCE Edit build.json to include qps tests and credentials Perform required MOE edits for running qps tests in VM Minor changes to underlying source files to support issues like more restrictive warning flags in cloud. Change on 2014/12/17 by vpai <vpai@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82349829
10 years ago
include/grpc++/server_credentials.h \
Remove endpoint/ directories. Fold endpoint interface into iomgr, move secure_endpoint into security/. This will make it easier for endpoint to rely on some iomgr defined types (like pollset). Change on 2014/12/17 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82338036
10 years ago
include/grpc++/server.h \
Initial import.
10 years ago
include/grpc++/status.h \
Remove endpoint/ directories. Fold endpoint interface into iomgr, move secure_endpoint into security/. This will make it easier for endpoint to rely on some iomgr defined types (like pollset). Change on 2014/12/17 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82338036
10 years ago
include/grpc++/stream_context_interface.h \
Initial import.
10 years ago
include/grpc++/stream.h \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBGRPC++_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBGRPC++_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure libraries if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
libs/$(CONFIG)/libgrpc++.a: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
ifeq ($(SYSTEM),MINGW32)
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
libs/$(CONFIG)/grpc++.$(SHARED_EXT): openssl_dep_error
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
else
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
libs/$(CONFIG)/libgrpc++.$(SHARED_EXT): openssl_dep_error
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Fixing parallel build. The source code of the libraries depends on openssl being built, as it will setup the headers. Since we don't know exactly which source depends on openssl (we could start flagging it in build.json, but that'd be painful), we just flag all the source for every ssl-enabled library to depend on openssl being built.
10 years ago
ifneq ($(OPENSSL_DEP),)
src/cpp/client/channel.cc: $(OPENSSL_DEP)
src/cpp/client/channel_arguments.cc: $(OPENSSL_DEP)
src/cpp/client/client_context.cc: $(OPENSSL_DEP)
src/cpp/client/create_channel.cc: $(OPENSSL_DEP)
src/cpp/client/credentials.cc: $(OPENSSL_DEP)
src/cpp/client/internal_stub.cc: $(OPENSSL_DEP)
src/cpp/proto/proto_utils.cc: $(OPENSSL_DEP)
src/cpp/common/rpc_method.cc: $(OPENSSL_DEP)
src/cpp/server/async_server.cc: $(OPENSSL_DEP)
src/cpp/server/async_server_context.cc: $(OPENSSL_DEP)
src/cpp/server/completion_queue.cc: $(OPENSSL_DEP)
src/cpp/server/server_builder.cc: $(OPENSSL_DEP)
src/cpp/server/server_context_impl.cc: $(OPENSSL_DEP)
src/cpp/server/server.cc: $(OPENSSL_DEP)
src/cpp/server/server_rpc_handler.cc: $(OPENSSL_DEP)
src/cpp/server/server_credentials.cc: $(OPENSSL_DEP)
src/cpp/server/thread_pool.cc: $(OPENSSL_DEP)
src/cpp/stream/stream_context.cc: $(OPENSSL_DEP)
src/cpp/util/status.cc: $(OPENSSL_DEP)
src/cpp/util/time.cc: $(OPENSSL_DEP)
endif
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libgrpc++.a: $(ZLIB_DEP) $(OPENSSL_DEP) $(LIBGRPC++_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libgrpc++.a $(LIBGRPC++_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
ifeq ($(SYSTEM),MINGW32)
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/grpc++.$(SHARED_EXT): $(LIBGRPC++_OBJS) $(ZLIB_DEP)libs/$(CONFIG)/grpc.$(SHARED_EXT) $(OPENSSL_DEP)
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(LDXX) $(LDFLAGS) -Llibs/$(CONFIG) -shared -Wl,--output-def=libs/$(CONFIG)/grpc++.def -Wl,--out-implib=libs/$(CONFIG)/libgrpc++-imp.a -o libs/$(CONFIG)/grpc++.$(SHARED_EXT) $(LIBGRPC++_OBJS) $(LDLIBS) $(LDLIBS_SECURE) $(OPENSSL_MERGE_LIBS) -lgrpc-imp
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
else
Add missing space
10 years ago
libs/$(CONFIG)/libgrpc++.$(SHARED_EXT): $(LIBGRPC++_OBJS) $(ZLIB_DEP) libs/$(CONFIG)/libgrpc.$(SHARED_EXT) $(OPENSSL_DEP)
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
ifeq ($(SYSTEM),Darwin)
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(LDXX) $(LDFLAGS) -Llibs/$(CONFIG) -dynamiclib -o libs/$(CONFIG)/libgrpc++.$(SHARED_EXT) $(LIBGRPC++_OBJS) $(LDLIBS) $(LDLIBS_SECURE) $(OPENSSL_MERGE_LIBS) -lgrpc
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
else
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(LDXX) $(LDFLAGS) -Llibs/$(CONFIG) -shared -Wl,-soname,libgrpc++.so.0 -o libs/$(CONFIG)/libgrpc++.$(SHARED_EXT) $(LIBGRPC++_OBJS) $(LDLIBS) $(LDLIBS_SECURE) $(OPENSSL_MERGE_LIBS) -lgrpc
$(Q) ln -sf libgrpc++.$(SHARED_EXT) libs/$(CONFIG)/libgrpc++.so
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
endif
endif
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBGRPC++_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/src/cpp/client/channel.o:
objs/$(CONFIG)/src/cpp/client/channel_arguments.o:
objs/$(CONFIG)/src/cpp/client/client_context.o:
objs/$(CONFIG)/src/cpp/client/create_channel.o:
objs/$(CONFIG)/src/cpp/client/credentials.o:
objs/$(CONFIG)/src/cpp/client/internal_stub.o:
objs/$(CONFIG)/src/cpp/proto/proto_utils.o:
objs/$(CONFIG)/src/cpp/common/rpc_method.o:
objs/$(CONFIG)/src/cpp/server/async_server.o:
objs/$(CONFIG)/src/cpp/server/async_server_context.o:
objs/$(CONFIG)/src/cpp/server/completion_queue.o:
objs/$(CONFIG)/src/cpp/server/server_builder.o:
objs/$(CONFIG)/src/cpp/server/server_context_impl.o:
objs/$(CONFIG)/src/cpp/server/server.o:
objs/$(CONFIG)/src/cpp/server/server_rpc_handler.o:
objs/$(CONFIG)/src/cpp/server/server_credentials.o:
objs/$(CONFIG)/src/cpp/server/thread_pool.o:
objs/$(CONFIG)/src/cpp/stream/stream_context.o:
objs/$(CONFIG)/src/cpp/util/status.o:
objs/$(CONFIG)/src/cpp/util/time.o:
Initial import.
10 years ago
LIBGRPC++_TEST_UTIL_SRC = \
Add a test to export two TestService on the same server, only with different package names. Make sure they work as intended. This is based on [] Change on 2015/01/08 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83558743
10 years ago
gens/test/cpp/util/messages.pb.cc \
Remove endpoint/ directories. Fold endpoint interface into iomgr, move secure_endpoint into security/. This will make it easier for endpoint to rely on some iomgr defined types (like pollset). Change on 2014/12/17 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82338036
10 years ago
gens/test/cpp/util/echo.pb.cc \
Add a test to export two TestService on the same server, only with different package names. Make sure they work as intended. This is based on [] Change on 2015/01/08 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83558743
10 years ago
gens/test/cpp/util/echo_duplicate.pb.cc \
Client side implementation of creating channel with credentials. The old test_ssl_channel related code is deleted and the new create channel call is used for interop tests. Change on 2014/12/19 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82540921
10 years ago
test/cpp/util/create_test_channel.cc \
Changing compilation ordering to fix dependency issue between generated files and their inclusion. Change on 2014/12/17 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82363995
10 years ago
test/cpp/end2end/async_test_server.cc \
Initial import.
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBGRPC++_TEST_UTIL_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBGRPC++_TEST_UTIL_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure libraries if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
libs/$(CONFIG)/libgrpc++_test_util.a: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Fixing parallel build. The source code of the libraries depends on openssl being built, as it will setup the headers. Since we don't know exactly which source depends on openssl (we could start flagging it in build.json, but that'd be painful), we just flag all the source for every ssl-enabled library to depend on openssl being built.
10 years ago
ifneq ($(OPENSSL_DEP),)
test/cpp/util/messages.proto: $(OPENSSL_DEP)
test/cpp/util/echo.proto: $(OPENSSL_DEP)
test/cpp/util/echo_duplicate.proto: $(OPENSSL_DEP)
test/cpp/util/create_test_channel.cc: $(OPENSSL_DEP)
test/cpp/end2end/async_test_server.cc: $(OPENSSL_DEP)
endif
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libgrpc++_test_util.a: $(ZLIB_DEP) $(OPENSSL_DEP) $(LIBGRPC++_TEST_UTIL_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libgrpc++_test_util.a $(LIBGRPC++_TEST_UTIL_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBGRPC++_TEST_UTIL_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/cpp/util/create_test_channel.o: gens/test/cpp/util/messages.pb.cc gens/test/cpp/util/echo.pb.cc gens/test/cpp/util/echo_duplicate.pb.cc
objs/$(CONFIG)/test/cpp/end2end/async_test_server.o: gens/test/cpp/util/messages.pb.cc gens/test/cpp/util/echo.pb.cc gens/test/cpp/util/echo_duplicate.pb.cc
Initial import.
10 years ago
TIPS client implementation.
10 years ago
LIBTIPS_CLIENT_LIB_SRC = \
gens/examples/tips/label.pb.cc \
gens/examples/tips/empty.pb.cc \
gens/examples/tips/pubsub.pb.cc \
examples/tips/client.cc \
LIBTIPS_CLIENT_LIB_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBTIPS_CLIENT_LIB_SRC))))
ifeq ($(NO_SECURE),true)
# You can't build secure libraries if you don't have OpenSSL with ALPN.
libs/$(CONFIG)/libtips_client_lib.a: openssl_dep_error
else
ifneq ($(OPENSSL_DEP),)
examples/tips/label.proto: $(OPENSSL_DEP)
examples/tips/empty.proto: $(OPENSSL_DEP)
examples/tips/pubsub.proto: $(OPENSSL_DEP)
examples/tips/client.cc: $(OPENSSL_DEP)
endif
libs/$(CONFIG)/libtips_client_lib.a: $(ZLIB_DEP) $(OPENSSL_DEP) $(LIBTIPS_CLIENT_LIB_OBJS)
$(E) "[AR] Creating $@"
$(Q) mkdir -p `dirname $@`
$(Q) $(AR) rcs libs/$(CONFIG)/libtips_client_lib.a $(LIBTIPS_CLIENT_LIB_OBJS)
endif
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
-include $(LIBTIPS_CLIENT_LIB_OBJS:.o=.dep)
endif
endif
objs/$(CONFIG)/examples/tips/client.o: gens/examples/tips/label.pb.cc gens/examples/tips/empty.pb.cc gens/examples/tips/pubsub.pb.cc
Initial import.
10 years ago
LIBEND2END_FIXTURE_CHTTP2_FAKE_SECURITY_SRC = \
test/core/end2end/fixtures/chttp2_fake_security.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_FIXTURE_CHTTP2_FAKE_SECURITY_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_FIXTURE_CHTTP2_FAKE_SECURITY_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure libraries if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Fixing parallel build. The source code of the libraries depends on openssl being built, as it will setup the headers. Since we don't know exactly which source depends on openssl (we could start flagging it in build.json, but that'd be painful), we just flag all the source for every ssl-enabled library to depend on openssl being built.
10 years ago
ifneq ($(OPENSSL_DEP),)
test/core/end2end/fixtures/chttp2_fake_security.c: $(OPENSSL_DEP)
endif
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a: $(ZLIB_DEP) $(OPENSSL_DEP) $(LIBEND2END_FIXTURE_CHTTP2_FAKE_SECURITY_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a $(LIBEND2END_FIXTURE_CHTTP2_FAKE_SECURITY_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_FIXTURE_CHTTP2_FAKE_SECURITY_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/fixtures/chttp2_fake_security.o:
Initial import.
10 years ago
LIBEND2END_FIXTURE_CHTTP2_FULLSTACK_SRC = \
test/core/end2end/fixtures/chttp2_fullstack.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_FIXTURE_CHTTP2_FULLSTACK_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_FIXTURE_CHTTP2_FULLSTACK_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure libraries if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Fixing parallel build. The source code of the libraries depends on openssl being built, as it will setup the headers. Since we don't know exactly which source depends on openssl (we could start flagging it in build.json, but that'd be painful), we just flag all the source for every ssl-enabled library to depend on openssl being built.
10 years ago
ifneq ($(OPENSSL_DEP),)
test/core/end2end/fixtures/chttp2_fullstack.c: $(OPENSSL_DEP)
endif
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a: $(ZLIB_DEP) $(OPENSSL_DEP) $(LIBEND2END_FIXTURE_CHTTP2_FULLSTACK_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a $(LIBEND2END_FIXTURE_CHTTP2_FULLSTACK_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_FIXTURE_CHTTP2_FULLSTACK_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/fixtures/chttp2_fullstack.o:
Initial import.
10 years ago
LIBEND2END_FIXTURE_CHTTP2_SIMPLE_SSL_FULLSTACK_SRC = \
test/core/end2end/fixtures/chttp2_simple_ssl_fullstack.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_FIXTURE_CHTTP2_SIMPLE_SSL_FULLSTACK_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_FIXTURE_CHTTP2_SIMPLE_SSL_FULLSTACK_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure libraries if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Fixing parallel build. The source code of the libraries depends on openssl being built, as it will setup the headers. Since we don't know exactly which source depends on openssl (we could start flagging it in build.json, but that'd be painful), we just flag all the source for every ssl-enabled library to depend on openssl being built.
10 years ago
ifneq ($(OPENSSL_DEP),)
test/core/end2end/fixtures/chttp2_simple_ssl_fullstack.c: $(OPENSSL_DEP)
endif
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a: $(ZLIB_DEP) $(OPENSSL_DEP) $(LIBEND2END_FIXTURE_CHTTP2_SIMPLE_SSL_FULLSTACK_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a $(LIBEND2END_FIXTURE_CHTTP2_SIMPLE_SSL_FULLSTACK_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_FIXTURE_CHTTP2_SIMPLE_SSL_FULLSTACK_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/fixtures/chttp2_simple_ssl_fullstack.o:
Initial import.
10 years ago
LIBEND2END_FIXTURE_CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_SRC = \
test/core/end2end/fixtures/chttp2_simple_ssl_with_oauth2_fullstack.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_FIXTURE_CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_FIXTURE_CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure libraries if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Fixing parallel build. The source code of the libraries depends on openssl being built, as it will setup the headers. Since we don't know exactly which source depends on openssl (we could start flagging it in build.json, but that'd be painful), we just flag all the source for every ssl-enabled library to depend on openssl being built.
10 years ago
ifneq ($(OPENSSL_DEP),)
test/core/end2end/fixtures/chttp2_simple_ssl_with_oauth2_fullstack.c: $(OPENSSL_DEP)
endif
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a: $(ZLIB_DEP) $(OPENSSL_DEP) $(LIBEND2END_FIXTURE_CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a $(LIBEND2END_FIXTURE_CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_FIXTURE_CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/fixtures/chttp2_simple_ssl_with_oauth2_fullstack.o:
Initial import.
10 years ago
LIBEND2END_FIXTURE_CHTTP2_SOCKET_PAIR_SRC = \
test/core/end2end/fixtures/chttp2_socket_pair.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_FIXTURE_CHTTP2_SOCKET_PAIR_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_FIXTURE_CHTTP2_SOCKET_PAIR_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure libraries if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Fixing parallel build. The source code of the libraries depends on openssl being built, as it will setup the headers. Since we don't know exactly which source depends on openssl (we could start flagging it in build.json, but that'd be painful), we just flag all the source for every ssl-enabled library to depend on openssl being built.
10 years ago
ifneq ($(OPENSSL_DEP),)
test/core/end2end/fixtures/chttp2_socket_pair.c: $(OPENSSL_DEP)
endif
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a: $(ZLIB_DEP) $(OPENSSL_DEP) $(LIBEND2END_FIXTURE_CHTTP2_SOCKET_PAIR_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a $(LIBEND2END_FIXTURE_CHTTP2_SOCKET_PAIR_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_FIXTURE_CHTTP2_SOCKET_PAIR_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/fixtures/chttp2_socket_pair.o:
Initial import.
10 years ago
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
LIBEND2END_FIXTURE_CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_SRC = \
test/core/end2end/fixtures/chttp2_socket_pair_one_byte_at_a_time.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_FIXTURE_CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_FIXTURE_CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure libraries if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Fixing parallel build. The source code of the libraries depends on openssl being built, as it will setup the headers. Since we don't know exactly which source depends on openssl (we could start flagging it in build.json, but that'd be painful), we just flag all the source for every ssl-enabled library to depend on openssl being built.
10 years ago
ifneq ($(OPENSSL_DEP),)
test/core/end2end/fixtures/chttp2_socket_pair_one_byte_at_a_time.c: $(OPENSSL_DEP)
endif
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a: $(ZLIB_DEP) $(OPENSSL_DEP) $(LIBEND2END_FIXTURE_CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_OBJS)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a $(LIBEND2END_FIXTURE_CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_OBJS)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_FIXTURE_CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/fixtures/chttp2_socket_pair_one_byte_at_a_time.o:
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Initial import.
10 years ago
LIBEND2END_TEST_CANCEL_AFTER_ACCEPT_SRC = \
test/core/end2end/tests/cancel_after_accept.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_TEST_CANCEL_AFTER_ACCEPT_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_CANCEL_AFTER_ACCEPT_SRC))))
Initial import.
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_test_cancel_after_accept.a: $(ZLIB_DEP) $(LIBEND2END_TEST_CANCEL_AFTER_ACCEPT_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_cancel_after_accept.a $(LIBEND2END_TEST_CANCEL_AFTER_ACCEPT_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_TEST_CANCEL_AFTER_ACCEPT_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/tests/cancel_after_accept.o:
Initial import.
10 years ago
LIBEND2END_TEST_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_SRC = \
test/core/end2end/tests/cancel_after_accept_and_writes_closed.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_TEST_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_SRC))))
Initial import.
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_test_cancel_after_accept_and_writes_closed.a: $(ZLIB_DEP) $(LIBEND2END_TEST_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_cancel_after_accept_and_writes_closed.a $(LIBEND2END_TEST_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_TEST_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/tests/cancel_after_accept_and_writes_closed.o:
Initial import.
10 years ago
LIBEND2END_TEST_CANCEL_AFTER_INVOKE_SRC = \
test/core/end2end/tests/cancel_after_invoke.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_TEST_CANCEL_AFTER_INVOKE_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_CANCEL_AFTER_INVOKE_SRC))))
Initial import.
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_test_cancel_after_invoke.a: $(ZLIB_DEP) $(LIBEND2END_TEST_CANCEL_AFTER_INVOKE_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_cancel_after_invoke.a $(LIBEND2END_TEST_CANCEL_AFTER_INVOKE_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_TEST_CANCEL_AFTER_INVOKE_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/tests/cancel_after_invoke.o:
Initial import.
10 years ago
LIBEND2END_TEST_CANCEL_BEFORE_INVOKE_SRC = \
test/core/end2end/tests/cancel_before_invoke.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_TEST_CANCEL_BEFORE_INVOKE_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_CANCEL_BEFORE_INVOKE_SRC))))
Initial import.
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_test_cancel_before_invoke.a: $(ZLIB_DEP) $(LIBEND2END_TEST_CANCEL_BEFORE_INVOKE_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_cancel_before_invoke.a $(LIBEND2END_TEST_CANCEL_BEFORE_INVOKE_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_TEST_CANCEL_BEFORE_INVOKE_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/tests/cancel_before_invoke.o:
Initial import.
10 years ago
LIBEND2END_TEST_CANCEL_IN_A_VACUUM_SRC = \
test/core/end2end/tests/cancel_in_a_vacuum.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_TEST_CANCEL_IN_A_VACUUM_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_CANCEL_IN_A_VACUUM_SRC))))
Initial import.
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_test_cancel_in_a_vacuum.a: $(ZLIB_DEP) $(LIBEND2END_TEST_CANCEL_IN_A_VACUUM_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_cancel_in_a_vacuum.a $(LIBEND2END_TEST_CANCEL_IN_A_VACUUM_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_TEST_CANCEL_IN_A_VACUUM_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/tests/cancel_in_a_vacuum.o:
Initial import.
10 years ago
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
LIBEND2END_TEST_CENSUS_SIMPLE_REQUEST_SRC = \
test/core/end2end/tests/census_simple_request.c \
LIBEND2END_TEST_CENSUS_SIMPLE_REQUEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_CENSUS_SIMPLE_REQUEST_SRC))))
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_test_census_simple_request.a: $(ZLIB_DEP) $(LIBEND2END_TEST_CENSUS_SIMPLE_REQUEST_OBJS)
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
$(E) "[AR] Creating $@"
$(Q) mkdir -p `dirname $@`
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_census_simple_request.a $(LIBEND2END_TEST_CENSUS_SIMPLE_REQUEST_OBJS)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_TEST_CENSUS_SIMPLE_REQUEST_OBJS:.o=.dep)
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/tests/census_simple_request.o:
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
LIBEND2END_TEST_DISAPPEARING_SERVER_SRC = \
test/core/end2end/tests/disappearing_server.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_TEST_DISAPPEARING_SERVER_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_DISAPPEARING_SERVER_SRC))))
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_test_disappearing_server.a: $(ZLIB_DEP) $(LIBEND2END_TEST_DISAPPEARING_SERVER_OBJS)
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
$(E) "[AR] Creating $@"
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_disappearing_server.a $(LIBEND2END_TEST_DISAPPEARING_SERVER_OBJS)
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_TEST_DISAPPEARING_SERVER_OBJS:.o=.dep)
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/tests/disappearing_server.o:
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
Initial import.
10 years ago
LIBEND2END_TEST_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_SRC = \
test/core/end2end/tests/early_server_shutdown_finishes_inflight_calls.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_TEST_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_SRC))))
Initial import.
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_inflight_calls.a: $(ZLIB_DEP) $(LIBEND2END_TEST_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_inflight_calls.a $(LIBEND2END_TEST_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_TEST_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/tests/early_server_shutdown_finishes_inflight_calls.o:
Initial import.
10 years ago
LIBEND2END_TEST_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_SRC = \
test/core/end2end/tests/early_server_shutdown_finishes_tags.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_TEST_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_SRC))))
Initial import.
