Without this fix it can be the case in the filter based stack that a
unary request that sees the response payload from the wire, prior to
being cancelled, can see the response payload even though we report
cancelled - since we unilaterally remove the error status from the batch
op just prior to evaluating whether to scrub the payload or not.
(also tweaks a couple of log lines that were useful in diagnosing this
bug)
<!--
If you know who should review your pull request, please assign it to
that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the
appropriate
lang label.
-->
This PR adds lookup cancellation to the client channel resolver fuzzer,
and adds an optional call to `resolver->Orphan()` at a few specific
points.
I also significantly sped up the fuzzer by removing all mutexes, waits,
and sleeps. It's single-threaded, after all.
---------
Co-authored-by: drfloob <drfloob@users.noreply.github.com>
Use an index instead of a string to select tests (and use that index
module total test count to ensure whatever the fuzzer selects we always
run a test).
This will make the fuzzer corpus unstable when the test count changes,
which I think is fine - it'll regenerate.
<!--
If you know who should review your pull request, please assign it to
that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the
appropriate
lang label.
-->
<!--
If you know who should review your pull request, please assign it to
that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the
appropriate
lang label.
-->
ChannelArgs fuzz configuration is expected to be used in other fuzzing
targets as well. This PR extracts the common code from the API fuzzer
and converts to use the C++ types.
Allow for multiple `--grpc_experiments`, `--grpc_trace` command line
arguments to be added, accumulate them, and provide them to gRPC as one
thing.
<!--
If you know who should review your pull request, please assign it to
that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the
appropriate
lang label.
-->
---------
Co-authored-by: ctiller <ctiller@users.noreply.github.com>
Reverts grpc/grpc#33088
The internal test is failing because after the change, transitive
dependencies will be included in response too, thus we need add
`empty2_pb2` to response since it's a dependency of
`empty2_extensions_pb2`.
We already did that in OSS test, we need do the same for internal test
case.
<!--
If you know who should review your pull request, please assign it to
that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the
appropriate
lang label.
-->
<!--
If you know who should review your pull request, please assign it to
that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the
appropriate
lang label.
-->
This is nearly identical to the helloworld example's greeter callback
server, but without the health check service enabled. Having a separate
example for reflection may help folks find this code.
Early out evaluating this function where we can, and use macros to
eliminate function calls in debug builds.
Takes per-example time from 5400ms to 1200ms in debug asan builds.
<!--
If you know who should review your pull request, please assign it to
that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the
appropriate
lang label.
-->
---------
Co-authored-by: ctiller <ctiller@users.noreply.github.com>
<!--
If you know who should review your pull request, please assign it to
that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the
appropriate
lang label.
-->
---------
Co-authored-by: ctiller <ctiller@users.noreply.github.com>
When we deprecate services in proto-file:
```proto
syntax = "proto3";
option php_namespace = "Test";
service MyService {
option deprecated = true;
rpc Call(Test) returns (Test) {
option deprecated = true;
};
}
message Test {
}
```
It might be a good idea to propagate this flat in the generated classes
so that the final users will be aware of the changes coming soon.
Example of the generated code:
```php
<?php
// GENERATED CODE -- DO NOT EDIT!
namespace Test;
/**
* @deprecated
*/
class MyServiceClient extends \Grpc\BaseStub {
/**
* @param string $hostname hostname
* @param array $opts channel options
* @param \Grpc\Channel $channel (optional) re-use channel object
*/
public function __construct($hostname, $opts, $channel = null) {
parent::__construct($hostname, $opts, $channel);
}
/**
* @deprecated
* @param \Test\Test $argument input argument
* @param array $metadata metadata
* @param array $options call options
* @return \Grpc\UnaryCall
*/
public function Call(\Test\Test $argument,
$metadata = [], $options = []) {
return $this->_simpleRequest('/MyService/Call',
$argument,
['\Test\Test', 'decode'],
$metadata, $options);
}
}
```
These tags will cause IDE to highlight the class/method to bring the
developer's attention.
