* separate slice into internal and refcount, with refcount being the minimal base
* python3-ize
* Separate slices from mdelems for static data
Allows us to separate a small build target with just some slice basics
from the rest, which will help simplify the build system longer term.
Additionally exposed an opportunity to eliminate an init function.
Should also help in a month or so when it's time to jettison interned
metadata and slices entirely.
* fix compilation
* Automated change: Fix sanity tests
* fixes
* fixfixfix
Co-authored-by: ctiller <ctiller@users.noreply.github.com>
Some changes:
* OnTransactCb now takes a non-const ReadableParcel* so that testing
codes no longer have to rely on mutable.
* Remove GetReadableParcel() interface from binder since we only sent
one-way transaction and the output (readable) parcel is never used.
* Remove GetDataPosition() / SetDataPosition() interfaces since they are
both unused.
* Some changes that should've been made to #27257 but was somehow
missing...
Android-related binder classes are only available if
GPR_SUPPORT_BINDER_TRANSPORT is defined. Thus, BinderServerCredentials
should only work if GPR_SUPPORT_BINDER_TRANSPORT (instead of
GPR_ANDROID) is defined as well.
* Fix ReadableParcelAndroid::ReadString interface
Also uses implementation from android/binder_parcel_utils.h to read
ByteArray and string from Parcel
Test example app on device, works correctly
There was a bug found by the fuzzer where we might access wire_writer_ before
finishing SETUP_TRANSPORT (and thus constructing wire_writer_). This PR
fixes such issue by making sure that we won't proceed with any requests
until the connection is fully established.
Since binder transactions may be coming from multiple different threads,
this PRs guard some of the WireReaderImpl's member with a mutex to make
sure there's no races between threads.
Add `grpc::BinderServerCredentials()` and other related functionalities for the server to listen to binder transactions through a phony "binder port".
The APIs are temporarily placed in internal headers until the corresponding gRFC is merged.
This PR imports unit tests and end-to-end tests of binder transport from the internal repository. No further changes will be made to internal repository.
* Buffer HPACK parsing until the end of a header boundary
HTTP2 headers are sent in (potentially) many frames, but all must be
sent sequentially with no traffic intervening.
This was not clear when I wrote the HPACK parser, and still indeed quite
contentious on the HTTP2 mailing lists.
Now that matter is well settled (years ago!) take advantage of the fact
by delaying parsing until all bytes are available.
A future change will leverage this to avoid having to store and verify
partial parse state, completely eliminating indirect calls within the
parser.
* maybe fixes
* xx
* fix boundary detection
* clang-format
* Revert "xx"
This reverts commit 258d712ed3.
* fix tests
* add missed check
* fixes
* fix
* update tests
* fix benchmark
* properly unref
* optimize final slice refcounting
* cleanup bm_chttp2_hpack
* start
* new parser progress
* refinement
* get it compiling
* bug-fix
* build files
* clang-tidy
* fixes
* fixes
* fixes
* fix-leaks
* clang-tidy
* comments
* fix merge error
* Revert "Buffer HPACK parsing until the end of a header boundary (#26700)"
This reverts commit 8bab3e4bf4.
* streaming hpack parser start
* streaming parser
* clang-format
* Rework HPackTable into C++
* clang-tidy
* fix merge
* actually set the size of the entries array
* better
* Limit initial window size increases/decreases by flow control windows of active streams
* Limit stream flow control window updates to maximum allowable
* Reviewer comments
* Alternative way
* Clean-up
* Add tests
* Remove unnecessary cq_verifier
* Generate projects
* Initialize recv_message
* Get around compilation issue
* Test size large
HTTP2 headers are sent in (potentially) many frames, but all must be
sent sequentially with no traffic intervening.
This was not clear when I wrote the HPACK parser, and still indeed quite
contentious on the HTTP2 mailing lists.
Now that matter is well settled (years ago!) take advantage of the fact
by delaying parsing until all bytes are available.
A future change will leverage this to avoid having to store and verify
partial parse state, completely eliminating indirect calls within the
parser.
Ming-Chuan
Esun Kim
Craig Tiller
Ta-Wei Tu
Ta-Wei Tu
AJ Heller
Yash Tibrewal