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_tags.a: $(ZLIB_DEP) $(LIBEND2END_TEST_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_tags.a $(LIBEND2END_TEST_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_TEST_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/tests/early_server_shutdown_finishes_tags.o:
Initial import.
10 years ago
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
LIBEND2END_TEST_GRACEFUL_SERVER_SHUTDOWN_SRC = \
test/core/end2end/tests/graceful_server_shutdown.c \
LIBEND2END_TEST_GRACEFUL_SERVER_SHUTDOWN_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_GRACEFUL_SERVER_SHUTDOWN_SRC))))
libs/$(CONFIG)/libend2end_test_graceful_server_shutdown.a: $(ZLIB_DEP) $(LIBEND2END_TEST_GRACEFUL_SERVER_SHUTDOWN_OBJS)
$(E) "[AR] Creating $@"
$(Q) mkdir -p `dirname $@`
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_graceful_server_shutdown.a $(LIBEND2END_TEST_GRACEFUL_SERVER_SHUTDOWN_OBJS)
ifneq ($(NO_DEPS),true)
-include $(LIBEND2END_TEST_GRACEFUL_SERVER_SHUTDOWN_OBJS:.o=.dep)
endif
objs/$(CONFIG)/test/core/end2end/tests/graceful_server_shutdown.o:
Initial import.
10 years ago
LIBEND2END_TEST_INVOKE_LARGE_REQUEST_SRC = \
test/core/end2end/tests/invoke_large_request.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_TEST_INVOKE_LARGE_REQUEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_INVOKE_LARGE_REQUEST_SRC))))
Initial import.
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_test_invoke_large_request.a: $(ZLIB_DEP) $(LIBEND2END_TEST_INVOKE_LARGE_REQUEST_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_invoke_large_request.a $(LIBEND2END_TEST_INVOKE_LARGE_REQUEST_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_TEST_INVOKE_LARGE_REQUEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/tests/invoke_large_request.o:
Initial import.
10 years ago
LIBEND2END_TEST_MAX_CONCURRENT_STREAMS_SRC = \
test/core/end2end/tests/max_concurrent_streams.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_TEST_MAX_CONCURRENT_STREAMS_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_MAX_CONCURRENT_STREAMS_SRC))))
Initial import.
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_test_max_concurrent_streams.a: $(ZLIB_DEP) $(LIBEND2END_TEST_MAX_CONCURRENT_STREAMS_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_max_concurrent_streams.a $(LIBEND2END_TEST_MAX_CONCURRENT_STREAMS_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_TEST_MAX_CONCURRENT_STREAMS_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/tests/max_concurrent_streams.o:
Initial import.
10 years ago
LIBEND2END_TEST_NO_OP_SRC = \
test/core/end2end/tests/no_op.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_TEST_NO_OP_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_NO_OP_SRC))))
Initial import.
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_test_no_op.a: $(ZLIB_DEP) $(LIBEND2END_TEST_NO_OP_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_no_op.a $(LIBEND2END_TEST_NO_OP_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_TEST_NO_OP_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/tests/no_op.o:
Initial import.
10 years ago
LIBEND2END_TEST_PING_PONG_STREAMING_SRC = \
test/core/end2end/tests/ping_pong_streaming.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_TEST_PING_PONG_STREAMING_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_PING_PONG_STREAMING_SRC))))
Initial import.
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_test_ping_pong_streaming.a: $(ZLIB_DEP) $(LIBEND2END_TEST_PING_PONG_STREAMING_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_ping_pong_streaming.a $(LIBEND2END_TEST_PING_PONG_STREAMING_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_TEST_PING_PONG_STREAMING_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/tests/ping_pong_streaming.o:
Initial import.
10 years ago
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
LIBEND2END_TEST_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_SRC = \
test/core/end2end/tests/request_response_with_binary_metadata_and_payload.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_TEST_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_SRC))))
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_test_request_response_with_binary_metadata_and_payload.a: $(ZLIB_DEP) $(LIBEND2END_TEST_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_OBJS)
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
$(E) "[AR] Creating $@"
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_request_response_with_binary_metadata_and_payload.a $(LIBEND2END_TEST_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_OBJS)
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_TEST_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_OBJS:.o=.dep)
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/tests/request_response_with_binary_metadata_and_payload.o:
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
Initial import.
10 years ago
LIBEND2END_TEST_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_SRC = \
test/core/end2end/tests/request_response_with_metadata_and_payload.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_TEST_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_SRC))))
Initial import.
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_test_request_response_with_metadata_and_payload.a: $(ZLIB_DEP) $(LIBEND2END_TEST_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_request_response_with_metadata_and_payload.a $(LIBEND2END_TEST_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_TEST_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/tests/request_response_with_metadata_and_payload.o:
Initial import.
10 years ago
LIBEND2END_TEST_REQUEST_RESPONSE_WITH_PAYLOAD_SRC = \
test/core/end2end/tests/request_response_with_payload.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_TEST_REQUEST_RESPONSE_WITH_PAYLOAD_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_REQUEST_RESPONSE_WITH_PAYLOAD_SRC))))
Initial import.
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_test_request_response_with_payload.a: $(ZLIB_DEP) $(LIBEND2END_TEST_REQUEST_RESPONSE_WITH_PAYLOAD_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_request_response_with_payload.a $(LIBEND2END_TEST_REQUEST_RESPONSE_WITH_PAYLOAD_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_TEST_REQUEST_RESPONSE_WITH_PAYLOAD_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/tests/request_response_with_payload.o:
Initial import.
10 years ago
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
LIBEND2END_TEST_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_SRC = \
test/core/end2end/tests/request_response_with_trailing_metadata_and_payload.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_TEST_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_SRC))))
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_test_request_response_with_trailing_metadata_and_payload.a: $(ZLIB_DEP) $(LIBEND2END_TEST_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_OBJS)
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
$(E) "[AR] Creating $@"
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_request_response_with_trailing_metadata_and_payload.a $(LIBEND2END_TEST_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_OBJS)
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_TEST_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_OBJS:.o=.dep)
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/tests/request_response_with_trailing_metadata_and_payload.o:
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
Initial import.
10 years ago
LIBEND2END_TEST_SIMPLE_DELAYED_REQUEST_SRC = \
test/core/end2end/tests/simple_delayed_request.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_TEST_SIMPLE_DELAYED_REQUEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_SIMPLE_DELAYED_REQUEST_SRC))))
Initial import.
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_test_simple_delayed_request.a: $(ZLIB_DEP) $(LIBEND2END_TEST_SIMPLE_DELAYED_REQUEST_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_simple_delayed_request.a $(LIBEND2END_TEST_SIMPLE_DELAYED_REQUEST_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_TEST_SIMPLE_DELAYED_REQUEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/tests/simple_delayed_request.o:
Initial import.
10 years ago
LIBEND2END_TEST_SIMPLE_REQUEST_SRC = \
test/core/end2end/tests/simple_request.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_TEST_SIMPLE_REQUEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_SIMPLE_REQUEST_SRC))))
Initial import.
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_test_simple_request.a: $(ZLIB_DEP) $(LIBEND2END_TEST_SIMPLE_REQUEST_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_simple_request.a $(LIBEND2END_TEST_SIMPLE_REQUEST_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_TEST_SIMPLE_REQUEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/tests/simple_request.o:
Initial import.
10 years ago
Strip "_test" from the name of thread_stress_test.c. Having a BUILD target named chttp2_fullstack_thread_stress_test_test is deceptive as it is a direct test of the system rather than a test of a test of the system. Change on 2014/12/09 by nathaniel <nathaniel@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81688649
10 years ago
LIBEND2END_TEST_THREAD_STRESS_SRC = \
test/core/end2end/tests/thread_stress.c \
Initial import.
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_TEST_THREAD_STRESS_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_THREAD_STRESS_SRC))))
Initial import.
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_test_thread_stress.a: $(ZLIB_DEP) $(LIBEND2END_TEST_THREAD_STRESS_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_thread_stress.a $(LIBEND2END_TEST_THREAD_STRESS_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_TEST_THREAD_STRESS_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/tests/thread_stress.o:
Initial import.
10 years ago
LIBEND2END_TEST_WRITES_DONE_HANGS_WITH_PENDING_READ_SRC = \
test/core/end2end/tests/writes_done_hangs_with_pending_read.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_TEST_WRITES_DONE_HANGS_WITH_PENDING_READ_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_TEST_WRITES_DONE_HANGS_WITH_PENDING_READ_SRC))))
Initial import.
10 years ago
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_test_writes_done_hangs_with_pending_read.a: $(ZLIB_DEP) $(LIBEND2END_TEST_WRITES_DONE_HANGS_WITH_PENDING_READ_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_test_writes_done_hangs_with_pending_read.a $(LIBEND2END_TEST_WRITES_DONE_HANGS_WITH_PENDING_READ_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_TEST_WRITES_DONE_HANGS_WITH_PENDING_READ_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/tests/writes_done_hangs_with_pending_read.o:
Initial import.
10 years ago
LIBEND2END_CERTS_SRC = \
Add SSL root for production GFE. Change on 2014/12/19 by chenw <chenw@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82554526
10 years ago
test/core/end2end/data/test_root_cert.c \
test/core/end2end/data/prod_roots_certs.c \
Initial import.
10 years ago
test/core/end2end/data/server1_cert.c \
test/core/end2end/data/server1_key.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LIBEND2END_CERTS_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LIBEND2END_CERTS_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure libraries if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
libs/$(CONFIG)/libend2end_certs.a: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Fixing parallel build. The source code of the libraries depends on openssl being built, as it will setup the headers. Since we don't know exactly which source depends on openssl (we could start flagging it in build.json, but that'd be painful), we just flag all the source for every ssl-enabled library to depend on openssl being built.
10 years ago
ifneq ($(OPENSSL_DEP),)
test/core/end2end/data/test_root_cert.c: $(OPENSSL_DEP)
test/core/end2end/data/prod_roots_certs.c: $(OPENSSL_DEP)
test/core/end2end/data/server1_cert.c: $(OPENSSL_DEP)
test/core/end2end/data/server1_key.c: $(OPENSSL_DEP)
endif
Fixing zlib dependency build.
10 years ago
libs/$(CONFIG)/libend2end_certs.a: $(ZLIB_DEP) $(OPENSSL_DEP) $(LIBEND2END_CERTS_OBJS)
Initial import.
10 years ago
$(E) "[AR] Creating $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(AR) rcs libs/$(CONFIG)/libend2end_certs.a $(LIBEND2END_CERTS_OBJS)
Initial import.
10 years ago
Improving Makefile support for shared libraries. -) Properly linking built-in OpenSSL into the grpc library. -) grpc now properly depends on gpr when linking shared code. -) Properly naming the shared library with all their aliases. -) Properly installing the shared library aliases on the system. -) Potentially supporting Darwin and MINGW32 targets for shared libraries. Caveat: if using shared libraries, some tests will not compile anymore if they want to use OpenSSL functions, as they are no longer publically available externally. Which is the feature we were seeking. The Makefile currently does it properly, by linking the tests statically. This only applies when using external code and Makefiles, if said external code improperly assumes SSL is available through grpc. Change on 2014/12/22 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82656438
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LIBEND2END_CERTS_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Allow C++ tests to be built in parallel
10 years ago
objs/$(CONFIG)/test/core/end2end/data/test_root_cert.o:
objs/$(CONFIG)/test/core/end2end/data/prod_roots_certs.o:
objs/$(CONFIG)/test/core/end2end/data/server1_cert.o:
objs/$(CONFIG)/test/core/end2end/data/server1_key.o:
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
# All of the test targets, and protoc plugins
Initial import.
10 years ago
GEN_HPACK_TABLES_SRC = \
src/core/transport/chttp2/gen_hpack_tables.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GEN_HPACK_TABLES_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GEN_HPACK_TABLES_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/gen_hpack_tables: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/gen_hpack_tables: $(GEN_HPACK_TABLES_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgpr.a libs/$(CONFIG)/libgrpc.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GEN_HPACK_TABLES_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgpr.a libs/$(CONFIG)/libgrpc.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/gen_hpack_tables
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
objs/$(CONFIG)/src/core/transport/chttp2/gen_hpack_tables.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgpr.a libs/$(CONFIG)/libgrpc.a
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_gen_hpack_tables: $(GEN_HPACK_TABLES_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GEN_HPACK_TABLES_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
First pass at making the compilers open-sourcable. Change on 2014/12/10 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81829127
10 years ago
CPP_PLUGIN_SRC = \
Fixing typos in build.json: .cpp -> .cc This was working so far because of GNU make's magic that automatically finds the sources depending on its output, and not on its input. This also explains why running "make" would always rebuild the plugins: because GNU make think some dependencies are phony and aren't there.
10 years ago
src/compiler/cpp_plugin.cc \
src/compiler/cpp_generator.cc \
First pass at making the compilers open-sourcable. Change on 2014/12/10 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81829127
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CPP_PLUGIN_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CPP_PLUGIN_SRC))))
First pass at making the compilers open-sourcable. Change on 2014/12/10 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81829127
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/cpp_plugin: $(CPP_PLUGIN_OBJS)
Getting the generated Makefile to support .proto files. -) Compiling the proto plugin using the HOSTCC compiler. -) Set up proper rules to invoke the plugin from protoc. -) Few various renaming hacks to fully get out of []. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82000361
10 years ago
$(E) "[HOSTLD] Linking $@"
First pass at making the compilers open-sourcable. Change on 2014/12/10 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81829127
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(HOST_LDXX) $(HOST_LDFLAGS) $(CPP_PLUGIN_OBJS) $(HOST_LDLIBSXX) $(HOST_LDLIBS) $(HOST_LDLIBS_PROTOC) -o bins/$(CONFIG)/cpp_plugin
First pass at making the compilers open-sourcable. Change on 2014/12/10 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81829127
10 years ago
Fix some C++ test dependencies
10 years ago
objs/$(CONFIG)/src/compiler/cpp_plugin.o:
objs/$(CONFIG)/src/compiler/cpp_generator.o:
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_cpp_plugin: $(CPP_PLUGIN_OBJS:.o=.dep)
First pass at making the compilers open-sourcable. Change on 2014/12/10 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81829127
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CPP_PLUGIN_OBJS:.o=.dep)
First pass at making the compilers open-sourcable. Change on 2014/12/10 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81829127
10 years ago
endif
RUBY_PLUGIN_SRC = \
Fixing typos in build.json: .cpp -> .cc This was working so far because of GNU make's magic that automatically finds the sources depending on its output, and not on its input. This also explains why running "make" would always rebuild the plugins: because GNU make think some dependencies are phony and aren't there.
10 years ago
src/compiler/ruby_plugin.cc \
src/compiler/ruby_generator.cc \
First pass at making the compilers open-sourcable. Change on 2014/12/10 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81829127
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
RUBY_PLUGIN_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(RUBY_PLUGIN_SRC))))
First pass at making the compilers open-sourcable. Change on 2014/12/10 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81829127
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/ruby_plugin: $(RUBY_PLUGIN_OBJS)
Getting the generated Makefile to support .proto files. -) Compiling the proto plugin using the HOSTCC compiler. -) Set up proper rules to invoke the plugin from protoc. -) Few various renaming hacks to fully get out of []. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82000361
10 years ago
$(E) "[HOSTLD] Linking $@"
First pass at making the compilers open-sourcable. Change on 2014/12/10 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81829127
10 years ago
$(Q) mkdir -p `dirname $@`
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
$(Q) $(HOST_LDXX) $(HOST_LDFLAGS) $(RUBY_PLUGIN_OBJS) $(HOST_LDLIBSXX) $(HOST_LDLIBS) $(HOST_LDLIBS_PROTOC) -o bins/$(CONFIG)/ruby_plugin
First pass at making the compilers open-sourcable. Change on 2014/12/10 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81829127
10 years ago
Fix some C++ test dependencies
10 years ago
objs/$(CONFIG)/src/compiler/ruby_plugin.o:
objs/$(CONFIG)/src/compiler/ruby_generator.o:
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_ruby_plugin: $(RUBY_PLUGIN_OBJS:.o=.dep)
First pass at making the compilers open-sourcable. Change on 2014/12/10 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81829127
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(RUBY_PLUGIN_OBJS:.o=.dep)
First pass at making the compilers open-sourcable. Change on 2014/12/10 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81829127
10 years ago
endif
Initial import.
10 years ago
GRPC_BYTE_BUFFER_READER_TEST_SRC = \
test/core/surface/byte_buffer_reader_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GRPC_BYTE_BUFFER_READER_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GRPC_BYTE_BUFFER_READER_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/grpc_byte_buffer_reader_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/grpc_byte_buffer_reader_test: $(GRPC_BYTE_BUFFER_READER_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GRPC_BYTE_BUFFER_READER_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/grpc_byte_buffer_reader_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/surface/byte_buffer_reader_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_grpc_byte_buffer_reader_test: $(GRPC_BYTE_BUFFER_READER_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GRPC_BYTE_BUFFER_READER_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
GPR_CANCELLABLE_TEST_SRC = \
test/core/support/cancellable_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GPR_CANCELLABLE_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GPR_CANCELLABLE_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/gpr_cancellable_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/gpr_cancellable_test: $(GPR_CANCELLABLE_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GPR_CANCELLABLE_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/gpr_cancellable_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/support/cancellable_test.o: libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_gpr_cancellable_test: $(GPR_CANCELLABLE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GPR_CANCELLABLE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
GPR_LOG_TEST_SRC = \
test/core/support/log_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GPR_LOG_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GPR_LOG_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/gpr_log_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/gpr_log_test: $(GPR_LOG_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GPR_LOG_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/gpr_log_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/support/log_test.o: libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_gpr_log_test: $(GPR_LOG_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GPR_LOG_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Add a test for useful.h. Additionally add rotl, rotr which will be needed to implement some hashing functions shortly. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82149559
10 years ago
GPR_USEFUL_TEST_SRC = \
test/core/support/useful_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GPR_USEFUL_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GPR_USEFUL_TEST_SRC))))
Add a test for useful.h. Additionally add rotl, rotr which will be needed to implement some hashing functions shortly. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82149559
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/gpr_useful_test: openssl_dep_error
Add a test for useful.h. Additionally add rotl, rotr which will be needed to implement some hashing functions shortly. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82149559
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/gpr_useful_test: $(GPR_USEFUL_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Add a test for useful.h. Additionally add rotl, rotr which will be needed to implement some hashing functions shortly. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82149559
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GPR_USEFUL_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/gpr_useful_test
Add a test for useful.h. Additionally add rotl, rotr which will be needed to implement some hashing functions shortly. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82149559
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/support/useful_test.o: libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_gpr_useful_test: $(GPR_USEFUL_TEST_OBJS:.o=.dep)
Add a test for useful.h. Additionally add rotl, rotr which will be needed to implement some hashing functions shortly. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82149559
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GPR_USEFUL_TEST_OBJS:.o=.dep)
Add a test for useful.h. Additionally add rotl, rotr which will be needed to implement some hashing functions shortly. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82149559
10 years ago
endif
endif
Initial import.
10 years ago
GPR_CMDLINE_TEST_SRC = \
test/core/support/cmdline_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GPR_CMDLINE_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GPR_CMDLINE_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/gpr_cmdline_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/gpr_cmdline_test: $(GPR_CMDLINE_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GPR_CMDLINE_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/gpr_cmdline_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/support/cmdline_test.o: libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_gpr_cmdline_test: $(GPR_CMDLINE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GPR_CMDLINE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
GPR_HISTOGRAM_TEST_SRC = \
test/core/support/histogram_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GPR_HISTOGRAM_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GPR_HISTOGRAM_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/gpr_histogram_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/gpr_histogram_test: $(GPR_HISTOGRAM_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GPR_HISTOGRAM_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/gpr_histogram_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/support/histogram_test.o: libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_gpr_histogram_test: $(GPR_HISTOGRAM_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GPR_HISTOGRAM_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
GPR_HOST_PORT_TEST_SRC = \
test/core/support/host_port_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GPR_HOST_PORT_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GPR_HOST_PORT_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/gpr_host_port_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/gpr_host_port_test: $(GPR_HOST_PORT_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GPR_HOST_PORT_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/gpr_host_port_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/support/host_port_test.o: libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_gpr_host_port_test: $(GPR_HOST_PORT_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GPR_HOST_PORT_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
GPR_SLICE_BUFFER_TEST_SRC = \
test/core/support/slice_buffer_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GPR_SLICE_BUFFER_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GPR_SLICE_BUFFER_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/gpr_slice_buffer_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/gpr_slice_buffer_test: $(GPR_SLICE_BUFFER_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GPR_SLICE_BUFFER_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/gpr_slice_buffer_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/support/slice_buffer_test.o: libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_gpr_slice_buffer_test: $(GPR_SLICE_BUFFER_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GPR_SLICE_BUFFER_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
GPR_SLICE_TEST_SRC = \
test/core/support/slice_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GPR_SLICE_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GPR_SLICE_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/gpr_slice_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/gpr_slice_test: $(GPR_SLICE_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GPR_SLICE_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/gpr_slice_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/support/slice_test.o: libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_gpr_slice_test: $(GPR_SLICE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GPR_SLICE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
GPR_STRING_TEST_SRC = \
test/core/support/string_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GPR_STRING_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GPR_STRING_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/gpr_string_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/gpr_string_test: $(GPR_STRING_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GPR_STRING_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/gpr_string_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/support/string_test.o: libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_gpr_string_test: $(GPR_STRING_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GPR_STRING_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
GPR_SYNC_TEST_SRC = \
test/core/support/sync_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GPR_SYNC_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GPR_SYNC_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/gpr_sync_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/gpr_sync_test: $(GPR_SYNC_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GPR_SYNC_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/gpr_sync_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/support/sync_test.o: libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_gpr_sync_test: $(GPR_SYNC_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GPR_SYNC_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
GPR_THD_TEST_SRC = \
test/core/support/thd_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GPR_THD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GPR_THD_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/gpr_thd_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/gpr_thd_test: $(GPR_THD_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GPR_THD_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/gpr_thd_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/support/thd_test.o: libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_gpr_thd_test: $(GPR_THD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GPR_THD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
GPR_TIME_TEST_SRC = \
test/core/support/time_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GPR_TIME_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GPR_TIME_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/gpr_time_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/gpr_time_test: $(GPR_TIME_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GPR_TIME_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/gpr_time_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/support/time_test.o: libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_gpr_time_test: $(GPR_TIME_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GPR_TIME_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
MURMUR_HASH_TEST_SRC = \
test/core/support/murmur_hash_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
MURMUR_HASH_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(MURMUR_HASH_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/murmur_hash_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/murmur_hash_test: $(MURMUR_HASH_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(MURMUR_HASH_TEST_OBJS) libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/murmur_hash_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/support/murmur_hash_test.o: libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_murmur_hash_test: $(MURMUR_HASH_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(MURMUR_HASH_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
GRPC_STREAM_OP_TEST_SRC = \
test/core/transport/stream_op_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GRPC_STREAM_OP_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GRPC_STREAM_OP_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/grpc_stream_op_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/grpc_stream_op_test: $(GRPC_STREAM_OP_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GRPC_STREAM_OP_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/grpc_stream_op_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/transport/stream_op_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_grpc_stream_op_test: $(GRPC_STREAM_OP_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GRPC_STREAM_OP_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
ALPN_TEST_SRC = \
test/core/transport/chttp2/alpn_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
ALPN_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(ALPN_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/alpn_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/alpn_test: $(ALPN_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(ALPN_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/alpn_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/transport/chttp2/alpn_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_alpn_test: $(ALPN_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(ALPN_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Add time averaged stats for future alarm list work. This is a trivial C89-ification of the []2 implementation of the same idea. Indeed the implementation files and tests have been branched from the [] versions. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82168822
10 years ago
TIME_AVERAGED_STATS_TEST_SRC = \
test/core/iomgr/time_averaged_stats_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
TIME_AVERAGED_STATS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(TIME_AVERAGED_STATS_TEST_SRC))))
Add time averaged stats for future alarm list work. This is a trivial C89-ification of the []2 implementation of the same idea. Indeed the implementation files and tests have been branched from the [] versions. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82168822
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/time_averaged_stats_test: openssl_dep_error
Add time averaged stats for future alarm list work. This is a trivial C89-ification of the []2 implementation of the same idea. Indeed the implementation files and tests have been branched from the [] versions. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82168822
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/time_averaged_stats_test: $(TIME_AVERAGED_STATS_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Add time averaged stats for future alarm list work. This is a trivial C89-ification of the []2 implementation of the same idea. Indeed the implementation files and tests have been branched from the [] versions. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82168822
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(TIME_AVERAGED_STATS_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/time_averaged_stats_test
Add time averaged stats for future alarm list work. This is a trivial C89-ification of the []2 implementation of the same idea. Indeed the implementation files and tests have been branched from the [] versions. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82168822
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/iomgr/time_averaged_stats_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_time_averaged_stats_test: $(TIME_AVERAGED_STATS_TEST_OBJS:.o=.dep)
Add time averaged stats for future alarm list work. This is a trivial C89-ification of the []2 implementation of the same idea. Indeed the implementation files and tests have been branched from the [] versions. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82168822
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(TIME_AVERAGED_STATS_TEST_OBJS:.o=.dep)
Add time averaged stats for future alarm list work. This is a trivial C89-ification of the []2 implementation of the same idea. Indeed the implementation files and tests have been branched from the [] versions. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82168822
10 years ago
endif
endif
Initial import.