This is consistent with how "protoc" handles the "deprecated" option for
PHP
<!--
If you know who should review your pull request, please assign it to
that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the
appropriate
lang label.
-->
This reverts commit 1624542ea4, relanding
https://github.com/grpc/grpc/pull/32956
Because of some proto dependency and build problems internally, I've
removed the ServiceConfig proto fuzzing component. These build issues
can hopefully be resolved soon, and then we can re-add the deleted
implementation from commit
[b078c9c](b078c9c015)
in this PR.
each oob test runs at least 2s (2 verifications with each server sending
report each 1s), current configuration allows at most 3 clients run in
parallel. In reality, probably only the two clients against one server.
Currently 8 jobs are running each batch, and it happens that each batch
has at most 2 oob clients against one server.
Since 3 languages support orca, so we should allow 3 clients running
against one server so the timeout is at least 4s. We give some buffers
to allow tests running more reliably.
10s in theory support 5-6 clients running against one server.
`tools/run_tests/sanity/check_absl_mutex.sh` was broken, a missing paren
crashed the script if run locally. It's unclear yet how our sanity
checks were not complaining about this, `run_tests.py` does not save the
log.
Rare bug: server initial metadata gets stranded in the outbound pipe.
(fix is a little unpleasant, but we'll do better at the five pipes
stage)
<!--
If you know who should review your pull request, please assign it to
that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the
appropriate
lang label.
-->
re2 previously failed to compile if:
1. An old `re2` version is installed with a non-standard system prefix,
such as `/opt/local`.
2. The environment variable is set: `CPPFLAGS=-I/opt/local/include`.
Running `make` would result in function prototype mismatches because the
Makefile would previously attempt to use the headers from
`/opt/local/include/re2` before the `third_party/re2/re2` directory.
https://github.com/grpc/grpc/pull/27660 caused `CPPFLAGS` to inherit
from the environment, but this can cause the Makefile to use external
include files for re2 and other libraries if `-I` flags are defined.
This commit reverts to the original behavior of only using
`RbConfig::CONFIG` values to avoid using the wrong headers.
I've noticed we add the cleanup hook after setting up the
infrastructure. Thus, if infra setup failed, the cleanup won't work.
This fixes it, and adds extra checks to not call
`cls.test_client_runner` if it's not set.
The logger uses `absl::FPrintF` to write to stdout. After reading a
number of sources online, I got the impression that `std::fwrite` which
is used by `absl::FPrintF` is atomic so there is no locking required
here.
---------
Co-authored-by: rockspore <rockspore@users.noreply.github.com>
Fail test if client or server pods restarted during test.
#### Testing
Tested locally, test will fail with message similar to:
```
----------------------------------------------------------------------
Traceback (most recent call last):
File "/usr/local/google/home/xuanwn/workspace/xds/grpc/tools/run_tests/xds_k8s_test_driver/framework/xds_k8s_testcase.py", line 501, in tearDown
))
AssertionError: 5 != 0 : Server pods unexpectedly restarted {sever_restarts} times during test.
----------------------------------------------------------------------
Ran 1 test in 886.867s
```
This is a mistake made in https://github.com/grpc/grpc/pull/33030.
`sizeof()` would count the null byte terminated the C string and would
cause us to skip a byte if it is used as the index to
`result->substr()`. This would also crash if `result` only contains
`grpc_config=` as @drfloob pointed out.
<!--
If you know who should review your pull request, please assign it to
that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the
appropriate
lang label.
-->
This metadata doesn't actually encode so passing it through from an app
will force a crash.
<!--
If you know who should review your pull request, please assign it to
that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the
appropriate
lang label.
-->
Craig Tiller
Eugene Ostroukhov
AJ Heller
Yash Tibrewal
Mark D. Roth
Vignesh Babu
Esun Kim
Luwei Ge
Xuan Wang
mojh7
Yijie Ma
Mikhail Galanin
Sergii Tkachenko
yifeizhuang
Stan Hu