10 years ago
CHTTP2_STREAM_ENCODER_TEST_SRC = \
test/core/transport/chttp2/stream_encoder_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_STREAM_ENCODER_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_STREAM_ENCODER_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_stream_encoder_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_stream_encoder_test: $(CHTTP2_STREAM_ENCODER_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_STREAM_ENCODER_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_stream_encoder_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/transport/chttp2/stream_encoder_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_stream_encoder_test: $(CHTTP2_STREAM_ENCODER_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_STREAM_ENCODER_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
HPACK_TABLE_TEST_SRC = \
test/core/transport/chttp2/hpack_table_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
HPACK_TABLE_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(HPACK_TABLE_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/hpack_table_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/hpack_table_test: $(HPACK_TABLE_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(HPACK_TABLE_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/hpack_table_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/transport/chttp2/hpack_table_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_hpack_table_test: $(HPACK_TABLE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(HPACK_TABLE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_STREAM_MAP_TEST_SRC = \
test/core/transport/chttp2/stream_map_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_STREAM_MAP_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_STREAM_MAP_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_stream_map_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_stream_map_test: $(CHTTP2_STREAM_MAP_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_STREAM_MAP_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_stream_map_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/transport/chttp2/stream_map_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_stream_map_test: $(CHTTP2_STREAM_MAP_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_STREAM_MAP_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
HPACK_PARSER_TEST_SRC = \
test/core/transport/chttp2/hpack_parser_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
HPACK_PARSER_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(HPACK_PARSER_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/hpack_parser_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/hpack_parser_test: $(HPACK_PARSER_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(HPACK_PARSER_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/hpack_parser_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/transport/chttp2/hpack_parser_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_hpack_parser_test: $(HPACK_PARSER_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(HPACK_PARSER_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
TRANSPORT_METADATA_TEST_SRC = \
test/core/transport/metadata_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
TRANSPORT_METADATA_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(TRANSPORT_METADATA_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/transport_metadata_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/transport_metadata_test: $(TRANSPORT_METADATA_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(TRANSPORT_METADATA_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/transport_metadata_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/transport/metadata_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_transport_metadata_test: $(TRANSPORT_METADATA_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(TRANSPORT_METADATA_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_STATUS_CONVERSION_TEST_SRC = \
test/core/transport/chttp2/status_conversion_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_STATUS_CONVERSION_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_STATUS_CONVERSION_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_status_conversion_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_status_conversion_test: $(CHTTP2_STATUS_CONVERSION_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_STATUS_CONVERSION_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_status_conversion_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/transport/chttp2/status_conversion_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_status_conversion_test: $(CHTTP2_STATUS_CONVERSION_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_STATUS_CONVERSION_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_TRANSPORT_END2END_TEST_SRC = \
test/core/transport/chttp2_transport_end2end_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_TRANSPORT_END2END_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_TRANSPORT_END2END_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_transport_end2end_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_transport_end2end_test: $(CHTTP2_TRANSPORT_END2END_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_TRANSPORT_END2END_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_transport_end2end_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/transport/chttp2_transport_end2end_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_transport_end2end_test: $(CHTTP2_TRANSPORT_END2END_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_TRANSPORT_END2END_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Introducing iomgr. Move eventmanager and platform dependent endpoint functionality into a single library called 'iomgr'. This is primarily to prepare for a Windows port - where posix socket semantics lead to poor quality code. Mostly this is a code movement CL, with some small changes to help prepare the way for porting: - em style fd objects can only be held internally in iomgr, and own their memory - added grpc_iomgr_create_endpoint_pair() to accomodate the common pattern of creating a tcp endpoint from the output of socketpair - this will help keep our tests portable - separated em alarm interface into a separate file, as this part of event manager is needed higher up the stack - made the eventmanager bits a true singleton, simplifying API's across the stack as there's no longer a reason to carry a pointer there. Initial design document is here: https://docs.google.com/document/d/1VmafcHvvrP5kwtQkz84R5yXF7u7fW-9Pn0bkSUQHDt8/edit?disco=AAAAARNByxg Change on 2014/12/09 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81716456
10 years ago
TCP_POSIX_TEST_SRC = \
test/core/iomgr/tcp_posix_test.c \
Initial import.
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
TCP_POSIX_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(TCP_POSIX_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/tcp_posix_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/tcp_posix_test: $(TCP_POSIX_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(TCP_POSIX_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/tcp_posix_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/iomgr/tcp_posix_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_tcp_posix_test: $(TCP_POSIX_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(TCP_POSIX_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
DUALSTACK_SOCKET_TEST_SRC = \
test/core/end2end/dualstack_socket_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
DUALSTACK_SOCKET_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(DUALSTACK_SOCKET_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/dualstack_socket_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/dualstack_socket_test: $(DUALSTACK_SOCKET_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(DUALSTACK_SOCKET_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/dualstack_socket_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/end2end/dualstack_socket_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_dualstack_socket_test: $(DUALSTACK_SOCKET_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(DUALSTACK_SOCKET_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
NO_SERVER_TEST_SRC = \
test/core/end2end/no_server_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
NO_SERVER_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(NO_SERVER_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/no_server_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/no_server_test: $(NO_SERVER_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(NO_SERVER_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/no_server_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/end2end/no_server_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_no_server_test: $(NO_SERVER_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(NO_SERVER_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Initial import.
10 years ago
RESOLVE_ADDRESS_TEST_SRC = \
Introducing iomgr. Move eventmanager and platform dependent endpoint functionality into a single library called 'iomgr'. This is primarily to prepare for a Windows port - where posix socket semantics lead to poor quality code. Mostly this is a code movement CL, with some small changes to help prepare the way for porting: - em style fd objects can only be held internally in iomgr, and own their memory - added grpc_iomgr_create_endpoint_pair() to accomodate the common pattern of creating a tcp endpoint from the output of socketpair - this will help keep our tests portable - separated em alarm interface into a separate file, as this part of event manager is needed higher up the stack - made the eventmanager bits a true singleton, simplifying API's across the stack as there's no longer a reason to carry a pointer there. Initial design document is here: https://docs.google.com/document/d/1VmafcHvvrP5kwtQkz84R5yXF7u7fW-9Pn0bkSUQHDt8/edit?disco=AAAAARNByxg Change on 2014/12/09 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81716456
10 years ago
test/core/iomgr/resolve_address_test.c \
Initial import.
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
RESOLVE_ADDRESS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(RESOLVE_ADDRESS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/resolve_address_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/resolve_address_test: $(RESOLVE_ADDRESS_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(RESOLVE_ADDRESS_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/resolve_address_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/iomgr/resolve_address_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_resolve_address_test: $(RESOLVE_ADDRESS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(RESOLVE_ADDRESS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Introducing iomgr. Move eventmanager and platform dependent endpoint functionality into a single library called 'iomgr'. This is primarily to prepare for a Windows port - where posix socket semantics lead to poor quality code. Mostly this is a code movement CL, with some small changes to help prepare the way for porting: - em style fd objects can only be held internally in iomgr, and own their memory - added grpc_iomgr_create_endpoint_pair() to accomodate the common pattern of creating a tcp endpoint from the output of socketpair - this will help keep our tests portable - separated em alarm interface into a separate file, as this part of event manager is needed higher up the stack - made the eventmanager bits a true singleton, simplifying API's across the stack as there's no longer a reason to carry a pointer there. Initial design document is here: https://docs.google.com/document/d/1VmafcHvvrP5kwtQkz84R5yXF7u7fW-9Pn0bkSUQHDt8/edit?disco=AAAAARNByxg Change on 2014/12/09 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81716456
10 years ago
SOCKADDR_UTILS_TEST_SRC = \
test/core/iomgr/sockaddr_utils_test.c \
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
SOCKADDR_UTILS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(SOCKADDR_UTILS_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/sockaddr_utils_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/sockaddr_utils_test: $(SOCKADDR_UTILS_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(SOCKADDR_UTILS_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/sockaddr_utils_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/iomgr/sockaddr_utils_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_sockaddr_utils_test: $(SOCKADDR_UTILS_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(SOCKADDR_UTILS_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Introducing iomgr. Move eventmanager and platform dependent endpoint functionality into a single library called 'iomgr'. This is primarily to prepare for a Windows port - where posix socket semantics lead to poor quality code. Mostly this is a code movement CL, with some small changes to help prepare the way for porting: - em style fd objects can only be held internally in iomgr, and own their memory - added grpc_iomgr_create_endpoint_pair() to accomodate the common pattern of creating a tcp endpoint from the output of socketpair - this will help keep our tests portable - separated em alarm interface into a separate file, as this part of event manager is needed higher up the stack - made the eventmanager bits a true singleton, simplifying API's across the stack as there's no longer a reason to carry a pointer there. Initial design document is here: https://docs.google.com/document/d/1VmafcHvvrP5kwtQkz84R5yXF7u7fW-9Pn0bkSUQHDt8/edit?disco=AAAAARNByxg Change on 2014/12/09 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81716456
10 years ago
TCP_SERVER_POSIX_TEST_SRC = \
test/core/iomgr/tcp_server_posix_test.c \
Initial import.
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
TCP_SERVER_POSIX_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(TCP_SERVER_POSIX_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/tcp_server_posix_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/tcp_server_posix_test: $(TCP_SERVER_POSIX_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(TCP_SERVER_POSIX_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/tcp_server_posix_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/iomgr/tcp_server_posix_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_tcp_server_posix_test: $(TCP_SERVER_POSIX_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(TCP_SERVER_POSIX_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Introducing iomgr. Move eventmanager and platform dependent endpoint functionality into a single library called 'iomgr'. This is primarily to prepare for a Windows port - where posix socket semantics lead to poor quality code. Mostly this is a code movement CL, with some small changes to help prepare the way for porting: - em style fd objects can only be held internally in iomgr, and own their memory - added grpc_iomgr_create_endpoint_pair() to accomodate the common pattern of creating a tcp endpoint from the output of socketpair - this will help keep our tests portable - separated em alarm interface into a separate file, as this part of event manager is needed higher up the stack - made the eventmanager bits a true singleton, simplifying API's across the stack as there's no longer a reason to carry a pointer there. Initial design document is here: https://docs.google.com/document/d/1VmafcHvvrP5kwtQkz84R5yXF7u7fW-9Pn0bkSUQHDt8/edit?disco=AAAAARNByxg Change on 2014/12/09 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81716456
10 years ago
TCP_CLIENT_POSIX_TEST_SRC = \
test/core/iomgr/tcp_client_posix_test.c \
Initial import.
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
TCP_CLIENT_POSIX_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(TCP_CLIENT_POSIX_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/tcp_client_posix_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/tcp_client_posix_test: $(TCP_CLIENT_POSIX_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(TCP_CLIENT_POSIX_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/tcp_client_posix_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/iomgr/tcp_client_posix_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_tcp_client_posix_test: $(TCP_CLIENT_POSIX_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(TCP_CLIENT_POSIX_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
GRPC_CHANNEL_STACK_TEST_SRC = \
test/core/channel/channel_stack_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GRPC_CHANNEL_STACK_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GRPC_CHANNEL_STACK_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/grpc_channel_stack_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/grpc_channel_stack_test: $(GRPC_CHANNEL_STACK_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GRPC_CHANNEL_STACK_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/grpc_channel_stack_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/channel/channel_stack_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_grpc_channel_stack_test: $(GRPC_CHANNEL_STACK_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GRPC_CHANNEL_STACK_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
METADATA_BUFFER_TEST_SRC = \
test/core/channel/metadata_buffer_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
METADATA_BUFFER_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(METADATA_BUFFER_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/metadata_buffer_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/metadata_buffer_test: $(METADATA_BUFFER_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(METADATA_BUFFER_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/metadata_buffer_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/channel/metadata_buffer_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_metadata_buffer_test: $(METADATA_BUFFER_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(METADATA_BUFFER_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
GRPC_COMPLETION_QUEUE_TEST_SRC = \
test/core/surface/completion_queue_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GRPC_COMPLETION_QUEUE_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GRPC_COMPLETION_QUEUE_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/grpc_completion_queue_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/grpc_completion_queue_test: $(GRPC_COMPLETION_QUEUE_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GRPC_COMPLETION_QUEUE_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/grpc_completion_queue_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/surface/completion_queue_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_grpc_completion_queue_test: $(GRPC_COMPLETION_QUEUE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GRPC_COMPLETION_QUEUE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
GRPC_COMPLETION_QUEUE_BENCHMARK_SRC = \
test/core/surface/completion_queue_benchmark.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GRPC_COMPLETION_QUEUE_BENCHMARK_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GRPC_COMPLETION_QUEUE_BENCHMARK_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/grpc_completion_queue_benchmark: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/grpc_completion_queue_benchmark: $(GRPC_COMPLETION_QUEUE_BENCHMARK_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GRPC_COMPLETION_QUEUE_BENCHMARK_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/grpc_completion_queue_benchmark
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/surface/completion_queue_benchmark.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_grpc_completion_queue_benchmark: $(GRPC_COMPLETION_QUEUE_BENCHMARK_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GRPC_COMPLETION_QUEUE_BENCHMARK_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
CENSUS_TRACE_STORE_TEST_SRC = \
test/core/statistics/trace_test.c \
CENSUS_TRACE_STORE_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CENSUS_TRACE_STORE_TEST_SRC))))
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
bins/$(CONFIG)/census_trace_store_test: openssl_dep_error
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/census_trace_store_test: $(CENSUS_TRACE_STORE_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CENSUS_TRACE_STORE_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/census_trace_store_test
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/statistics/trace_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_census_trace_store_test: $(CENSUS_TRACE_STORE_TEST_OBJS:.o=.dep)
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CENSUS_TRACE_STORE_TEST_OBJS:.o=.dep)
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
endif
endif
CENSUS_STATS_STORE_TEST_SRC = \
test/core/statistics/rpc_stats_test.c \
CENSUS_STATS_STORE_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CENSUS_STATS_STORE_TEST_SRC))))
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
bins/$(CONFIG)/census_stats_store_test: openssl_dep_error
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/census_stats_store_test: $(CENSUS_STATS_STORE_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CENSUS_STATS_STORE_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/census_stats_store_test
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/statistics/rpc_stats_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_census_stats_store_test: $(CENSUS_STATS_STORE_TEST_OBJS:.o=.dep)
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CENSUS_STATS_STORE_TEST_OBJS:.o=.dep)
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
endif
endif
Initial import.
10 years ago
CENSUS_WINDOW_STATS_TEST_SRC = \
test/core/statistics/window_stats_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CENSUS_WINDOW_STATS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CENSUS_WINDOW_STATS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/census_window_stats_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/census_window_stats_test: $(CENSUS_WINDOW_STATS_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CENSUS_WINDOW_STATS_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/census_window_stats_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/statistics/window_stats_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_census_window_stats_test: $(CENSUS_WINDOW_STATS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CENSUS_WINDOW_STATS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CENSUS_STATISTICS_QUICK_TEST_SRC = \
test/core/statistics/quick_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CENSUS_STATISTICS_QUICK_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CENSUS_STATISTICS_QUICK_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/census_statistics_quick_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/census_statistics_quick_test: $(CENSUS_STATISTICS_QUICK_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CENSUS_STATISTICS_QUICK_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/census_statistics_quick_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/statistics/quick_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_census_statistics_quick_test: $(CENSUS_STATISTICS_QUICK_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CENSUS_STATISTICS_QUICK_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Fix a bug in base-log, and add test. The tests compute usable space in the log, but do so using subtraction on unsigned values and did not correctly account for number of blocks used per-core. This could lead underflow, and an incorrect space calculation. In addition, testing showed that with current defaults, tests would fail on a system with > ~64 cores. This can be alleviated through use of a larger log. An explicit check in the log initialization has been added for this case, and the log size increased if necessary. Change on 2014/12/15 by aveitch <aveitch@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82154432
10 years ago
CENSUS_STATISTICS_SMALL_LOG_TEST_SRC = \
test/core/statistics/small_log_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CENSUS_STATISTICS_SMALL_LOG_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CENSUS_STATISTICS_SMALL_LOG_TEST_SRC))))
Fix a bug in base-log, and add test. The tests compute usable space in the log, but do so using subtraction on unsigned values and did not correctly account for number of blocks used per-core. This could lead underflow, and an incorrect space calculation. In addition, testing showed that with current defaults, tests would fail on a system with > ~64 cores. This can be alleviated through use of a larger log. An explicit check in the log initialization has been added for this case, and the log size increased if necessary. Change on 2014/12/15 by aveitch <aveitch@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82154432
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/census_statistics_small_log_test: openssl_dep_error
Fix a bug in base-log, and add test. The tests compute usable space in the log, but do so using subtraction on unsigned values and did not correctly account for number of blocks used per-core. This could lead underflow, and an incorrect space calculation. In addition, testing showed that with current defaults, tests would fail on a system with > ~64 cores. This can be alleviated through use of a larger log. An explicit check in the log initialization has been added for this case, and the log size increased if necessary. Change on 2014/12/15 by aveitch <aveitch@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82154432
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/census_statistics_small_log_test: $(CENSUS_STATISTICS_SMALL_LOG_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix a bug in base-log, and add test. The tests compute usable space in the log, but do so using subtraction on unsigned values and did not correctly account for number of blocks used per-core. This could lead underflow, and an incorrect space calculation. In addition, testing showed that with current defaults, tests would fail on a system with > ~64 cores. This can be alleviated through use of a larger log. An explicit check in the log initialization has been added for this case, and the log size increased if necessary. Change on 2014/12/15 by aveitch <aveitch@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82154432
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CENSUS_STATISTICS_SMALL_LOG_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/census_statistics_small_log_test
Fix a bug in base-log, and add test. The tests compute usable space in the log, but do so using subtraction on unsigned values and did not correctly account for number of blocks used per-core. This could lead underflow, and an incorrect space calculation. In addition, testing showed that with current defaults, tests would fail on a system with > ~64 cores. This can be alleviated through use of a larger log. An explicit check in the log initialization has been added for this case, and the log size increased if necessary. Change on 2014/12/15 by aveitch <aveitch@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82154432
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/statistics/small_log_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_census_statistics_small_log_test: $(CENSUS_STATISTICS_SMALL_LOG_TEST_OBJS:.o=.dep)
Fix a bug in base-log, and add test. The tests compute usable space in the log, but do so using subtraction on unsigned values and did not correctly account for number of blocks used per-core. This could lead underflow, and an incorrect space calculation. In addition, testing showed that with current defaults, tests would fail on a system with > ~64 cores. This can be alleviated through use of a larger log. An explicit check in the log initialization has been added for this case, and the log size increased if necessary. Change on 2014/12/15 by aveitch <aveitch@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82154432
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CENSUS_STATISTICS_SMALL_LOG_TEST_OBJS:.o=.dep)
Fix a bug in base-log, and add test. The tests compute usable space in the log, but do so using subtraction on unsigned values and did not correctly account for number of blocks used per-core. This could lead underflow, and an incorrect space calculation. In addition, testing showed that with current defaults, tests would fail on a system with > ~64 cores. This can be alleviated through use of a larger log. An explicit check in the log initialization has been added for this case, and the log size increased if necessary. Change on 2014/12/15 by aveitch <aveitch@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82154432
10 years ago
endif
endif
Initial import.
10 years ago
CENSUS_STATISTICS_PERFORMANCE_TEST_SRC = \
test/core/statistics/performance_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CENSUS_STATISTICS_PERFORMANCE_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CENSUS_STATISTICS_PERFORMANCE_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/census_statistics_performance_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/census_statistics_performance_test: $(CENSUS_STATISTICS_PERFORMANCE_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CENSUS_STATISTICS_PERFORMANCE_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/census_statistics_performance_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/statistics/performance_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_census_statistics_performance_test: $(CENSUS_STATISTICS_PERFORMANCE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CENSUS_STATISTICS_PERFORMANCE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CENSUS_STATISTICS_MULTIPLE_WRITERS_TEST_SRC = \
test/core/statistics/multiple_writers_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CENSUS_STATISTICS_MULTIPLE_WRITERS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CENSUS_STATISTICS_MULTIPLE_WRITERS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/census_statistics_multiple_writers_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/census_statistics_multiple_writers_test: $(CENSUS_STATISTICS_MULTIPLE_WRITERS_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CENSUS_STATISTICS_MULTIPLE_WRITERS_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/census_statistics_multiple_writers_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/statistics/multiple_writers_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_census_statistics_multiple_writers_test: $(CENSUS_STATISTICS_MULTIPLE_WRITERS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CENSUS_STATISTICS_MULTIPLE_WRITERS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CENSUS_STATISTICS_MULTIPLE_WRITERS_CIRCULAR_BUFFER_TEST_SRC = \
test/core/statistics/multiple_writers_circular_buffer_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CENSUS_STATISTICS_MULTIPLE_WRITERS_CIRCULAR_BUFFER_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CENSUS_STATISTICS_MULTIPLE_WRITERS_CIRCULAR_BUFFER_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/census_statistics_multiple_writers_circular_buffer_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/census_statistics_multiple_writers_circular_buffer_test: $(CENSUS_STATISTICS_MULTIPLE_WRITERS_CIRCULAR_BUFFER_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CENSUS_STATISTICS_MULTIPLE_WRITERS_CIRCULAR_BUFFER_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/census_statistics_multiple_writers_circular_buffer_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/statistics/multiple_writers_circular_buffer_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_census_statistics_multiple_writers_circular_buffer_test: $(CENSUS_STATISTICS_MULTIPLE_WRITERS_CIRCULAR_BUFFER_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CENSUS_STATISTICS_MULTIPLE_WRITERS_CIRCULAR_BUFFER_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CENSUS_STUB_TEST_SRC = \
test/core/statistics/census_stub_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CENSUS_STUB_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CENSUS_STUB_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/census_stub_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/census_stub_test: $(CENSUS_STUB_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CENSUS_STUB_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/census_stub_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/statistics/census_stub_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_census_stub_test: $(CENSUS_STUB_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CENSUS_STUB_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CENSUS_HASH_TABLE_TEST_SRC = \
test/core/statistics/hash_table_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CENSUS_HASH_TABLE_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CENSUS_HASH_TABLE_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/census_hash_table_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/census_hash_table_test: $(CENSUS_HASH_TABLE_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CENSUS_HASH_TABLE_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/census_hash_table_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/statistics/hash_table_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_census_hash_table_test: $(CENSUS_HASH_TABLE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CENSUS_HASH_TABLE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
FLING_SERVER_SRC = \
test/core/fling/server.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
FLING_SERVER_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(FLING_SERVER_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/fling_server: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/fling_server: $(FLING_SERVER_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(FLING_SERVER_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/fling_server
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/fling/server.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_fling_server: $(FLING_SERVER_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(FLING_SERVER_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
FLING_CLIENT_SRC = \
test/core/fling/client.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
FLING_CLIENT_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(FLING_CLIENT_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/fling_client: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/fling_client: $(FLING_CLIENT_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(FLING_CLIENT_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/fling_client
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/fling/client.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_fling_client: $(FLING_CLIENT_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(FLING_CLIENT_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
FLING_TEST_SRC = \
test/core/fling/fling_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
FLING_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(FLING_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/fling_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/fling_test: $(FLING_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(FLING_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/fling_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/fling/fling_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_fling_test: $(FLING_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(FLING_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
ECHO_SERVER_SRC = \
test/core/echo/server.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
ECHO_SERVER_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(ECHO_SERVER_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/echo_server: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/echo_server: $(ECHO_SERVER_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(ECHO_SERVER_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/echo_server
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/echo/server.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_echo_server: $(ECHO_SERVER_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(ECHO_SERVER_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
ECHO_CLIENT_SRC = \
test/core/echo/client.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
ECHO_CLIENT_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(ECHO_CLIENT_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/echo_client: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/echo_client: $(ECHO_CLIENT_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(ECHO_CLIENT_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/echo_client
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/echo/client.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_echo_client: $(ECHO_CLIENT_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(ECHO_CLIENT_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
ECHO_TEST_SRC = \
test/core/echo/echo_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
ECHO_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(ECHO_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/echo_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/echo_test: $(ECHO_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(ECHO_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/echo_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/echo/echo_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_echo_test: $(ECHO_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(ECHO_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
LOW_LEVEL_PING_PONG_BENCHMARK_SRC = \
test/core/network_benchmarks/low_level_ping_pong.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LOW_LEVEL_PING_PONG_BENCHMARK_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LOW_LEVEL_PING_PONG_BENCHMARK_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/low_level_ping_pong_benchmark: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/low_level_ping_pong_benchmark: $(LOW_LEVEL_PING_PONG_BENCHMARK_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(LOW_LEVEL_PING_PONG_BENCHMARK_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/low_level_ping_pong_benchmark
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/network_benchmarks/low_level_ping_pong.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_low_level_ping_pong_benchmark: $(LOW_LEVEL_PING_PONG_BENCHMARK_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LOW_LEVEL_PING_PONG_BENCHMARK_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
MESSAGE_COMPRESS_TEST_SRC = \
test/core/compression/message_compress_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
MESSAGE_COMPRESS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(MESSAGE_COMPRESS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/message_compress_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/message_compress_test: $(MESSAGE_COMPRESS_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(MESSAGE_COMPRESS_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/message_compress_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/compression/message_compress_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_message_compress_test: $(MESSAGE_COMPRESS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(MESSAGE_COMPRESS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
BIN_ENCODER_TEST_SRC = \
test/core/transport/chttp2/bin_encoder_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
BIN_ENCODER_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(BIN_ENCODER_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/bin_encoder_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/bin_encoder_test: $(BIN_ENCODER_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(BIN_ENCODER_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/bin_encoder_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/transport/chttp2/bin_encoder_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_bin_encoder_test: $(BIN_ENCODER_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(BIN_ENCODER_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Initial import.
10 years ago
SECURE_ENDPOINT_TEST_SRC = \
Remove endpoint/ directories. Fold endpoint interface into iomgr, move secure_endpoint into security/. This will make it easier for endpoint to rely on some iomgr defined types (like pollset). Change on 2014/12/17 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82338036
10 years ago
test/core/security/secure_endpoint_test.c \
Initial import.
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
SECURE_ENDPOINT_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(SECURE_ENDPOINT_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/secure_endpoint_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/secure_endpoint_test: $(SECURE_ENDPOINT_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(SECURE_ENDPOINT_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/secure_endpoint_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/security/secure_endpoint_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_secure_endpoint_test: $(SECURE_ENDPOINT_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(SECURE_ENDPOINT_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
HTTPCLI_FORMAT_REQUEST_TEST_SRC = \
test/core/httpcli/format_request_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
HTTPCLI_FORMAT_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(HTTPCLI_FORMAT_REQUEST_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/httpcli_format_request_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/httpcli_format_request_test: $(HTTPCLI_FORMAT_REQUEST_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(HTTPCLI_FORMAT_REQUEST_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/httpcli_format_request_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/httpcli/format_request_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_httpcli_format_request_test: $(HTTPCLI_FORMAT_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(HTTPCLI_FORMAT_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
HTTPCLI_PARSER_TEST_SRC = \
test/core/httpcli/parser_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
HTTPCLI_PARSER_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(HTTPCLI_PARSER_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/httpcli_parser_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/httpcli_parser_test: $(HTTPCLI_PARSER_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(HTTPCLI_PARSER_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/httpcli_parser_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/httpcli/parser_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_httpcli_parser_test: $(HTTPCLI_PARSER_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(HTTPCLI_PARSER_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
HTTPCLI_TEST_SRC = \
test/core/httpcli/httpcli_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
HTTPCLI_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(HTTPCLI_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/httpcli_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/httpcli_test: $(HTTPCLI_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(HTTPCLI_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/httpcli_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/httpcli/httpcli_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_httpcli_test: $(HTTPCLI_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(HTTPCLI_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
GRPC_CREDENTIALS_TEST_SRC = \
test/core/security/credentials_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GRPC_CREDENTIALS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GRPC_CREDENTIALS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/grpc_credentials_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/grpc_credentials_test: $(GRPC_CREDENTIALS_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GRPC_CREDENTIALS_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/grpc_credentials_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/security/credentials_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_grpc_credentials_test: $(GRPC_CREDENTIALS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GRPC_CREDENTIALS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Adding support for service account credentials. - Tested end to end with a JSON key I generated for my account using the fetch_oauth2 binary. - The same fetch_oauth2 binary can get a token from the GCE metadata service on a VM in cloud. Change on 2014/12/19 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82548689
10 years ago
GRPC_FETCH_OAUTH2_SRC = \
test/core/security/fetch_oauth2.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GRPC_FETCH_OAUTH2_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GRPC_FETCH_OAUTH2_SRC))))
Adding support for service account credentials. - Tested end to end with a JSON key I generated for my account using the fetch_oauth2 binary. - The same fetch_oauth2 binary can get a token from the GCE metadata service on a VM in cloud. Change on 2014/12/19 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82548689
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/grpc_fetch_oauth2: openssl_dep_error
Adding support for service account credentials. - Tested end to end with a JSON key I generated for my account using the fetch_oauth2 binary. - The same fetch_oauth2 binary can get a token from the GCE metadata service on a VM in cloud. Change on 2014/12/19 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82548689
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/grpc_fetch_oauth2: $(GRPC_FETCH_OAUTH2_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Adding support for service account credentials. - Tested end to end with a JSON key I generated for my account using the fetch_oauth2 binary. - The same fetch_oauth2 binary can get a token from the GCE metadata service on a VM in cloud. Change on 2014/12/19 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82548689
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GRPC_FETCH_OAUTH2_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/grpc_fetch_oauth2
Adding support for service account credentials. - Tested end to end with a JSON key I generated for my account using the fetch_oauth2 binary. - The same fetch_oauth2 binary can get a token from the GCE metadata service on a VM in cloud. Change on 2014/12/19 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82548689
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/security/fetch_oauth2.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_grpc_fetch_oauth2: $(GRPC_FETCH_OAUTH2_OBJS:.o=.dep)
Adding support for service account credentials. - Tested end to end with a JSON key I generated for my account using the fetch_oauth2 binary. - The same fetch_oauth2 binary can get a token from the GCE metadata service on a VM in cloud. Change on 2014/12/19 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82548689
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GRPC_FETCH_OAUTH2_OBJS:.o=.dep)
Adding support for service account credentials. - Tested end to end with a JSON key I generated for my account using the fetch_oauth2 binary. - The same fetch_oauth2 binary can get a token from the GCE metadata service on a VM in cloud. Change on 2014/12/19 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82548689
10 years ago
endif
endif
Adding base utils for JWT service account workflow. OpenSSL base64 decoding is a disaster and does not support url_safe which we need for the JWT. Change on 2014/12/12 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020601
10 years ago
GRPC_BASE64_TEST_SRC = \
test/core/security/base64_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GRPC_BASE64_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GRPC_BASE64_TEST_SRC))))
Adding base utils for JWT service account workflow. OpenSSL base64 decoding is a disaster and does not support url_safe which we need for the JWT. Change on 2014/12/12 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020601
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/grpc_base64_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/grpc_base64_test: $(GRPC_BASE64_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Adding base utils for JWT service account workflow. OpenSSL base64 decoding is a disaster and does not support url_safe which we need for the JWT. Change on 2014/12/12 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020601
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GRPC_BASE64_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/grpc_base64_test
Adding base utils for JWT service account workflow. OpenSSL base64 decoding is a disaster and does not support url_safe which we need for the JWT. Change on 2014/12/12 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020601
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/security/base64_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_grpc_base64_test: $(GRPC_BASE64_TEST_OBJS:.o=.dep)
Adding base utils for JWT service account workflow. OpenSSL base64 decoding is a disaster and does not support url_safe which we need for the JWT. Change on 2014/12/12 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020601
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GRPC_BASE64_TEST_OBJS:.o=.dep)
Adding base utils for JWT service account workflow. OpenSSL base64 decoding is a disaster and does not support url_safe which we need for the JWT. Change on 2014/12/12 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020601
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Adding base utils for JWT service account workflow. OpenSSL base64 decoding is a disaster and does not support url_safe which we need for the JWT. Change on 2014/12/12 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020601
10 years ago
GRPC_JSON_TOKEN_TEST_SRC = \
test/core/security/json_token_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
GRPC_JSON_TOKEN_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(GRPC_JSON_TOKEN_TEST_SRC))))
Adding base utils for JWT service account workflow. OpenSSL base64 decoding is a disaster and does not support url_safe which we need for the JWT. Change on 2014/12/12 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020601
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/grpc_json_token_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/grpc_json_token_test: $(GRPC_JSON_TOKEN_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Adding base utils for JWT service account workflow. OpenSSL base64 decoding is a disaster and does not support url_safe which we need for the JWT. Change on 2014/12/12 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020601
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(GRPC_JSON_TOKEN_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/grpc_json_token_test
Adding base utils for JWT service account workflow. OpenSSL base64 decoding is a disaster and does not support url_safe which we need for the JWT. Change on 2014/12/12 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020601
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/security/json_token_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_grpc_json_token_test: $(GRPC_JSON_TOKEN_TEST_OBJS:.o=.dep)
Adding base utils for JWT service account workflow. OpenSSL base64 decoding is a disaster and does not support url_safe which we need for the JWT. Change on 2014/12/12 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020601
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(GRPC_JSON_TOKEN_TEST_OBJS:.o=.dep)
Adding base utils for JWT service account workflow. OpenSSL base64 decoding is a disaster and does not support url_safe which we need for the JWT. Change on 2014/12/12 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020601
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Adding base utils for JWT service account workflow. OpenSSL base64 decoding is a disaster and does not support url_safe which we need for the JWT. Change on 2014/12/12 by jboeuf <jboeuf@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020601
10 years ago
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
TIMEOUT_ENCODING_TEST_SRC = \
test/core/transport/chttp2/timeout_encoding_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
TIMEOUT_ENCODING_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(TIMEOUT_ENCODING_TEST_SRC))))
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/timeout_encoding_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/timeout_encoding_test: $(TIMEOUT_ENCODING_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(TIMEOUT_ENCODING_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/timeout_encoding_test
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/transport/chttp2/timeout_encoding_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_timeout_encoding_test: $(TIMEOUT_ENCODING_TEST_OBJS:.o=.dep)
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(TIMEOUT_ENCODING_TEST_OBJS:.o=.dep)
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
FD_POSIX_TEST_SRC = \
test/core/iomgr/fd_posix_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
FD_POSIX_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(FD_POSIX_TEST_SRC))))
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/fd_posix_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/fd_posix_test: $(FD_POSIX_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(FD_POSIX_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/fd_posix_test
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/iomgr/fd_posix_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_fd_posix_test: $(FD_POSIX_TEST_OBJS:.o=.dep)
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(FD_POSIX_TEST_OBJS:.o=.dep)
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Initial import.
10 years ago
FLING_STREAM_TEST_SRC = \
test/core/fling/fling_stream_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
FLING_STREAM_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(FLING_STREAM_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/fling_stream_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/fling_stream_test: $(FLING_STREAM_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(FLING_STREAM_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/fling_stream_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/fling/fling_stream_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_fling_stream_test: $(FLING_STREAM_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(FLING_STREAM_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
LAME_CLIENT_TEST_SRC = \
test/core/surface/lame_client_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
LAME_CLIENT_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(LAME_CLIENT_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/lame_client_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/lame_client_test: $(LAME_CLIENT_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(LAME_CLIENT_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/lame_client_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/surface/lame_client_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_lame_client_test: $(LAME_CLIENT_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(LAME_CLIENT_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
THREAD_POOL_TEST_SRC = \
test/cpp/server/thread_pool_test.cc \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
THREAD_POOL_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(THREAD_POOL_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/thread_pool_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/thread_pool_test: $(THREAD_POOL_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LDXX) $(LDFLAGS) $(THREAD_POOL_TEST_OBJS) $(GTEST_LIB) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBSXX) $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/thread_pool_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/cpp/server/thread_pool_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_thread_pool_test: $(THREAD_POOL_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(THREAD_POOL_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
STATUS_TEST_SRC = \
test/cpp/util/status_test.cc \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
STATUS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(STATUS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/status_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/status_test: $(STATUS_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LDXX) $(LDFLAGS) $(STATUS_TEST_OBJS) $(GTEST_LIB) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBSXX) $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/status_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/cpp/util/status_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_status_test: $(STATUS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(STATUS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
SYNC_CLIENT_ASYNC_SERVER_TEST_SRC = \
test/cpp/end2end/sync_client_async_server_test.cc \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
SYNC_CLIENT_ASYNC_SERVER_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(SYNC_CLIENT_ASYNC_SERVER_TEST_SRC))))
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/sync_client_async_server_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Merge github.com:google/grpc into testcpp
10 years ago
bins/$(CONFIG)/sync_client_async_server_test: $(SYNC_CLIENT_ASYNC_SERVER_TEST_OBJS) libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Merge github.com:google/grpc into testcpp
10 years ago
$(Q) $(LDXX) $(LDFLAGS) $(SYNC_CLIENT_ASYNC_SERVER_TEST_OBJS) $(GTEST_LIB) libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBSXX) $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/sync_client_async_server_test
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/cpp/end2end/sync_client_async_server_test.o: libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_sync_client_async_server_test: $(SYNC_CLIENT_ASYNC_SERVER_TEST_OBJS:.o=.dep)
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(SYNC_CLIENT_ASYNC_SERVER_TEST_OBJS:.o=.dep)
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
QPS_CLIENT_SRC = \
Fix qps test. This was failing to compile.
10 years ago
gens/test/cpp/qps/qpstest.pb.cc \
Prepare QPS tests for running in GCE Edit build.json to include qps tests and credentials Perform required MOE edits for running qps tests in VM Minor changes to underlying source files to support issues like more restrictive warning flags in cloud. Change on 2014/12/17 by vpai <vpai@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82349829
10 years ago
test/cpp/qps/client.cc \
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
QPS_CLIENT_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(QPS_CLIENT_SRC))))
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/qps_client: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/qps_client: $(QPS_CLIENT_OBJS) libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LDXX) $(LDFLAGS) $(QPS_CLIENT_OBJS) $(GTEST_LIB) libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBSXX) $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/qps_client
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix qps test. This was failing to compile.
10 years ago
objs/$(CONFIG)/test/cpp/qps/qpstest.o: libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/cpp/qps/client.o: libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_qps_client: $(QPS_CLIENT_OBJS:.o=.dep)
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(QPS_CLIENT_OBJS:.o=.dep)
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
QPS_SERVER_SRC = \
Fix qps test. This was failing to compile.
10 years ago
gens/test/cpp/qps/qpstest.pb.cc \
Prepare QPS tests for running in GCE Edit build.json to include qps tests and credentials Perform required MOE edits for running qps tests in VM Minor changes to underlying source files to support issues like more restrictive warning flags in cloud. Change on 2014/12/17 by vpai <vpai@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82349829
10 years ago
test/cpp/qps/server.cc \
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
QPS_SERVER_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(QPS_SERVER_SRC))))
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/qps_server: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/qps_server: $(QPS_SERVER_OBJS) libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LDXX) $(LDFLAGS) $(QPS_SERVER_OBJS) $(GTEST_LIB) libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBSXX) $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/qps_server
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix qps test. This was failing to compile.
10 years ago
objs/$(CONFIG)/test/cpp/qps/qpstest.o: libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/cpp/qps/server.o: libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_qps_server: $(QPS_SERVER_OBJS:.o=.dep)
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(QPS_SERVER_OBJS:.o=.dep)
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
INTEROP_SERVER_SRC = \
Getting the generated Makefile to support .proto files. -) Compiling the proto plugin using the HOSTCC compiler. -) Set up proper rules to invoke the plugin from protoc. -) Few various renaming hacks to fully get out of []. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82000361
10 years ago
gens/test/cpp/interop/empty.pb.cc \
gens/test/cpp/interop/messages.pb.cc \
gens/test/cpp/interop/test.pb.cc \
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
test/cpp/interop/server.cc \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
INTEROP_SERVER_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(INTEROP_SERVER_SRC))))
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/interop_server: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/interop_server: $(INTEROP_SERVER_OBJS) libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LDXX) $(LDFLAGS) $(INTEROP_SERVER_OBJS) $(GTEST_LIB) libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBSXX) $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/interop_server
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/cpp/interop/empty.o: libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
objs/$(CONFIG)/test/cpp/interop/messages.o: libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
objs/$(CONFIG)/test/cpp/interop/test.o: libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
objs/$(CONFIG)/test/cpp/interop/server.o: libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_interop_server: $(INTEROP_SERVER_OBJS:.o=.dep)
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(INTEROP_SERVER_OBJS:.o=.dep)
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
INTEROP_CLIENT_SRC = \
Getting the generated Makefile to support .proto files. -) Compiling the proto plugin using the HOSTCC compiler. -) Set up proper rules to invoke the plugin from protoc. -) Few various renaming hacks to fully get out of []. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82000361
10 years ago
gens/test/cpp/interop/empty.pb.cc \
gens/test/cpp/interop/messages.pb.cc \
gens/test/cpp/interop/test.pb.cc \
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
test/cpp/interop/client.cc \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
INTEROP_CLIENT_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(INTEROP_CLIENT_SRC))))
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/interop_client: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/interop_client: $(INTEROP_CLIENT_OBJS) libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LDXX) $(LDFLAGS) $(INTEROP_CLIENT_OBJS) $(GTEST_LIB) libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBSXX) $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/interop_client
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/cpp/interop/empty.o: libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
objs/$(CONFIG)/test/cpp/interop/messages.o: libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
objs/$(CONFIG)/test/cpp/interop/test.o: libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
objs/$(CONFIG)/test/cpp/interop/client.o: libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_interop_client: $(INTEROP_CLIENT_OBJS:.o=.dep)
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(INTEROP_CLIENT_OBJS:.o=.dep)
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
TIPS client implementation.
10 years ago
TIPS_CLIENT_SRC = \
examples/tips/client_main.cc \
TIPS_CLIENT_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(TIPS_CLIENT_SRC))))
ifeq ($(NO_SECURE),true)
# You can't build secure targets if you don't have OpenSSL with ALPN.
bins/$(CONFIG)/tips_client: openssl_dep_error
else
bins/$(CONFIG)/tips_client: $(TIPS_CLIENT_OBJS) libs/$(CONFIG)/libtips_client_lib.a libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
$(Q) $(LDXX) $(LDFLAGS) $(TIPS_CLIENT_OBJS) $(GTEST_LIB) libs/$(CONFIG)/libtips_client_lib.a libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBSXX) $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/tips_client
endif
objs/$(CONFIG)/examples/tips/client_main.o: libs/$(CONFIG)/libtips_client_lib.a libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
deps_tips_client: $(TIPS_CLIENT_OBJS:.o=.dep)
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
-include $(TIPS_CLIENT_OBJS:.o=.dep)
endif
endif
TIPS_CLIENT_TEST_SRC = \
examples/tips/client_test.cc \
TIPS_CLIENT_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(TIPS_CLIENT_TEST_SRC))))
ifeq ($(NO_SECURE),true)
# You can't build secure targets if you don't have OpenSSL with ALPN.
bins/$(CONFIG)/tips_client_test: openssl_dep_error
else
bins/$(CONFIG)/tips_client_test: $(TIPS_CLIENT_TEST_OBJS) libs/$(CONFIG)/libtips_client_lib.a libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
$(Q) $(LDXX) $(LDFLAGS) $(TIPS_CLIENT_TEST_OBJS) $(GTEST_LIB) libs/$(CONFIG)/libtips_client_lib.a libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBSXX) $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/tips_client_test
endif
objs/$(CONFIG)/examples/tips/client_test.o: libs/$(CONFIG)/libtips_client_lib.a libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
deps_tips_client_test: $(TIPS_CLIENT_TEST_OBJS:.o=.dep)
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
-include $(TIPS_CLIENT_TEST_OBJS:.o=.dep)
endif
endif
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
END2END_TEST_SRC = \
test/cpp/end2end/end2end_test.cc \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
END2END_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(END2END_TEST_SRC))))
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/end2end_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/end2end_test: $(END2END_TEST_OBJS) libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LDXX) $(LDFLAGS) $(END2END_TEST_OBJS) $(GTEST_LIB) libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBSXX) $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/end2end_test
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/cpp/end2end/end2end_test.o: libs/$(CONFIG)/libgrpc++_test_util.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_end2end_test: $(END2END_TEST_OBJS:.o=.dep)
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(END2END_TEST_OBJS:.o=.dep)
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Client side implementation of creating channel with credentials. The old test_ssl_channel related code is deleted and the new create channel call is used for interop tests. Change on 2014/12/19 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82540921
10 years ago
CHANNEL_ARGUMENTS_TEST_SRC = \
test/cpp/client/channel_arguments_test.cc \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHANNEL_ARGUMENTS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHANNEL_ARGUMENTS_TEST_SRC))))
Client side implementation of creating channel with credentials. The old test_ssl_channel related code is deleted and the new create channel call is used for interop tests. Change on 2014/12/19 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82540921
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/channel_arguments_test: openssl_dep_error
Client side implementation of creating channel with credentials. The old test_ssl_channel related code is deleted and the new create channel call is used for interop tests. Change on 2014/12/19 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82540921
10 years ago
else
Fix some C++ test dependencies
10 years ago
bins/$(CONFIG)/channel_arguments_test: $(CHANNEL_ARGUMENTS_TEST_OBJS) libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr.a
Client side implementation of creating channel with credentials. The old test_ssl_channel related code is deleted and the new create channel call is used for interop tests. Change on 2014/12/19 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82540921
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Fix some C++ test dependencies
10 years ago
$(Q) $(LDXX) $(LDFLAGS) $(CHANNEL_ARGUMENTS_TEST_OBJS) $(GTEST_LIB) libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr.a $(LDLIBSXX) $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/channel_arguments_test
Client side implementation of creating channel with credentials. The old test_ssl_channel related code is deleted and the new create channel call is used for interop tests. Change on 2014/12/19 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82540921
10 years ago
endif
Fix some C++ test dependencies
10 years ago
objs/$(CONFIG)/test/cpp/client/channel_arguments_test.o: libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr.a
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_channel_arguments_test: $(CHANNEL_ARGUMENTS_TEST_OBJS:.o=.dep)
Client side implementation of creating channel with credentials. The old test_ssl_channel related code is deleted and the new create channel call is used for interop tests. Change on 2014/12/19 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82540921
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHANNEL_ARGUMENTS_TEST_OBJS:.o=.dep)
Client side implementation of creating channel with credentials. The old test_ssl_channel related code is deleted and the new create channel call is used for interop tests. Change on 2014/12/19 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82540921
10 years ago
endif
endif
Add ServiceAccount Credentials wrapping and handle credentials creation failure. Change on 2015/01/09 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83634736
10 years ago
CREDENTIALS_TEST_SRC = \
test/cpp/client/credentials_test.cc \
CREDENTIALS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CREDENTIALS_TEST_SRC))))
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Add ServiceAccount Credentials wrapping and handle credentials creation failure. Change on 2015/01/09 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83634736
10 years ago
bins/$(CONFIG)/credentials_test: openssl_dep_error
else
Fix some C++ test dependencies
10 years ago
bins/$(CONFIG)/credentials_test: $(CREDENTIALS_TEST_OBJS) libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr.a
Add ServiceAccount Credentials wrapping and handle credentials creation failure. Change on 2015/01/09 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83634736
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Fix some C++ test dependencies
10 years ago
$(Q) $(LDXX) $(LDFLAGS) $(CREDENTIALS_TEST_OBJS) $(GTEST_LIB) libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr.a $(LDLIBSXX) $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/credentials_test
Add ServiceAccount Credentials wrapping and handle credentials creation failure. Change on 2015/01/09 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83634736
10 years ago
endif
Fix some C++ test dependencies
10 years ago
objs/$(CONFIG)/test/cpp/client/credentials_test.o: libs/$(CONFIG)/libgrpc++.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr.a
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_credentials_test: $(CREDENTIALS_TEST_OBJS:.o=.dep)
Add ServiceAccount Credentials wrapping and handle credentials creation failure. Change on 2015/01/09 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83634736
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CREDENTIALS_TEST_OBJS:.o=.dep)
Add ServiceAccount Credentials wrapping and handle credentials creation failure. Change on 2015/01/09 by yangg <yangg@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83634736
10 years ago
endif
endif
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
ALARM_TEST_SRC = \
test/core/iomgr/alarm_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
ALARM_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(ALARM_TEST_SRC))))
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/alarm_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/alarm_test: $(ALARM_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(ALARM_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/alarm_test
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/iomgr/alarm_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_alarm_test: $(ALARM_TEST_OBJS:.o=.dep)
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(ALARM_TEST_OBJS:.o=.dep)
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Port [] alarm management to GRPC. This change implements a platform independent alarm manager in alarm.c. It's integrated with iomgr, and some tests are cleaned up. The alarm implementation itself is a fairly direct port of LazyAlarmList from eventmanager. SpinLock has been replaced for now with gpr_mu, and other atomic operations have been dropped (again, for now). A majority of tests have been ported. Change on 2014/12/19 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82551363
10 years ago
ALARM_LIST_TEST_SRC = \
test/core/iomgr/alarm_list_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
ALARM_LIST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(ALARM_LIST_TEST_SRC))))
Port [] alarm management to GRPC. This change implements a platform independent alarm manager in alarm.c. It's integrated with iomgr, and some tests are cleaned up. The alarm implementation itself is a fairly direct port of LazyAlarmList from eventmanager. SpinLock has been replaced for now with gpr_mu, and other atomic operations have been dropped (again, for now). A majority of tests have been ported. Change on 2014/12/19 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82551363
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/alarm_list_test: openssl_dep_error
Port [] alarm management to GRPC. This change implements a platform independent alarm manager in alarm.c. It's integrated with iomgr, and some tests are cleaned up. The alarm implementation itself is a fairly direct port of LazyAlarmList from eventmanager. SpinLock has been replaced for now with gpr_mu, and other atomic operations have been dropped (again, for now). A majority of tests have been ported. Change on 2014/12/19 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82551363
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/alarm_list_test: $(ALARM_LIST_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Port [] alarm management to GRPC. This change implements a platform independent alarm manager in alarm.c. It's integrated with iomgr, and some tests are cleaned up. The alarm implementation itself is a fairly direct port of LazyAlarmList from eventmanager. SpinLock has been replaced for now with gpr_mu, and other atomic operations have been dropped (again, for now). A majority of tests have been ported. Change on 2014/12/19 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82551363
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(ALARM_LIST_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/alarm_list_test
Port [] alarm management to GRPC. This change implements a platform independent alarm manager in alarm.c. It's integrated with iomgr, and some tests are cleaned up. The alarm implementation itself is a fairly direct port of LazyAlarmList from eventmanager. SpinLock has been replaced for now with gpr_mu, and other atomic operations have been dropped (again, for now). A majority of tests have been ported. Change on 2014/12/19 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82551363
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/iomgr/alarm_list_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_alarm_list_test: $(ALARM_LIST_TEST_OBJS:.o=.dep)
Port [] alarm management to GRPC. This change implements a platform independent alarm manager in alarm.c. It's integrated with iomgr, and some tests are cleaned up. The alarm implementation itself is a fairly direct port of LazyAlarmList from eventmanager. SpinLock has been replaced for now with gpr_mu, and other atomic operations have been dropped (again, for now). A majority of tests have been ported. Change on 2014/12/19 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82551363
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(ALARM_LIST_TEST_OBJS:.o=.dep)
Port [] alarm management to GRPC. This change implements a platform independent alarm manager in alarm.c. It's integrated with iomgr, and some tests are cleaned up. The alarm implementation itself is a fairly direct port of LazyAlarmList from eventmanager. SpinLock has been replaced for now with gpr_mu, and other atomic operations have been dropped (again, for now). A majority of tests have been ported. Change on 2014/12/19 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82551363
10 years ago
endif
endif
ALARM_HEAP_TEST_SRC = \
test/core/iomgr/alarm_heap_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
ALARM_HEAP_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(ALARM_HEAP_TEST_SRC))))
Port [] alarm management to GRPC. This change implements a platform independent alarm manager in alarm.c. It's integrated with iomgr, and some tests are cleaned up. The alarm implementation itself is a fairly direct port of LazyAlarmList from eventmanager. SpinLock has been replaced for now with gpr_mu, and other atomic operations have been dropped (again, for now). A majority of tests have been ported. Change on 2014/12/19 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82551363
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/alarm_heap_test: openssl_dep_error
Port [] alarm management to GRPC. This change implements a platform independent alarm manager in alarm.c. It's integrated with iomgr, and some tests are cleaned up. The alarm implementation itself is a fairly direct port of LazyAlarmList from eventmanager. SpinLock has been replaced for now with gpr_mu, and other atomic operations have been dropped (again, for now). A majority of tests have been ported. Change on 2014/12/19 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82551363
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/alarm_heap_test: $(ALARM_HEAP_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Port [] alarm management to GRPC. This change implements a platform independent alarm manager in alarm.c. It's integrated with iomgr, and some tests are cleaned up. The alarm implementation itself is a fairly direct port of LazyAlarmList from eventmanager. SpinLock has been replaced for now with gpr_mu, and other atomic operations have been dropped (again, for now). A majority of tests have been ported. Change on 2014/12/19 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82551363
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(ALARM_HEAP_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/alarm_heap_test
Port [] alarm management to GRPC. This change implements a platform independent alarm manager in alarm.c. It's integrated with iomgr, and some tests are cleaned up. The alarm implementation itself is a fairly direct port of LazyAlarmList from eventmanager. SpinLock has been replaced for now with gpr_mu, and other atomic operations have been dropped (again, for now). A majority of tests have been ported. Change on 2014/12/19 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82551363
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/iomgr/alarm_heap_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_alarm_heap_test: $(ALARM_HEAP_TEST_OBJS:.o=.dep)
Port [] alarm management to GRPC. This change implements a platform independent alarm manager in alarm.c. It's integrated with iomgr, and some tests are cleaned up. The alarm implementation itself is a fairly direct port of LazyAlarmList from eventmanager. SpinLock has been replaced for now with gpr_mu, and other atomic operations have been dropped (again, for now). A majority of tests have been ported. Change on 2014/12/19 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82551363
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(ALARM_HEAP_TEST_OBJS:.o=.dep)
Port [] alarm management to GRPC. This change implements a platform independent alarm manager in alarm.c. It's integrated with iomgr, and some tests are cleaned up. The alarm implementation itself is a fairly direct port of LazyAlarmList from eventmanager. SpinLock has been replaced for now with gpr_mu, and other atomic operations have been dropped (again, for now). A majority of tests have been ported. Change on 2014/12/19 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82551363
10 years ago
endif
endif
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
TIME_TEST_SRC = \
test/core/support/time_test.c \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
TIME_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(TIME_TEST_SRC))))
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/time_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/time_test: $(TIME_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(TIME_TEST_OBJS) libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/time_test
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Merge github.com:google/grpc into testcpp
10 years ago
objs/$(CONFIG)/test/core/support/time_test.o: libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_time_test: $(TIME_TEST_OBJS:.o=.dep)
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(TIME_TEST_OBJS:.o=.dep)
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Reinstate build file validity checks. Also, make those checks pass. Change on 2014/12/10 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81792364
10 years ago
Initial import.
10 years ago
CHTTP2_FAKE_SECURITY_CANCEL_AFTER_ACCEPT_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FAKE_SECURITY_CANCEL_AFTER_ACCEPT_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_CANCEL_AFTER_ACCEPT_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fake_security_cancel_after_accept_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fake_security_cancel_after_accept_test: $(CHTTP2_FAKE_SECURITY_CANCEL_AFTER_ACCEPT_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_cancel_after_accept.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_CANCEL_AFTER_ACCEPT_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_cancel_after_accept.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_cancel_after_accept_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fake_security_cancel_after_accept_test: $(CHTTP2_FAKE_SECURITY_CANCEL_AFTER_ACCEPT_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FAKE_SECURITY_CANCEL_AFTER_ACCEPT_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FAKE_SECURITY_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FAKE_SECURITY_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fake_security_cancel_after_accept_and_writes_closed_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fake_security_cancel_after_accept_and_writes_closed_test: $(CHTTP2_FAKE_SECURITY_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_cancel_after_accept_and_writes_closed.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_cancel_after_accept_and_writes_closed.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_cancel_after_accept_and_writes_closed_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fake_security_cancel_after_accept_and_writes_closed_test: $(CHTTP2_FAKE_SECURITY_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FAKE_SECURITY_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FAKE_SECURITY_CANCEL_AFTER_INVOKE_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FAKE_SECURITY_CANCEL_AFTER_INVOKE_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_CANCEL_AFTER_INVOKE_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fake_security_cancel_after_invoke_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fake_security_cancel_after_invoke_test: $(CHTTP2_FAKE_SECURITY_CANCEL_AFTER_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_cancel_after_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_CANCEL_AFTER_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_cancel_after_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_cancel_after_invoke_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fake_security_cancel_after_invoke_test: $(CHTTP2_FAKE_SECURITY_CANCEL_AFTER_INVOKE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FAKE_SECURITY_CANCEL_AFTER_INVOKE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FAKE_SECURITY_CANCEL_BEFORE_INVOKE_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FAKE_SECURITY_CANCEL_BEFORE_INVOKE_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_CANCEL_BEFORE_INVOKE_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fake_security_cancel_before_invoke_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fake_security_cancel_before_invoke_test: $(CHTTP2_FAKE_SECURITY_CANCEL_BEFORE_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_cancel_before_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_CANCEL_BEFORE_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_cancel_before_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_cancel_before_invoke_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fake_security_cancel_before_invoke_test: $(CHTTP2_FAKE_SECURITY_CANCEL_BEFORE_INVOKE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FAKE_SECURITY_CANCEL_BEFORE_INVOKE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FAKE_SECURITY_CANCEL_IN_A_VACUUM_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FAKE_SECURITY_CANCEL_IN_A_VACUUM_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_CANCEL_IN_A_VACUUM_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fake_security_cancel_in_a_vacuum_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fake_security_cancel_in_a_vacuum_test: $(CHTTP2_FAKE_SECURITY_CANCEL_IN_A_VACUUM_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_cancel_in_a_vacuum.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_CANCEL_IN_A_VACUUM_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_cancel_in_a_vacuum.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_cancel_in_a_vacuum_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fake_security_cancel_in_a_vacuum_test: $(CHTTP2_FAKE_SECURITY_CANCEL_IN_A_VACUUM_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FAKE_SECURITY_CANCEL_IN_A_VACUUM_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
CHTTP2_FAKE_SECURITY_CENSUS_SIMPLE_REQUEST_TEST_SRC = \
CHTTP2_FAKE_SECURITY_CENSUS_SIMPLE_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_CENSUS_SIMPLE_REQUEST_TEST_SRC))))
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
bins/$(CONFIG)/chttp2_fake_security_census_simple_request_test: openssl_dep_error
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fake_security_census_simple_request_test: $(CHTTP2_FAKE_SECURITY_CENSUS_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_census_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_CENSUS_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_census_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_census_simple_request_test
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fake_security_census_simple_request_test: $(CHTTP2_FAKE_SECURITY_CENSUS_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FAKE_SECURITY_CENSUS_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
endif
endif
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
CHTTP2_FAKE_SECURITY_DISAPPEARING_SERVER_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FAKE_SECURITY_DISAPPEARING_SERVER_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_DISAPPEARING_SERVER_TEST_SRC))))
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fake_security_disappearing_server_test: openssl_dep_error
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fake_security_disappearing_server_test: $(CHTTP2_FAKE_SECURITY_DISAPPEARING_SERVER_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_disappearing_server.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_DISAPPEARING_SERVER_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_disappearing_server.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_disappearing_server_test
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fake_security_disappearing_server_test: $(CHTTP2_FAKE_SECURITY_DISAPPEARING_SERVER_TEST_OBJS:.o=.dep)
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FAKE_SECURITY_DISAPPEARING_SERVER_TEST_OBJS:.o=.dep)
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
endif
endif
Initial import.
10 years ago
CHTTP2_FAKE_SECURITY_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FAKE_SECURITY_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fake_security_early_server_shutdown_finishes_inflight_calls_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fake_security_early_server_shutdown_finishes_inflight_calls_test: $(CHTTP2_FAKE_SECURITY_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_inflight_calls.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_inflight_calls.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_early_server_shutdown_finishes_inflight_calls_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fake_security_early_server_shutdown_finishes_inflight_calls_test: $(CHTTP2_FAKE_SECURITY_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FAKE_SECURITY_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FAKE_SECURITY_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FAKE_SECURITY_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fake_security_early_server_shutdown_finishes_tags_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fake_security_early_server_shutdown_finishes_tags_test: $(CHTTP2_FAKE_SECURITY_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_tags.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_tags.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_early_server_shutdown_finishes_tags_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fake_security_early_server_shutdown_finishes_tags_test: $(CHTTP2_FAKE_SECURITY_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FAKE_SECURITY_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
CHTTP2_FAKE_SECURITY_GRACEFUL_SERVER_SHUTDOWN_TEST_SRC = \
CHTTP2_FAKE_SECURITY_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_GRACEFUL_SERVER_SHUTDOWN_TEST_SRC))))
ifeq ($(NO_SECURE),true)
TIPS client implementation.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
bins/$(CONFIG)/chttp2_fake_security_graceful_server_shutdown_test: openssl_dep_error
else
bins/$(CONFIG)/chttp2_fake_security_graceful_server_shutdown_test: $(CHTTP2_FAKE_SECURITY_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_graceful_server_shutdown.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_graceful_server_shutdown.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_graceful_server_shutdown_test
endif
deps_chttp2_fake_security_graceful_server_shutdown_test: $(CHTTP2_FAKE_SECURITY_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS:.o=.dep)
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
-include $(CHTTP2_FAKE_SECURITY_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS:.o=.dep)
endif
endif
Initial import.
10 years ago
CHTTP2_FAKE_SECURITY_INVOKE_LARGE_REQUEST_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FAKE_SECURITY_INVOKE_LARGE_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_INVOKE_LARGE_REQUEST_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fake_security_invoke_large_request_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fake_security_invoke_large_request_test: $(CHTTP2_FAKE_SECURITY_INVOKE_LARGE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_invoke_large_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_INVOKE_LARGE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_invoke_large_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_invoke_large_request_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fake_security_invoke_large_request_test: $(CHTTP2_FAKE_SECURITY_INVOKE_LARGE_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FAKE_SECURITY_INVOKE_LARGE_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FAKE_SECURITY_MAX_CONCURRENT_STREAMS_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FAKE_SECURITY_MAX_CONCURRENT_STREAMS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_MAX_CONCURRENT_STREAMS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fake_security_max_concurrent_streams_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fake_security_max_concurrent_streams_test: $(CHTTP2_FAKE_SECURITY_MAX_CONCURRENT_STREAMS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_max_concurrent_streams.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_MAX_CONCURRENT_STREAMS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_max_concurrent_streams.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_max_concurrent_streams_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fake_security_max_concurrent_streams_test: $(CHTTP2_FAKE_SECURITY_MAX_CONCURRENT_STREAMS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FAKE_SECURITY_MAX_CONCURRENT_STREAMS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FAKE_SECURITY_NO_OP_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FAKE_SECURITY_NO_OP_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_NO_OP_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fake_security_no_op_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fake_security_no_op_test: $(CHTTP2_FAKE_SECURITY_NO_OP_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_no_op.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_NO_OP_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_no_op.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_no_op_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fake_security_no_op_test: $(CHTTP2_FAKE_SECURITY_NO_OP_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FAKE_SECURITY_NO_OP_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FAKE_SECURITY_PING_PONG_STREAMING_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FAKE_SECURITY_PING_PONG_STREAMING_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_PING_PONG_STREAMING_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fake_security_ping_pong_streaming_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fake_security_ping_pong_streaming_test: $(CHTTP2_FAKE_SECURITY_PING_PONG_STREAMING_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_ping_pong_streaming.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_PING_PONG_STREAMING_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_ping_pong_streaming.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_ping_pong_streaming_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fake_security_ping_pong_streaming_test: $(CHTTP2_FAKE_SECURITY_PING_PONG_STREAMING_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FAKE_SECURITY_PING_PONG_STREAMING_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_SRC))))
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fake_security_request_response_with_binary_metadata_and_payload_test: openssl_dep_error
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fake_security_request_response_with_binary_metadata_and_payload_test: $(CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_request_response_with_binary_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_request_response_with_binary_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_request_response_with_binary_metadata_and_payload_test
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fake_security_request_response_with_binary_metadata_and_payload_test: $(CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
endif
endif
Initial import.
10 years ago
CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fake_security_request_response_with_metadata_and_payload_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fake_security_request_response_with_metadata_and_payload_test: $(CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_request_response_with_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_request_response_with_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_request_response_with_metadata_and_payload_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fake_security_request_response_with_metadata_and_payload_test: $(CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fake_security_request_response_with_payload_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fake_security_request_response_with_payload_test: $(CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_request_response_with_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_request_response_with_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_request_response_with_payload_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fake_security_request_response_with_payload_test: $(CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_SRC))))
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fake_security_request_response_with_trailing_metadata_and_payload_test: openssl_dep_error
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fake_security_request_response_with_trailing_metadata_and_payload_test: $(CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_request_response_with_trailing_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_request_response_with_trailing_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_request_response_with_trailing_metadata_and_payload_test
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fake_security_request_response_with_trailing_metadata_and_payload_test: $(CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FAKE_SECURITY_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
endif
endif
Initial import.
10 years ago
CHTTP2_FAKE_SECURITY_SIMPLE_DELAYED_REQUEST_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FAKE_SECURITY_SIMPLE_DELAYED_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_SIMPLE_DELAYED_REQUEST_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fake_security_simple_delayed_request_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fake_security_simple_delayed_request_test: $(CHTTP2_FAKE_SECURITY_SIMPLE_DELAYED_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_simple_delayed_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_SIMPLE_DELAYED_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_simple_delayed_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_simple_delayed_request_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fake_security_simple_delayed_request_test: $(CHTTP2_FAKE_SECURITY_SIMPLE_DELAYED_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FAKE_SECURITY_SIMPLE_DELAYED_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FAKE_SECURITY_SIMPLE_REQUEST_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FAKE_SECURITY_SIMPLE_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_SIMPLE_REQUEST_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fake_security_simple_request_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fake_security_simple_request_test: $(CHTTP2_FAKE_SECURITY_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_simple_request_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fake_security_simple_request_test: $(CHTTP2_FAKE_SECURITY_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FAKE_SECURITY_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Strip "_test" from the name of thread_stress_test.c. Having a BUILD target named chttp2_fullstack_thread_stress_test_test is deceptive as it is a direct test of the system rather than a test of a test of the system. Change on 2014/12/09 by nathaniel <nathaniel@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81688649
10 years ago
CHTTP2_FAKE_SECURITY_THREAD_STRESS_TEST_SRC = \
Initial import.
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FAKE_SECURITY_THREAD_STRESS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_THREAD_STRESS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fake_security_thread_stress_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fake_security_thread_stress_test: $(CHTTP2_FAKE_SECURITY_THREAD_STRESS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_thread_stress.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_THREAD_STRESS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_thread_stress.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_thread_stress_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fake_security_thread_stress_test: $(CHTTP2_FAKE_SECURITY_THREAD_STRESS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FAKE_SECURITY_THREAD_STRESS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FAKE_SECURITY_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FAKE_SECURITY_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FAKE_SECURITY_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fake_security_writes_done_hangs_with_pending_read_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fake_security_writes_done_hangs_with_pending_read_test: $(CHTTP2_FAKE_SECURITY_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_writes_done_hangs_with_pending_read.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FAKE_SECURITY_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fake_security.a libs/$(CONFIG)/libend2end_test_writes_done_hangs_with_pending_read.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fake_security_writes_done_hangs_with_pending_read_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fake_security_writes_done_hangs_with_pending_read_test: $(CHTTP2_FAKE_SECURITY_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FAKE_SECURITY_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FULLSTACK_CANCEL_AFTER_ACCEPT_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FULLSTACK_CANCEL_AFTER_ACCEPT_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_CANCEL_AFTER_ACCEPT_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fullstack_cancel_after_accept_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fullstack_cancel_after_accept_test: $(CHTTP2_FULLSTACK_CANCEL_AFTER_ACCEPT_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_after_accept.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_CANCEL_AFTER_ACCEPT_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_after_accept.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_cancel_after_accept_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fullstack_cancel_after_accept_test: $(CHTTP2_FULLSTACK_CANCEL_AFTER_ACCEPT_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FULLSTACK_CANCEL_AFTER_ACCEPT_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FULLSTACK_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FULLSTACK_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fullstack_cancel_after_accept_and_writes_closed_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fullstack_cancel_after_accept_and_writes_closed_test: $(CHTTP2_FULLSTACK_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_after_accept_and_writes_closed.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_after_accept_and_writes_closed.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_cancel_after_accept_and_writes_closed_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fullstack_cancel_after_accept_and_writes_closed_test: $(CHTTP2_FULLSTACK_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FULLSTACK_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FULLSTACK_CANCEL_AFTER_INVOKE_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FULLSTACK_CANCEL_AFTER_INVOKE_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_CANCEL_AFTER_INVOKE_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fullstack_cancel_after_invoke_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fullstack_cancel_after_invoke_test: $(CHTTP2_FULLSTACK_CANCEL_AFTER_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_after_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_CANCEL_AFTER_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_after_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_cancel_after_invoke_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fullstack_cancel_after_invoke_test: $(CHTTP2_FULLSTACK_CANCEL_AFTER_INVOKE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FULLSTACK_CANCEL_AFTER_INVOKE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FULLSTACK_CANCEL_BEFORE_INVOKE_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FULLSTACK_CANCEL_BEFORE_INVOKE_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_CANCEL_BEFORE_INVOKE_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fullstack_cancel_before_invoke_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fullstack_cancel_before_invoke_test: $(CHTTP2_FULLSTACK_CANCEL_BEFORE_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_before_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_CANCEL_BEFORE_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_before_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_cancel_before_invoke_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fullstack_cancel_before_invoke_test: $(CHTTP2_FULLSTACK_CANCEL_BEFORE_INVOKE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FULLSTACK_CANCEL_BEFORE_INVOKE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FULLSTACK_CANCEL_IN_A_VACUUM_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FULLSTACK_CANCEL_IN_A_VACUUM_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_CANCEL_IN_A_VACUUM_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fullstack_cancel_in_a_vacuum_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fullstack_cancel_in_a_vacuum_test: $(CHTTP2_FULLSTACK_CANCEL_IN_A_VACUUM_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_in_a_vacuum.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_CANCEL_IN_A_VACUUM_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_in_a_vacuum.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_cancel_in_a_vacuum_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fullstack_cancel_in_a_vacuum_test: $(CHTTP2_FULLSTACK_CANCEL_IN_A_VACUUM_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FULLSTACK_CANCEL_IN_A_VACUUM_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
CHTTP2_FULLSTACK_CENSUS_SIMPLE_REQUEST_TEST_SRC = \
CHTTP2_FULLSTACK_CENSUS_SIMPLE_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_CENSUS_SIMPLE_REQUEST_TEST_SRC))))
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
bins/$(CONFIG)/chttp2_fullstack_census_simple_request_test: openssl_dep_error
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fullstack_census_simple_request_test: $(CHTTP2_FULLSTACK_CENSUS_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_census_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_CENSUS_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_census_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_census_simple_request_test
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fullstack_census_simple_request_test: $(CHTTP2_FULLSTACK_CENSUS_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FULLSTACK_CENSUS_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
endif
endif
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
CHTTP2_FULLSTACK_DISAPPEARING_SERVER_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FULLSTACK_DISAPPEARING_SERVER_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_DISAPPEARING_SERVER_TEST_SRC))))
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fullstack_disappearing_server_test: openssl_dep_error
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fullstack_disappearing_server_test: $(CHTTP2_FULLSTACK_DISAPPEARING_SERVER_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_disappearing_server.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_DISAPPEARING_SERVER_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_disappearing_server.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_disappearing_server_test
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fullstack_disappearing_server_test: $(CHTTP2_FULLSTACK_DISAPPEARING_SERVER_TEST_OBJS:.o=.dep)
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FULLSTACK_DISAPPEARING_SERVER_TEST_OBJS:.o=.dep)
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
endif
endif
Initial import.
10 years ago
CHTTP2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fullstack_early_server_shutdown_finishes_inflight_calls_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fullstack_early_server_shutdown_finishes_inflight_calls_test: $(CHTTP2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_inflight_calls.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_inflight_calls.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_early_server_shutdown_finishes_inflight_calls_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fullstack_early_server_shutdown_finishes_inflight_calls_test: $(CHTTP2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fullstack_early_server_shutdown_finishes_tags_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fullstack_early_server_shutdown_finishes_tags_test: $(CHTTP2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_tags.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_tags.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_early_server_shutdown_finishes_tags_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fullstack_early_server_shutdown_finishes_tags_test: $(CHTTP2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
CHTTP2_FULLSTACK_GRACEFUL_SERVER_SHUTDOWN_TEST_SRC = \
CHTTP2_FULLSTACK_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_GRACEFUL_SERVER_SHUTDOWN_TEST_SRC))))
ifeq ($(NO_SECURE),true)
TIPS client implementation.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
bins/$(CONFIG)/chttp2_fullstack_graceful_server_shutdown_test: openssl_dep_error
else
bins/$(CONFIG)/chttp2_fullstack_graceful_server_shutdown_test: $(CHTTP2_FULLSTACK_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_graceful_server_shutdown.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_graceful_server_shutdown.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_graceful_server_shutdown_test
endif
deps_chttp2_fullstack_graceful_server_shutdown_test: $(CHTTP2_FULLSTACK_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS:.o=.dep)
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
-include $(CHTTP2_FULLSTACK_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS:.o=.dep)
endif
endif
Initial import.
10 years ago
CHTTP2_FULLSTACK_INVOKE_LARGE_REQUEST_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FULLSTACK_INVOKE_LARGE_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_INVOKE_LARGE_REQUEST_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fullstack_invoke_large_request_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fullstack_invoke_large_request_test: $(CHTTP2_FULLSTACK_INVOKE_LARGE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_invoke_large_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_INVOKE_LARGE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_invoke_large_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_invoke_large_request_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fullstack_invoke_large_request_test: $(CHTTP2_FULLSTACK_INVOKE_LARGE_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FULLSTACK_INVOKE_LARGE_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FULLSTACK_MAX_CONCURRENT_STREAMS_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FULLSTACK_MAX_CONCURRENT_STREAMS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_MAX_CONCURRENT_STREAMS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fullstack_max_concurrent_streams_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fullstack_max_concurrent_streams_test: $(CHTTP2_FULLSTACK_MAX_CONCURRENT_STREAMS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_max_concurrent_streams.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_MAX_CONCURRENT_STREAMS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_max_concurrent_streams.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_max_concurrent_streams_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fullstack_max_concurrent_streams_test: $(CHTTP2_FULLSTACK_MAX_CONCURRENT_STREAMS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FULLSTACK_MAX_CONCURRENT_STREAMS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FULLSTACK_NO_OP_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FULLSTACK_NO_OP_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_NO_OP_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fullstack_no_op_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fullstack_no_op_test: $(CHTTP2_FULLSTACK_NO_OP_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_no_op.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_NO_OP_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_no_op.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_no_op_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fullstack_no_op_test: $(CHTTP2_FULLSTACK_NO_OP_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FULLSTACK_NO_OP_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FULLSTACK_PING_PONG_STREAMING_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FULLSTACK_PING_PONG_STREAMING_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_PING_PONG_STREAMING_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fullstack_ping_pong_streaming_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fullstack_ping_pong_streaming_test: $(CHTTP2_FULLSTACK_PING_PONG_STREAMING_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_ping_pong_streaming.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_PING_PONG_STREAMING_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_ping_pong_streaming.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_ping_pong_streaming_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fullstack_ping_pong_streaming_test: $(CHTTP2_FULLSTACK_PING_PONG_STREAMING_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FULLSTACK_PING_PONG_STREAMING_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_SRC))))
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fullstack_request_response_with_binary_metadata_and_payload_test: openssl_dep_error
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fullstack_request_response_with_binary_metadata_and_payload_test: $(CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_binary_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_binary_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_request_response_with_binary_metadata_and_payload_test
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fullstack_request_response_with_binary_metadata_and_payload_test: $(CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
endif
endif
Initial import.
10 years ago
CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fullstack_request_response_with_metadata_and_payload_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fullstack_request_response_with_metadata_and_payload_test: $(CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_request_response_with_metadata_and_payload_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fullstack_request_response_with_metadata_and_payload_test: $(CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fullstack_request_response_with_payload_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fullstack_request_response_with_payload_test: $(CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_request_response_with_payload_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fullstack_request_response_with_payload_test: $(CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_SRC))))
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fullstack_request_response_with_trailing_metadata_and_payload_test: openssl_dep_error
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fullstack_request_response_with_trailing_metadata_and_payload_test: $(CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_trailing_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_trailing_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_request_response_with_trailing_metadata_and_payload_test
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fullstack_request_response_with_trailing_metadata_and_payload_test: $(CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FULLSTACK_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
endif
endif
Initial import.
10 years ago
CHTTP2_FULLSTACK_SIMPLE_DELAYED_REQUEST_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FULLSTACK_SIMPLE_DELAYED_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_SIMPLE_DELAYED_REQUEST_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fullstack_simple_delayed_request_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fullstack_simple_delayed_request_test: $(CHTTP2_FULLSTACK_SIMPLE_DELAYED_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_simple_delayed_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_SIMPLE_DELAYED_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_simple_delayed_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_simple_delayed_request_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fullstack_simple_delayed_request_test: $(CHTTP2_FULLSTACK_SIMPLE_DELAYED_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FULLSTACK_SIMPLE_DELAYED_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FULLSTACK_SIMPLE_REQUEST_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FULLSTACK_SIMPLE_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_SIMPLE_REQUEST_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fullstack_simple_request_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fullstack_simple_request_test: $(CHTTP2_FULLSTACK_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_simple_request_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fullstack_simple_request_test: $(CHTTP2_FULLSTACK_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FULLSTACK_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Strip "_test" from the name of thread_stress_test.c. Having a BUILD target named chttp2_fullstack_thread_stress_test_test is deceptive as it is a direct test of the system rather than a test of a test of the system. Change on 2014/12/09 by nathaniel <nathaniel@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81688649
10 years ago
CHTTP2_FULLSTACK_THREAD_STRESS_TEST_SRC = \
Initial import.
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FULLSTACK_THREAD_STRESS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_THREAD_STRESS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fullstack_thread_stress_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fullstack_thread_stress_test: $(CHTTP2_FULLSTACK_THREAD_STRESS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_thread_stress.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_THREAD_STRESS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_thread_stress.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_thread_stress_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fullstack_thread_stress_test: $(CHTTP2_FULLSTACK_THREAD_STRESS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FULLSTACK_THREAD_STRESS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_FULLSTACK_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_FULLSTACK_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_FULLSTACK_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_fullstack_writes_done_hangs_with_pending_read_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_fullstack_writes_done_hangs_with_pending_read_test: $(CHTTP2_FULLSTACK_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_writes_done_hangs_with_pending_read.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_FULLSTACK_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_fullstack.a libs/$(CONFIG)/libend2end_test_writes_done_hangs_with_pending_read.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_fullstack_writes_done_hangs_with_pending_read_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_fullstack_writes_done_hangs_with_pending_read_test: $(CHTTP2_FULLSTACK_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_FULLSTACK_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_AFTER_ACCEPT_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_AFTER_ACCEPT_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_AFTER_ACCEPT_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_after_accept_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_after_accept_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_AFTER_ACCEPT_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_after_accept.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_AFTER_ACCEPT_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_after_accept.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_after_accept_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_fullstack_cancel_after_accept_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_AFTER_ACCEPT_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_AFTER_ACCEPT_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_after_accept_and_writes_closed_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_after_accept_and_writes_closed_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_after_accept_and_writes_closed.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_after_accept_and_writes_closed.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_after_accept_and_writes_closed_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_fullstack_cancel_after_accept_and_writes_closed_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_AFTER_INVOKE_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_AFTER_INVOKE_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_AFTER_INVOKE_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_after_invoke_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_after_invoke_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_AFTER_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_after_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_AFTER_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_after_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_after_invoke_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_fullstack_cancel_after_invoke_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_AFTER_INVOKE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_AFTER_INVOKE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_BEFORE_INVOKE_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_BEFORE_INVOKE_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_BEFORE_INVOKE_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_before_invoke_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_before_invoke_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_BEFORE_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_before_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_BEFORE_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_before_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_before_invoke_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_fullstack_cancel_before_invoke_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_BEFORE_INVOKE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_BEFORE_INVOKE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_IN_A_VACUUM_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_IN_A_VACUUM_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_IN_A_VACUUM_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_in_a_vacuum_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_in_a_vacuum_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_IN_A_VACUUM_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_in_a_vacuum.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_IN_A_VACUUM_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_in_a_vacuum.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_cancel_in_a_vacuum_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_fullstack_cancel_in_a_vacuum_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_IN_A_VACUUM_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_CANCEL_IN_A_VACUUM_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_CENSUS_SIMPLE_REQUEST_TEST_SRC = \
CHTTP2_SIMPLE_SSL_FULLSTACK_CENSUS_SIMPLE_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_CENSUS_SIMPLE_REQUEST_TEST_SRC))))
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_census_simple_request_test: openssl_dep_error
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_census_simple_request_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_CENSUS_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_census_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_CENSUS_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_census_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_census_simple_request_test
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_fullstack_census_simple_request_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_CENSUS_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_CENSUS_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
endif
endif
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_DISAPPEARING_SERVER_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_DISAPPEARING_SERVER_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_DISAPPEARING_SERVER_TEST_SRC))))
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_disappearing_server_test: openssl_dep_error
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_disappearing_server_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_DISAPPEARING_SERVER_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_disappearing_server.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_DISAPPEARING_SERVER_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_disappearing_server.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_disappearing_server_test
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_fullstack_disappearing_server_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_DISAPPEARING_SERVER_TEST_OBJS:.o=.dep)
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_DISAPPEARING_SERVER_TEST_OBJS:.o=.dep)
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
endif
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_early_server_shutdown_finishes_inflight_calls_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_early_server_shutdown_finishes_inflight_calls_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_inflight_calls.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_inflight_calls.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_early_server_shutdown_finishes_inflight_calls_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_fullstack_early_server_shutdown_finishes_inflight_calls_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_early_server_shutdown_finishes_tags_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_early_server_shutdown_finishes_tags_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_tags.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_tags.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_early_server_shutdown_finishes_tags_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_fullstack_early_server_shutdown_finishes_tags_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_GRACEFUL_SERVER_SHUTDOWN_TEST_SRC = \
CHTTP2_SIMPLE_SSL_FULLSTACK_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_GRACEFUL_SERVER_SHUTDOWN_TEST_SRC))))
ifeq ($(NO_SECURE),true)
TIPS client implementation.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_graceful_server_shutdown_test: openssl_dep_error
else
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_graceful_server_shutdown_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_graceful_server_shutdown.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_graceful_server_shutdown.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_graceful_server_shutdown_test
endif
deps_chttp2_simple_ssl_fullstack_graceful_server_shutdown_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS:.o=.dep)
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS:.o=.dep)
endif
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_INVOKE_LARGE_REQUEST_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_INVOKE_LARGE_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_INVOKE_LARGE_REQUEST_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_invoke_large_request_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_invoke_large_request_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_INVOKE_LARGE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_invoke_large_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_INVOKE_LARGE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_invoke_large_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_invoke_large_request_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_fullstack_invoke_large_request_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_INVOKE_LARGE_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_INVOKE_LARGE_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_MAX_CONCURRENT_STREAMS_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_MAX_CONCURRENT_STREAMS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_MAX_CONCURRENT_STREAMS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_max_concurrent_streams_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_max_concurrent_streams_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_MAX_CONCURRENT_STREAMS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_max_concurrent_streams.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_MAX_CONCURRENT_STREAMS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_max_concurrent_streams.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_max_concurrent_streams_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_fullstack_max_concurrent_streams_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_MAX_CONCURRENT_STREAMS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_MAX_CONCURRENT_STREAMS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_NO_OP_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_NO_OP_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_NO_OP_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_no_op_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_no_op_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_NO_OP_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_no_op.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_NO_OP_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_no_op.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_no_op_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_fullstack_no_op_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_NO_OP_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_NO_OP_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_PING_PONG_STREAMING_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_PING_PONG_STREAMING_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_PING_PONG_STREAMING_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_ping_pong_streaming_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_ping_pong_streaming_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_PING_PONG_STREAMING_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_ping_pong_streaming.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_PING_PONG_STREAMING_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_ping_pong_streaming.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_ping_pong_streaming_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_fullstack_ping_pong_streaming_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_PING_PONG_STREAMING_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_PING_PONG_STREAMING_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_SRC))))
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_request_response_with_binary_metadata_and_payload_test: openssl_dep_error
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_request_response_with_binary_metadata_and_payload_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_binary_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_binary_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_request_response_with_binary_metadata_and_payload_test
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_fullstack_request_response_with_binary_metadata_and_payload_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
endif
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_request_response_with_metadata_and_payload_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_request_response_with_metadata_and_payload_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_request_response_with_metadata_and_payload_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_fullstack_request_response_with_metadata_and_payload_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_request_response_with_payload_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_request_response_with_payload_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_request_response_with_payload_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_fullstack_request_response_with_payload_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_SRC))))
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_request_response_with_trailing_metadata_and_payload_test: openssl_dep_error
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_request_response_with_trailing_metadata_and_payload_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_trailing_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_trailing_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_request_response_with_trailing_metadata_and_payload_test
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_fullstack_request_response_with_trailing_metadata_and_payload_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
endif
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_SIMPLE_DELAYED_REQUEST_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_SIMPLE_DELAYED_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_SIMPLE_DELAYED_REQUEST_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_simple_delayed_request_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_simple_delayed_request_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_SIMPLE_DELAYED_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_simple_delayed_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_SIMPLE_DELAYED_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_simple_delayed_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_simple_delayed_request_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_fullstack_simple_delayed_request_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_SIMPLE_DELAYED_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_SIMPLE_DELAYED_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_SIMPLE_REQUEST_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_SIMPLE_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_SIMPLE_REQUEST_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_simple_request_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_simple_request_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_simple_request_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_fullstack_simple_request_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Strip "_test" from the name of thread_stress_test.c. Having a BUILD target named chttp2_fullstack_thread_stress_test_test is deceptive as it is a direct test of the system rather than a test of a test of the system. Change on 2014/12/09 by nathaniel <nathaniel@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81688649
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_THREAD_STRESS_TEST_SRC = \
Initial import.
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_THREAD_STRESS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_THREAD_STRESS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_thread_stress_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_thread_stress_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_THREAD_STRESS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_thread_stress.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_THREAD_STRESS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_thread_stress.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_thread_stress_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_fullstack_thread_stress_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_THREAD_STRESS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_THREAD_STRESS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_FULLSTACK_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_FULLSTACK_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_writes_done_hangs_with_pending_read_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_fullstack_writes_done_hangs_with_pending_read_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_writes_done_hangs_with_pending_read.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_FULLSTACK_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_fullstack.a libs/$(CONFIG)/libend2end_test_writes_done_hangs_with_pending_read.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_fullstack_writes_done_hangs_with_pending_read_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_fullstack_writes_done_hangs_with_pending_read_test: $(CHTTP2_SIMPLE_SSL_FULLSTACK_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_FULLSTACK_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_AFTER_ACCEPT_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_AFTER_ACCEPT_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_AFTER_ACCEPT_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_accept_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_accept_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_AFTER_ACCEPT_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_after_accept.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_AFTER_ACCEPT_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_after_accept.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_accept_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_accept_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_AFTER_ACCEPT_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_AFTER_ACCEPT_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_accept_and_writes_closed_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_accept_and_writes_closed_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_after_accept_and_writes_closed.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_after_accept_and_writes_closed.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_accept_and_writes_closed_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_accept_and_writes_closed_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_AFTER_INVOKE_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_AFTER_INVOKE_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_AFTER_INVOKE_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_invoke_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_invoke_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_AFTER_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_after_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_AFTER_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_after_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_invoke_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_with_oauth2_fullstack_cancel_after_invoke_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_AFTER_INVOKE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_AFTER_INVOKE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_BEFORE_INVOKE_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_BEFORE_INVOKE_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_BEFORE_INVOKE_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_before_invoke_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_before_invoke_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_BEFORE_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_before_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_BEFORE_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_before_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_before_invoke_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_with_oauth2_fullstack_cancel_before_invoke_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_BEFORE_INVOKE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_BEFORE_INVOKE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_IN_A_VACUUM_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_IN_A_VACUUM_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_IN_A_VACUUM_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_in_a_vacuum_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_in_a_vacuum_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_IN_A_VACUUM_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_in_a_vacuum.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_IN_A_VACUUM_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_cancel_in_a_vacuum.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_cancel_in_a_vacuum_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_with_oauth2_fullstack_cancel_in_a_vacuum_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_IN_A_VACUUM_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CANCEL_IN_A_VACUUM_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CENSUS_SIMPLE_REQUEST_TEST_SRC = \
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CENSUS_SIMPLE_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CENSUS_SIMPLE_REQUEST_TEST_SRC))))
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_census_simple_request_test: openssl_dep_error
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_census_simple_request_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CENSUS_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_census_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CENSUS_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_census_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_census_simple_request_test
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_with_oauth2_fullstack_census_simple_request_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CENSUS_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_CENSUS_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
endif
endif
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_DISAPPEARING_SERVER_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_DISAPPEARING_SERVER_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_DISAPPEARING_SERVER_TEST_SRC))))
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_disappearing_server_test: openssl_dep_error
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_disappearing_server_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_DISAPPEARING_SERVER_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_disappearing_server.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_DISAPPEARING_SERVER_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_disappearing_server.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_disappearing_server_test
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_with_oauth2_fullstack_disappearing_server_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_DISAPPEARING_SERVER_TEST_OBJS:.o=.dep)
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_DISAPPEARING_SERVER_TEST_OBJS:.o=.dep)
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
endif
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_early_server_shutdown_finishes_inflight_calls_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_early_server_shutdown_finishes_inflight_calls_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_inflight_calls.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_inflight_calls.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_early_server_shutdown_finishes_inflight_calls_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_with_oauth2_fullstack_early_server_shutdown_finishes_inflight_calls_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_early_server_shutdown_finishes_tags_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_early_server_shutdown_finishes_tags_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_tags.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_tags.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_early_server_shutdown_finishes_tags_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_with_oauth2_fullstack_early_server_shutdown_finishes_tags_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_GRACEFUL_SERVER_SHUTDOWN_TEST_SRC = \
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_GRACEFUL_SERVER_SHUTDOWN_TEST_SRC))))
ifeq ($(NO_SECURE),true)
TIPS client implementation.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_graceful_server_shutdown_test: openssl_dep_error
else
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_graceful_server_shutdown_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_graceful_server_shutdown.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_graceful_server_shutdown.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_graceful_server_shutdown_test
endif
deps_chttp2_simple_ssl_with_oauth2_fullstack_graceful_server_shutdown_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS:.o=.dep)
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS:.o=.dep)
endif
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_INVOKE_LARGE_REQUEST_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_INVOKE_LARGE_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_INVOKE_LARGE_REQUEST_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_invoke_large_request_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_invoke_large_request_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_INVOKE_LARGE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_invoke_large_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_INVOKE_LARGE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_invoke_large_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_invoke_large_request_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_with_oauth2_fullstack_invoke_large_request_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_INVOKE_LARGE_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_INVOKE_LARGE_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_MAX_CONCURRENT_STREAMS_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_MAX_CONCURRENT_STREAMS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_MAX_CONCURRENT_STREAMS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_max_concurrent_streams_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_max_concurrent_streams_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_MAX_CONCURRENT_STREAMS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_max_concurrent_streams.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_MAX_CONCURRENT_STREAMS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_max_concurrent_streams.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_max_concurrent_streams_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_with_oauth2_fullstack_max_concurrent_streams_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_MAX_CONCURRENT_STREAMS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_MAX_CONCURRENT_STREAMS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_NO_OP_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_NO_OP_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_NO_OP_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_no_op_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_no_op_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_NO_OP_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_no_op.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_NO_OP_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_no_op.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_no_op_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_with_oauth2_fullstack_no_op_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_NO_OP_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_NO_OP_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_PING_PONG_STREAMING_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_PING_PONG_STREAMING_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_PING_PONG_STREAMING_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_ping_pong_streaming_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_ping_pong_streaming_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_PING_PONG_STREAMING_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_ping_pong_streaming.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_PING_PONG_STREAMING_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_ping_pong_streaming.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_ping_pong_streaming_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_with_oauth2_fullstack_ping_pong_streaming_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_PING_PONG_STREAMING_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_PING_PONG_STREAMING_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_SRC))))
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_binary_metadata_and_payload_test: openssl_dep_error
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_binary_metadata_and_payload_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_binary_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_binary_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_binary_metadata_and_payload_test
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_binary_metadata_and_payload_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
endif
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_metadata_and_payload_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_metadata_and_payload_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_metadata_and_payload_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_metadata_and_payload_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_payload_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_payload_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_payload_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_payload_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_SRC))))
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_trailing_metadata_and_payload_test: openssl_dep_error
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_trailing_metadata_and_payload_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_trailing_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_request_response_with_trailing_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_trailing_metadata_and_payload_test
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_with_oauth2_fullstack_request_response_with_trailing_metadata_and_payload_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
endif
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_SIMPLE_DELAYED_REQUEST_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_SIMPLE_DELAYED_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_SIMPLE_DELAYED_REQUEST_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_simple_delayed_request_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_simple_delayed_request_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_SIMPLE_DELAYED_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_simple_delayed_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_SIMPLE_DELAYED_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_simple_delayed_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_simple_delayed_request_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_with_oauth2_fullstack_simple_delayed_request_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_SIMPLE_DELAYED_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_SIMPLE_DELAYED_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_SIMPLE_REQUEST_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_SIMPLE_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_SIMPLE_REQUEST_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_simple_request_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_simple_request_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_simple_request_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_with_oauth2_fullstack_simple_request_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Strip "_test" from the name of thread_stress_test.c. Having a BUILD target named chttp2_fullstack_thread_stress_test_test is deceptive as it is a direct test of the system rather than a test of a test of the system. Change on 2014/12/09 by nathaniel <nathaniel@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81688649
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_THREAD_STRESS_TEST_SRC = \
Initial import.
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_THREAD_STRESS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_THREAD_STRESS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_thread_stress_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_thread_stress_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_THREAD_STRESS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_thread_stress.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_THREAD_STRESS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_thread_stress.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_thread_stress_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_with_oauth2_fullstack_thread_stress_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_THREAD_STRESS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_THREAD_STRESS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_writes_done_hangs_with_pending_read_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_writes_done_hangs_with_pending_read_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_writes_done_hangs_with_pending_read.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_simple_ssl_with_oauth2_fullstack.a libs/$(CONFIG)/libend2end_test_writes_done_hangs_with_pending_read.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_simple_ssl_with_oauth2_fullstack_writes_done_hangs_with_pending_read_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_simple_ssl_with_oauth2_fullstack_writes_done_hangs_with_pending_read_test: $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SIMPLE_SSL_WITH_OAUTH2_FULLSTACK_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SOCKET_PAIR_CANCEL_AFTER_ACCEPT_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_CANCEL_AFTER_ACCEPT_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_CANCEL_AFTER_ACCEPT_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_cancel_after_accept_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_cancel_after_accept_test: $(CHTTP2_SOCKET_PAIR_CANCEL_AFTER_ACCEPT_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_cancel_after_accept.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_CANCEL_AFTER_ACCEPT_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_cancel_after_accept.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_cancel_after_accept_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_cancel_after_accept_test: $(CHTTP2_SOCKET_PAIR_CANCEL_AFTER_ACCEPT_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_CANCEL_AFTER_ACCEPT_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SOCKET_PAIR_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_cancel_after_accept_and_writes_closed_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_cancel_after_accept_and_writes_closed_test: $(CHTTP2_SOCKET_PAIR_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_cancel_after_accept_and_writes_closed.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_cancel_after_accept_and_writes_closed.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_cancel_after_accept_and_writes_closed_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_cancel_after_accept_and_writes_closed_test: $(CHTTP2_SOCKET_PAIR_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SOCKET_PAIR_CANCEL_AFTER_INVOKE_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_CANCEL_AFTER_INVOKE_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_CANCEL_AFTER_INVOKE_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_cancel_after_invoke_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_cancel_after_invoke_test: $(CHTTP2_SOCKET_PAIR_CANCEL_AFTER_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_cancel_after_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_CANCEL_AFTER_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_cancel_after_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_cancel_after_invoke_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_cancel_after_invoke_test: $(CHTTP2_SOCKET_PAIR_CANCEL_AFTER_INVOKE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_CANCEL_AFTER_INVOKE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SOCKET_PAIR_CANCEL_BEFORE_INVOKE_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_CANCEL_BEFORE_INVOKE_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_CANCEL_BEFORE_INVOKE_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_cancel_before_invoke_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_cancel_before_invoke_test: $(CHTTP2_SOCKET_PAIR_CANCEL_BEFORE_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_cancel_before_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_CANCEL_BEFORE_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_cancel_before_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_cancel_before_invoke_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_cancel_before_invoke_test: $(CHTTP2_SOCKET_PAIR_CANCEL_BEFORE_INVOKE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_CANCEL_BEFORE_INVOKE_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SOCKET_PAIR_CANCEL_IN_A_VACUUM_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_CANCEL_IN_A_VACUUM_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_CANCEL_IN_A_VACUUM_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_cancel_in_a_vacuum_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_cancel_in_a_vacuum_test: $(CHTTP2_SOCKET_PAIR_CANCEL_IN_A_VACUUM_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_cancel_in_a_vacuum.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_CANCEL_IN_A_VACUUM_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_cancel_in_a_vacuum.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_cancel_in_a_vacuum_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_cancel_in_a_vacuum_test: $(CHTTP2_SOCKET_PAIR_CANCEL_IN_A_VACUUM_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_CANCEL_IN_A_VACUUM_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
CHTTP2_SOCKET_PAIR_CENSUS_SIMPLE_REQUEST_TEST_SRC = \
CHTTP2_SOCKET_PAIR_CENSUS_SIMPLE_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_CENSUS_SIMPLE_REQUEST_TEST_SRC))))
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_census_simple_request_test: openssl_dep_error
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_census_simple_request_test: $(CHTTP2_SOCKET_PAIR_CENSUS_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_census_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_CENSUS_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_census_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_census_simple_request_test
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_census_simple_request_test: $(CHTTP2_SOCKET_PAIR_CENSUS_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_CENSUS_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
endif
endif
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
CHTTP2_SOCKET_PAIR_DISAPPEARING_SERVER_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_DISAPPEARING_SERVER_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_DISAPPEARING_SERVER_TEST_SRC))))
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_disappearing_server_test: openssl_dep_error
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_disappearing_server_test: $(CHTTP2_SOCKET_PAIR_DISAPPEARING_SERVER_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_disappearing_server.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_DISAPPEARING_SERVER_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_disappearing_server.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_disappearing_server_test
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_disappearing_server_test: $(CHTTP2_SOCKET_PAIR_DISAPPEARING_SERVER_TEST_OBJS:.o=.dep)
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_DISAPPEARING_SERVER_TEST_OBJS:.o=.dep)
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
endif
endif
Initial import.
10 years ago
CHTTP2_SOCKET_PAIR_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_early_server_shutdown_finishes_inflight_calls_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_early_server_shutdown_finishes_inflight_calls_test: $(CHTTP2_SOCKET_PAIR_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_inflight_calls.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_inflight_calls.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_early_server_shutdown_finishes_inflight_calls_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_early_server_shutdown_finishes_inflight_calls_test: $(CHTTP2_SOCKET_PAIR_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SOCKET_PAIR_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_early_server_shutdown_finishes_tags_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_early_server_shutdown_finishes_tags_test: $(CHTTP2_SOCKET_PAIR_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_tags.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_tags.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_early_server_shutdown_finishes_tags_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_early_server_shutdown_finishes_tags_test: $(CHTTP2_SOCKET_PAIR_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
CHTTP2_SOCKET_PAIR_GRACEFUL_SERVER_SHUTDOWN_TEST_SRC = \
CHTTP2_SOCKET_PAIR_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_GRACEFUL_SERVER_SHUTDOWN_TEST_SRC))))
ifeq ($(NO_SECURE),true)
TIPS client implementation.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_graceful_server_shutdown_test: openssl_dep_error
else
bins/$(CONFIG)/chttp2_socket_pair_graceful_server_shutdown_test: $(CHTTP2_SOCKET_PAIR_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_graceful_server_shutdown.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_graceful_server_shutdown.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_graceful_server_shutdown_test
endif
deps_chttp2_socket_pair_graceful_server_shutdown_test: $(CHTTP2_SOCKET_PAIR_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS:.o=.dep)
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
-include $(CHTTP2_SOCKET_PAIR_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS:.o=.dep)
endif
endif
Initial import.
10 years ago
CHTTP2_SOCKET_PAIR_INVOKE_LARGE_REQUEST_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_INVOKE_LARGE_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_INVOKE_LARGE_REQUEST_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_invoke_large_request_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_invoke_large_request_test: $(CHTTP2_SOCKET_PAIR_INVOKE_LARGE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_invoke_large_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_INVOKE_LARGE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_invoke_large_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_invoke_large_request_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_invoke_large_request_test: $(CHTTP2_SOCKET_PAIR_INVOKE_LARGE_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_INVOKE_LARGE_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SOCKET_PAIR_MAX_CONCURRENT_STREAMS_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_MAX_CONCURRENT_STREAMS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_MAX_CONCURRENT_STREAMS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_max_concurrent_streams_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_max_concurrent_streams_test: $(CHTTP2_SOCKET_PAIR_MAX_CONCURRENT_STREAMS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_max_concurrent_streams.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_MAX_CONCURRENT_STREAMS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_max_concurrent_streams.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_max_concurrent_streams_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_max_concurrent_streams_test: $(CHTTP2_SOCKET_PAIR_MAX_CONCURRENT_STREAMS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_MAX_CONCURRENT_STREAMS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SOCKET_PAIR_NO_OP_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_NO_OP_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_NO_OP_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_no_op_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_no_op_test: $(CHTTP2_SOCKET_PAIR_NO_OP_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_no_op.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_NO_OP_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_no_op.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_no_op_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_no_op_test: $(CHTTP2_SOCKET_PAIR_NO_OP_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_NO_OP_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SOCKET_PAIR_PING_PONG_STREAMING_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_PING_PONG_STREAMING_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_PING_PONG_STREAMING_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_ping_pong_streaming_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_ping_pong_streaming_test: $(CHTTP2_SOCKET_PAIR_PING_PONG_STREAMING_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_ping_pong_streaming.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_PING_PONG_STREAMING_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_ping_pong_streaming.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_ping_pong_streaming_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_ping_pong_streaming_test: $(CHTTP2_SOCKET_PAIR_PING_PONG_STREAMING_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_PING_PONG_STREAMING_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_SRC))))
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_request_response_with_binary_metadata_and_payload_test: openssl_dep_error
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_request_response_with_binary_metadata_and_payload_test: $(CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_request_response_with_binary_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_request_response_with_binary_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_request_response_with_binary_metadata_and_payload_test
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_request_response_with_binary_metadata_and_payload_test: $(CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
endif
endif
Initial import.
10 years ago
CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_request_response_with_metadata_and_payload_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_request_response_with_metadata_and_payload_test: $(CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_request_response_with_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_request_response_with_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_request_response_with_metadata_and_payload_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_request_response_with_metadata_and_payload_test: $(CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_request_response_with_payload_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_request_response_with_payload_test: $(CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_request_response_with_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_request_response_with_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_request_response_with_payload_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_request_response_with_payload_test: $(CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_SRC))))
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_request_response_with_trailing_metadata_and_payload_test: openssl_dep_error
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_request_response_with_trailing_metadata_and_payload_test: $(CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_request_response_with_trailing_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_request_response_with_trailing_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_request_response_with_trailing_metadata_and_payload_test
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_request_response_with_trailing_metadata_and_payload_test: $(CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
endif
endif
Initial import.
10 years ago
CHTTP2_SOCKET_PAIR_SIMPLE_DELAYED_REQUEST_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_SIMPLE_DELAYED_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_SIMPLE_DELAYED_REQUEST_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_simple_delayed_request_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_simple_delayed_request_test: $(CHTTP2_SOCKET_PAIR_SIMPLE_DELAYED_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_simple_delayed_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_SIMPLE_DELAYED_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_simple_delayed_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_simple_delayed_request_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_simple_delayed_request_test: $(CHTTP2_SOCKET_PAIR_SIMPLE_DELAYED_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_SIMPLE_DELAYED_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SOCKET_PAIR_SIMPLE_REQUEST_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_SIMPLE_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_SIMPLE_REQUEST_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_simple_request_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_simple_request_test: $(CHTTP2_SOCKET_PAIR_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_simple_request_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_simple_request_test: $(CHTTP2_SOCKET_PAIR_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Strip "_test" from the name of thread_stress_test.c. Having a BUILD target named chttp2_fullstack_thread_stress_test_test is deceptive as it is a direct test of the system rather than a test of a test of the system. Change on 2014/12/09 by nathaniel <nathaniel@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81688649
10 years ago
CHTTP2_SOCKET_PAIR_THREAD_STRESS_TEST_SRC = \
Initial import.
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_THREAD_STRESS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_THREAD_STRESS_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_thread_stress_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_thread_stress_test: $(CHTTP2_SOCKET_PAIR_THREAD_STRESS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_thread_stress.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_THREAD_STRESS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_thread_stress.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_thread_stress_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_thread_stress_test: $(CHTTP2_SOCKET_PAIR_THREAD_STRESS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_THREAD_STRESS_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
CHTTP2_SOCKET_PAIR_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_SRC))))
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_writes_done_hangs_with_pending_read_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_writes_done_hangs_with_pending_read_test: $(CHTTP2_SOCKET_PAIR_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_writes_done_hangs_with_pending_read.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Initial import.
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair.a libs/$(CONFIG)/libend2end_test_writes_done_hangs_with_pending_read.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_writes_done_hangs_with_pending_read_test
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_writes_done_hangs_with_pending_read_test: $(CHTTP2_SOCKET_PAIR_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS:.o=.dep)
Initial import.
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Initial import.
10 years ago
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_AFTER_ACCEPT_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_AFTER_ACCEPT_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_AFTER_ACCEPT_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_after_accept_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_after_accept_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_AFTER_ACCEPT_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_cancel_after_accept.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_AFTER_ACCEPT_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_cancel_after_accept.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_after_accept_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_one_byte_at_a_time_cancel_after_accept_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_AFTER_ACCEPT_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_AFTER_ACCEPT_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_after_accept_and_writes_closed_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_after_accept_and_writes_closed_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_cancel_after_accept_and_writes_closed.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_cancel_after_accept_and_writes_closed.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_after_accept_and_writes_closed_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_one_byte_at_a_time_cancel_after_accept_and_writes_closed_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_AFTER_ACCEPT_AND_WRITES_CLOSED_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_AFTER_INVOKE_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_AFTER_INVOKE_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_AFTER_INVOKE_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_after_invoke_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_after_invoke_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_AFTER_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_cancel_after_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_AFTER_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_cancel_after_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_after_invoke_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_one_byte_at_a_time_cancel_after_invoke_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_AFTER_INVOKE_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_AFTER_INVOKE_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_BEFORE_INVOKE_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_BEFORE_INVOKE_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_BEFORE_INVOKE_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_before_invoke_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_before_invoke_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_BEFORE_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_cancel_before_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_BEFORE_INVOKE_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_cancel_before_invoke.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_before_invoke_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_one_byte_at_a_time_cancel_before_invoke_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_BEFORE_INVOKE_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_BEFORE_INVOKE_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_IN_A_VACUUM_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_IN_A_VACUUM_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_IN_A_VACUUM_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_in_a_vacuum_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_in_a_vacuum_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_IN_A_VACUUM_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_cancel_in_a_vacuum.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_IN_A_VACUUM_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_cancel_in_a_vacuum.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_cancel_in_a_vacuum_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_one_byte_at_a_time_cancel_in_a_vacuum_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_IN_A_VACUUM_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CANCEL_IN_A_VACUUM_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CENSUS_SIMPLE_REQUEST_TEST_SRC = \
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CENSUS_SIMPLE_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CENSUS_SIMPLE_REQUEST_TEST_SRC))))
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_census_simple_request_test: openssl_dep_error
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_census_simple_request_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CENSUS_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_census_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CENSUS_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_census_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_census_simple_request_test
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_one_byte_at_a_time_census_simple_request_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CENSUS_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_CENSUS_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
C implementation of Census trace store and stats store for grpc C lib. Change on 2015/01/08 by hongyu <hongyu@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83556470
10 years ago
endif
endif
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_DISAPPEARING_SERVER_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_DISAPPEARING_SERVER_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_DISAPPEARING_SERVER_TEST_SRC))))
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_disappearing_server_test: openssl_dep_error
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_disappearing_server_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_DISAPPEARING_SERVER_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_disappearing_server.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_DISAPPEARING_SERVER_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_disappearing_server.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_disappearing_server_test
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_one_byte_at_a_time_disappearing_server_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_DISAPPEARING_SERVER_TEST_OBJS:.o=.dep)
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_DISAPPEARING_SERVER_TEST_OBJS:.o=.dep)
GOAWAY & Reconnection support. Clients stay connected to a server after it shutdowns until all active calls have completed, and then they drop. After a GOAWAY or a disconnect, clients will attempt to re-resolve and reconnect to the server. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82178515
10 years ago
endif
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_early_server_shutdown_finishes_inflight_calls_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_early_server_shutdown_finishes_inflight_calls_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_inflight_calls.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_inflight_calls.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_early_server_shutdown_finishes_inflight_calls_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_one_byte_at_a_time_early_server_shutdown_finishes_inflight_calls_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_EARLY_SERVER_SHUTDOWN_FINISHES_INFLIGHT_CALLS_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_early_server_shutdown_finishes_tags_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_early_server_shutdown_finishes_tags_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_tags.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_early_server_shutdown_finishes_tags.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_early_server_shutdown_finishes_tags_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_one_byte_at_a_time_early_server_shutdown_finishes_tags_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_EARLY_SERVER_SHUTDOWN_FINISHES_TAGS_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_GRACEFUL_SERVER_SHUTDOWN_TEST_SRC = \
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_GRACEFUL_SERVER_SHUTDOWN_TEST_SRC))))
ifeq ($(NO_SECURE),true)
TIPS client implementation.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Add an optional server shutdown event. Allows servers to expedite shutdown once all in-progress calls are completed. We may want to eventually remove grpc_server_shutdown in preference to this.
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_graceful_server_shutdown_test: openssl_dep_error
else
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_graceful_server_shutdown_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_graceful_server_shutdown.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_graceful_server_shutdown.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_graceful_server_shutdown_test
endif
deps_chttp2_socket_pair_one_byte_at_a_time_graceful_server_shutdown_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS:.o=.dep)
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_GRACEFUL_SERVER_SHUTDOWN_TEST_OBJS:.o=.dep)
endif
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_INVOKE_LARGE_REQUEST_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_INVOKE_LARGE_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_INVOKE_LARGE_REQUEST_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_invoke_large_request_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_invoke_large_request_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_INVOKE_LARGE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_invoke_large_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_INVOKE_LARGE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_invoke_large_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_invoke_large_request_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_one_byte_at_a_time_invoke_large_request_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_INVOKE_LARGE_REQUEST_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_INVOKE_LARGE_REQUEST_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_MAX_CONCURRENT_STREAMS_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_MAX_CONCURRENT_STREAMS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_MAX_CONCURRENT_STREAMS_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_max_concurrent_streams_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_max_concurrent_streams_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_MAX_CONCURRENT_STREAMS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_max_concurrent_streams.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_MAX_CONCURRENT_STREAMS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_max_concurrent_streams.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_max_concurrent_streams_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_one_byte_at_a_time_max_concurrent_streams_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_MAX_CONCURRENT_STREAMS_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_MAX_CONCURRENT_STREAMS_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_NO_OP_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_NO_OP_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_NO_OP_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_no_op_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_no_op_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_NO_OP_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_no_op.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_NO_OP_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_no_op.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_no_op_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_one_byte_at_a_time_no_op_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_NO_OP_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_NO_OP_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_PING_PONG_STREAMING_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_PING_PONG_STREAMING_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_PING_PONG_STREAMING_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_ping_pong_streaming_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_ping_pong_streaming_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_PING_PONG_STREAMING_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_ping_pong_streaming.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_PING_PONG_STREAMING_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_ping_pong_streaming.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_ping_pong_streaming_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_one_byte_at_a_time_ping_pong_streaming_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_PING_PONG_STREAMING_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_PING_PONG_STREAMING_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_SRC))))
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_request_response_with_binary_metadata_and_payload_test: openssl_dep_error
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_request_response_with_binary_metadata_and_payload_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_request_response_with_binary_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_request_response_with_binary_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_request_response_with_binary_metadata_and_payload_test
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_one_byte_at_a_time_request_response_with_binary_metadata_and_payload_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_BINARY_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Binary header encoding. Also fixes a rather embarrassing bug in bin_encoder.c. Change on 2014/12/12 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82024795
10 years ago
endif
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_request_response_with_metadata_and_payload_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_request_response_with_metadata_and_payload_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_request_response_with_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_request_response_with_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_request_response_with_metadata_and_payload_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_one_byte_at_a_time_request_response_with_metadata_and_payload_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_request_response_with_payload_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_request_response_with_payload_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_request_response_with_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_request_response_with_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_request_response_with_payload_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_one_byte_at_a_time_request_response_with_payload_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_PAYLOAD_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_SRC))))
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_request_response_with_trailing_metadata_and_payload_test: openssl_dep_error
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_request_response_with_trailing_metadata_and_payload_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_request_response_with_trailing_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
$(E) "[LD] Linking $@"
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_request_response_with_trailing_metadata_and_payload.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_request_response_with_trailing_metadata_and_payload_test
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_one_byte_at_a_time_request_response_with_trailing_metadata_and_payload_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_REQUEST_RESPONSE_WITH_TRAILING_METADATA_AND_PAYLOAD_TEST_OBJS:.o=.dep)
Trailing metadata support. Actually exposing it from the C API. Also removing grpc_status, since it's mostly useless. Change on 2014/12/15 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82180546
10 years ago
endif
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_SIMPLE_DELAYED_REQUEST_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_SIMPLE_DELAYED_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_SIMPLE_DELAYED_REQUEST_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_simple_delayed_request_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_simple_delayed_request_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_SIMPLE_DELAYED_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_simple_delayed_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_SIMPLE_DELAYED_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_simple_delayed_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_simple_delayed_request_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_one_byte_at_a_time_simple_delayed_request_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_SIMPLE_DELAYED_REQUEST_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_SIMPLE_DELAYED_REQUEST_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_SIMPLE_REQUEST_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_SIMPLE_REQUEST_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_SIMPLE_REQUEST_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_simple_request_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_simple_request_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_SIMPLE_REQUEST_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_simple_request.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_simple_request_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_one_byte_at_a_time_simple_request_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_SIMPLE_REQUEST_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Strip "_test" from the name of thread_stress_test.c. Having a BUILD target named chttp2_fullstack_thread_stress_test_test is deceptive as it is a direct test of the system rather than a test of a test of the system. Change on 2014/12/09 by nathaniel <nathaniel@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81688649
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_THREAD_STRESS_TEST_SRC = \
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_THREAD_STRESS_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_THREAD_STRESS_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_thread_stress_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_thread_stress_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_THREAD_STRESS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_thread_stress.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_THREAD_STRESS_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_thread_stress.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_thread_stress_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_one_byte_at_a_time_thread_stress_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_THREAD_STRESS_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_THREAD_STRESS_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_SRC = \
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS = $(addprefix objs/$(CONFIG)/, $(addsuffix .o, $(basename $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_SRC))))
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifeq ($(NO_SECURE),true)
Adding a few more comments in the Makefile template, to explain some of its more obscure features.
10 years ago
# You can't build secure targets if you don't have OpenSSL with ALPN.
Allow specifying CONFIG as a string, instead of DEBUG=1. This will open the way for CONFIG=tsan, for instance. Change on 2015/01/06 by ctiller <ctiller@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83368093
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_writes_done_hangs_with_pending_read_test: openssl_dep_error
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
else
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_writes_done_hangs_with_pending_read_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_writes_done_hangs_with_pending_read.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
$(E) "[LD] Linking $@"
C/C++ separation of the install target. -) install, static, shared and strip are now fully dissociated -) make_dirs is gone Change on 2014/12/08 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81631734
10 years ago
$(Q) mkdir -p `dirname $@`
Splitting grpc_test_util into gpr_test_util. Change on 2015/01/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=83801564
10 years ago
$(Q) $(LD) $(LDFLAGS) $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS) libs/$(CONFIG)/libend2end_fixture_chttp2_socket_pair_one_byte_at_a_time.a libs/$(CONFIG)/libend2end_test_writes_done_hangs_with_pending_read.a libs/$(CONFIG)/libend2end_certs.a libs/$(CONFIG)/libgrpc_test_util.a libs/$(CONFIG)/libgrpc.a libs/$(CONFIG)/libgpr_test_util.a libs/$(CONFIG)/libgpr.a $(LDLIBS) $(LDLIBS_SECURE) -o bins/$(CONFIG)/chttp2_socket_pair_one_byte_at_a_time_writes_done_hangs_with_pending_read_test
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Fix some C++ test dependencies
10 years ago
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
deps_chttp2_socket_pair_one_byte_at_a_time_writes_done_hangs_with_pending_read_test: $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
ifneq ($(NO_SECURE),true)
ifneq ($(NO_DEPS),true)
Remove _DEPS variables They're trivially recreatable from the _OBJS variables now.
10 years ago
-include $(CHTTP2_SOCKET_PAIR_ONE_BYTE_AT_A_TIME_WRITES_DONE_HANGS_WITH_PENDING_READ_TEST_OBJS:.o=.dep)
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
endif
Makefile will now build zlib and/or OpenSSL if needed. -) Detecting system and embedded presence of zlib and OpenSSL with ALPN. -) Automatically disabling secure targets if no OpenSSL present --> make all won't work if no OpenSSL is present, forcing the users to select the nonsecure Makefile targets explicitely. -) Removing build instructions for OpenSSL - this isn't really necessary anymore. -) Adding more blurb about OpenSSL and the new Makefile features. Change on 2014/12/12 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=82020890
10 years ago
endif
Updating moe_db.txt with the latest equivalence since the ruby import changed the exported structure. Change on 2014/12/01 by nnoble <nnoble@google.com> ------------- new [] file for grpc testing. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Fix unfinished calls in thread_stress_test. Previously we had an early return if we cancelled a stream part way through a message. Correct this, so that half close and full close signals are propagated up the stack correctly so that higher level state machines can see the termination. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Remove dependency on internal C code. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Turn off the flaky bit from thread_stress_test. Change on 2014/12/02 by ctiller <ctiller@google.com> ------------- Add test cases of empty request/response, request streaming, response streaming, and half duplex streaming. Bring up the GFE/ESF for mannual test: [] build java/com/google/net/[]/testing/integration/hexa:server_components_env []-bin/java/com/google/net/[]/testing/integration/hexa/server_components_env --manual --rpc_port=25000 --use_autobahn Change on 2014/12/02 by chenw <chenw@google.com> ------------- Make echo/server.c and fling/server.c shutdown cleanly on SIGINT, and update the relevant tests to exercise this mechanism. Now "[] coverage" and the memory leak detector are able to see into the server processes. Change on 2014/12/02 by pmarks <pmarks@google.com> ------------- Allow the # of channels to be configurable in this performance test. The threads will use the channels in statically-defined round-robin order (not based on when RPCs complete on any channel). The interesting cases are #channels=1 or #channels=#threads (we previously only had the latter case) Change on 2014/12/02 by vpai <vpai@google.com> ------------- Fixed a typo and reworded a comment. Change on 2014/12/02 by gnezdo <gnezdo@google.com> ------------- Require the grpc_call in this ClientContext to be NULL before allowing set_call to be invoked. Otherwise, it's an indication of a leak somewhere. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Correctly return status other than ok and add a test for it. Change on 2014/12/02 by yangg <yangg@google.com> ------------- Better C++ guards for grpc_security.h Change on 2014/12/02 by nnoble <nnoble@google.com> ------------- Use nullptr instead of NULL for consistency. Change on 2014/12/02 by vpai <vpai@google.com> ------------- Updates the ruby gRPC service class to require the serialization method to be a static method - this brings it inline with the proto3 ruby API - it adds a monkey patch to allow existing proto (beefcake) to continue working. Change on 2014/12/02 by temiola <temiola@google.com> ------------- Adding a buildable unit to the blue print file. Added the buildable unit as its name will be usesd as tap project id. This test will fail right away in tap until tests are actually added. Change on 2014/12/02 by donnadionne <donnadionne@google.com> ------------- Move interop ESF C++ server from Java to grpc directory. Tests passed: [] test javatests/com/google/net/[]/testing/integration/hexa/... [] test net/grpc/testing/interop/esf_server/... Change on 2014/12/02 by chenw <chenw@google.com> ------------- Return a lame channel as opposed to NULL when secure channel creation fails. - Looks like we're going to need something similar server-side. - I changed the prototype of the lame client channel factory to take an explicit void as I think this is better practice in C. Let me know if you disagree and I will revert these changes. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- Putting ALPN support where it belongs. Change on 2014/12/02 by jboeuf <jboeuf@google.com> ------------- GOAWAY send path. Sends a GOAWAY frame when shutting down. This is not read and understood yet. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Adds support for secure channels and servers. - wraps new C apis (credentials, server_credentials) and Server#add_secure_http_port - adds tests to ensure credentials and server credentials can be created - updates client_server_spec to run the client_server wrapper layer end-to-end tests using a secure channel Change on 2014/12/03 by temiola <temiola@google.com> ------------- Fix existing issues regarding out of order events. At the client side, using pluck as the client_metadata_read can happen anytime after invoke. At the server side, allow halfclose_ok and rpc_end to come in reverse order. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Don't track coverage of tests. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Change UnaryCall to conform standard test requirement of LargeUnaryCall. Change on 2014/12/03 by yangg <yangg@google.com> ------------- updating alpn version to h2-15 ensure all interop are on the same version and working. Java and go are not ready for h2-16 yet. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Add config to bring echo server in []. This is used to test production GFE as its bckend. Change on 2014/12/03 by chenw <chenw@google.com> ------------- In preparation for fixing shutdown race issues, change em to take ownership of the file descriptor. Add an API to grpc_tcp to take an already created grpc_em_fd object, and change tcp_client to use that API. This is needed because otherwise an em user's close() of the file descriptor may race with libevent internals. That's not an issue yet because destroy() frees the events inline, but that can't be done safely if there is a concurrent poller. Change on 2014/12/03 by klempner <klempner@google.com> ------------- Fixing TAP opensource build We don't want to compile and run C++ tests in the C target. Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Move and separate interop tests by languages. Small fixes to the test runner. Improving logging. Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Fixing the opensource build: -) The C/C++ split wasn't done up to the 'dep' target level -) The alpn.c file was missing from build.json Change on 2014/12/03 by nnoble <nnoble@google.com> ------------- Adding blue print files after projects exist Change on 2014/12/03 by donnadionne <donnadionne@google.com> ------------- Refactor StreamContext using the new completion_queue_pluck API. The dedicated the poller thread has been removed. This CL keeps the current behavior to make it short. There is one following to make it usable for both client and server. The tags for pluck is based on the address of this StreamContext object for potential debug use. The Read/Write and Wait cannot be called concurrently now and this might need to be fixed. Change on 2014/12/03 by yangg <yangg@google.com> ------------- Binary encoding utilities. Support base64 encoding, HPACK static huffman encoding, and doing both at once. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Enforce Makefile regeneration in presubmits. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Make CloseSend() send a real zero-length control message to indicate EOS. Change on 2014/12/03 by zhaoq <zhaoq@google.com> ------------- Prefer to create dualstack sockets for TCP clients and servers, with automatic fallback for environments where IPV6_V6ONLY can't be turned off. Change on 2014/12/03 by pmarks <pmarks@google.com> ------------- Add opensource path to build targets. Ensure that MOE is going to run. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- Add PingPong test case. Delete FullDuplex test case. The latter is not specified for client in https://docs.google.com/document/d/1dwrPpIu5EqiKVsquZfoOqTj7vP8fa1i49gornJo50Qw/edit# Change on 2014/12/03 by chenw <chenw@google.com> ------------- Make generate_projects.sh check out the generated targets. Change on 2014/12/03 by ctiller <ctiller@google.com> ------------- rspec cleanup - stops declaring specs within the GRPC module - splits Bidi streaming specs into a separate test suite adding tests in the GRPC module was a mistake, it pollutes the module and can affect other tests that run later by the test runner the bidi tests are currently flaky, having them run in their own test suite allows having two separate continuous builds (once ruby gRPC is on GitHub), one that includes bidi where we tolerate flakiness, and another that does not, where there should be no flakiness at all Change on 2014/12/03 by temiola <temiola@google.com> ------------- Adding support for composite and IAM credentials. - For now, we don't do any checks on credentials compatibility in the composite credentials. Maybe we'll add that later. - Refactored the end to end security tests so that we always use the public API (except for the fake security context which is not exposed). Change on 2014/12/03 by jboeuf <jboeuf@google.com> ------------- Make GPR library buildable in Visual Studio 2013. Change on 2014/12/04 by jtattermusch <jtattermusch@google.com> ------------- Adds codegen for ruby This is being added now that ruby's proto and grpc apis are defined and stable Change on 2014/12/04 by temiola <temiola@google.com> ------------- Prevent NewStream() from sending negative or 0 timeout. Change on 2014/12/04 by zhaoq <zhaoq@google.com> ------------- Add a grpc_sockaddr_to_string() function, and use it when logging bind failures. Also improve const-correctness in some earlier code. I'm not certain whether inet_ntop() will need any platform-specific implementations, but for now the compiler offers no complaints. Demo: $ []-bin/net/grpc/c/echo_server 1.2.3.4:80 ... tcp_server.c:139] bind addr=[::ffff:1.2.3.4]:80: Permission denied Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Refactoring - moves c wrapped classes to a submodule Google::RPC::Core - this allows for an explicit rule when reading through gRPC ruby code for telling when an object is pure ruby or wrapped C Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixes the bidi_call [] Change on 2014/12/04 by temiola <temiola@google.com> ------------- Fixing dev build when activating surface traces. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Updates the tests to reflect that fact that some Credentials compose works. Change on 2014/12/04 by temiola <temiola@google.com> ------------- Making the generate_project_test actually do something. Change on 2014/12/04 by nnoble <nnoble@google.com> ------------- Rename "esf_server" to "[]4_server". Delete "test_sever" from Java directory. Change on 2014/12/04 by chenw <chenw@google.com> ------------- Added PHP client interop tests. Tested large_unary against the C++ server. Change on 2014/12/04 by mlumish <mlumish@google.com> ------------- Refactor grpc_create_dualstack_socket() by pulling the setsockopt into its own function. This separates the magic test flag from the real fallback logic. Change on 2014/12/04 by pmarks <pmarks@google.com> ------------- Fixes the type of the constant used for test cert hostname Change on 2014/12/04 by temiola <temiola@google.com> ------------- Disabling these tests as they're causing flakiness. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Change intptr --> uintptr. Handles the case where a void* turns into a negative number, which then gets hashed into a negative bucket and segfaults. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Add a test fixture to force parsers to handle one byte at a time. This should expand coverage and hopefully prevent errors at some point (it seems to pass out of the box for now though). Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- The code generator isn't +x. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Updates math_client and math_server to allow construction using crednetials By: - Extending rpc_server constructor so that it takes a credentials keyword param - Extending client_stub constructor so that it takes a credentials keyword param Change on 2014/12/04 by temiola <temiola@google.com> ------------- Format output a little more nicely. Print each line of output separately - previously logging.info was truncating this at some maximum length, and logs were getting lost. Change on 2014/12/04 by ctiller <ctiller@google.com> ------------- Up timeout for this test. Under TSAN, if we process one byte at a time, this timeout can be reached - and I think this is the cause of the following flake: [] Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Adding more error logging for ssl. Change on 2014/12/05 by jboeuf <jboeuf@google.com> ------------- Read path for goaway. Still need to add hooks to deprecate a channel on the client side when goaway is received. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Separate accept() into server_accept() and server_end_of_initial_metadata(). This allows servers to initiate reads before finishing writing metadata. Change on 2014/12/05 by ctiller <ctiller@google.com> ------------- Fix for breakage 11512317 - adding missing test files. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- grpc c++ server side streaming support. This is based on [] There is a lot of room to clean up the internal implementation which may require refactoring of CompletionQueue. The current cl serves as a working implementation with the missing interfaces. The sample generated files are included and will be removed before submitting. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Changed to the latest timeout format again (search "grpc-timeout" in [] for the spec). Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Fixing opensource build. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Making absolutely sure we can do the moe export by adding a sh_test for it. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Change :scheme psuedo-header from "grpc" to "http" or "https". Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Add server credential wrapping for c++ server. It only wraps ssl and []2 for now. The ServerCredentials class and the factory class are in a similar fashion as client side wrapping. The difference is the factory method returns shared_ptr instead of unique_ptr as the server builder needs to keep a reference to it for actually creating the server later. The integration will happen in a following cl. Change on 2014/12/05 by yangg <yangg@google.com> ------------- Fixed bugs in new_grpc_docker_builder.sh Change on 2014/12/05 by mlumish <mlumish@google.com> ------------- In secure endpoint, hold a refcount for the life of a write callback if the write does not complete immediately. Change on 2014/12/05 by klempner <klempner@google.com> ------------- Add migration support to MOE and have TAP verify it doesn't break. Migration support allows mirroring commits from [] into the git repo, instead of just a dump of the current source. Change on 2014/12/05 by ejona <ejona@google.com> ------------- Change initial window size to 65535 according http2 draft 15. Change on 2014/12/05 by zhaoq <zhaoq@google.com> ------------- Re-enable the flaky cases in dualstack_socket_test, with additional logging to help track down the problem if it surfaces again. This also seems like a good opportunity to make grpc_socket_utils a separate library, as it's not really specific to TCP. Example output: logspam: [], 26570) resolved 2 addrs in 37ms: logspam: [0] [::1]:26570 logspam: [1] 127.0.0.1:26570 Change on 2014/12/05 by pmarks <pmarks@google.com> ------------- Opensource build fixes. -) A function that has a return type should actually return something. -) Don't pass unsigned chars to strlen and strncmp. Change on 2014/12/05 by nnoble <nnoble@google.com> ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=81458281
10 years ago
Initial import.
10 years ago
Add missing space
10 years ago
.PHONY: all strip tools dep_error openssl_dep_error openssl_dep_message git_update stop buildtests buildtests_c buildtests_cxx test test_c test_cxx install install_c install_cxx install-headers install-headers_c install-headers_cxx install-shared install-shared_c install-shared_cxx install-static install-static_c install-static_cxx strip strip-shared strip-static strip_c strip-shared_c strip-static_c strip_cxx strip-shared_cxx strip-static_cxx dep_c dep_cxx bins_dep_c bins_dep_cxx clean
Initial import.
10 years ago