* Use a unique_ptr with custom deleter for grpc_transport in
SubchannelConnector::Result
* comment
* fix
* iwyu
* std::ignore
* review
* explicit release ownership
* Bump v1.51.x to 1.51.0-pre1 (#31622)
* bump version to 1.51.0-pre1
* regenerate projects
* Branch 1.51.x: Cherrypick "Adding token url validation cases for psc endpoints #31616" and "Added url validation for aws metadata endpoints in aws external account #31626" (#31646)
* Adding token url validation cases for psc endpoints (#31616)
* Adding validation case for psc endpoint
* formatting fix
* Added url validation for aws metadata endpoints in aws external account (#31626)
* Added url validation for aws metadata endpoints in aws external account
* addressing review comments
* fix error message back
* Fix broken test
* xds resolver: fix edge cases in interactions between LDS and RDS (#31668) (#31672)
* xds resolver: fix edge cases in interactions between LDS and RDS
* improve SwitchFromInlineRouteConfigToRds test
* clang-tidy
* Automated change: Fix sanity tests
Co-authored-by: markdroth <markdroth@users.noreply.github.com>
Co-authored-by: markdroth <markdroth@users.noreply.github.com>
* xDS: fix crash on wrong listener type (both client and server side) (#31684) (#31691)
* xds resolver: fix crash on wrong listener type
* fix same bug on server side
* fix clang-tidy and add requested TODO
* Bump v1.51.x to v1.51.0 (#31705)
* bump version to 1.51.0
* regenerate projects
* Revert "Build with System OpenSSL on Mac OS arm64 (#31096)" (#31739)
This reverts commit b3d9833166.
* Bump v1.51.x to 1.51.1 (#31740)
* Bump to 1.51.1
* Regenerate projects
Co-authored-by: Richard Belleville <rbellevi@google.com>
Co-authored-by: aeitzman <12433791+aeitzman@users.noreply.github.com>
Co-authored-by: Mark D. Roth <roth@google.com>
Co-authored-by: markdroth <markdroth@users.noreply.github.com>
* Inital Grpc.Tools unit tests
* refining the tests
* Get tests to run on Linux
* further grpc.tools tests changes
* changes for relative/absoulte paths
* Simplify running on just one framework
* Add Protobuf.MSBuild.dll for tests to use
* fix typo in comment
* fix copyright note
* Revert "Add Protobuf.MSBuild.dll for tests to use"
This reverts commit 542756cade.
* cleanup Directory.Build.props
* fix location of tasksAssembly
* cleanup in preparing paths
* add test-out to .gitignore
* many simplifications to MsBuildIntegrationTest
* update expected test results data
* many improvements to fakeprotoc.py
* yapf code
* fix copyright licenses
* reintroduce normalized paths
* minor fixes
* attempt simplify Directory.Build.targets
* newline at EOF
* simplify result comparison
* add a TODO
* further simplify MsBuildIntegrationTest test harness
* avoid internal symbol conflict by diambiguating test protos
* add TODO
* style fixes
* one more fixup
Co-authored-by: Jan Tattermusch <jtattermusch@google.com>
Updates the ProtoReflectionDescriptorDatabase ctor to take a reference
to std::shared_ptr<grpc::ChannelInterface> rather than
std::shared_ptr<grpc::Channel>. This helps code that is making use of
grpc::ChannelInterface from having to perform a cast from the Base to
the Derived when creating the refelction db.
* initial
* Intermediate
* Another try
* Try multiple necesary pulls
* Filter works other than client half close
* Fixes
* Add a cancelled RPC test
* Handle trailer only responses
* Tests for disabled logging and truncated payloads
* Fix authority and peer
* Add TODOs and asserts for half-close
* Fix tests for half-close and cancel
* 2d748fcb1cf45cac62729b8346ad15e6abc79e97
* Fix sanity checks
* Strict bazel build
* Fix package
* IWYU
* Fix cmake
* Explicit cast to string
* Size casts
* Fix Arena leak and disable macos build for now
* Reviewer comments
In Ruby 3.2 the default allocator warns if used with `Data_Wrap_Struct`. This removed the default allocator in those cases.
Ruby discussion: https://bugs.ruby-lang.org/issues/18007
* [༺ EventEngine༻ ] Specify requirements for Run* immediate execution
Also adds test suite tests that may catch non-conforming implementations
(Run called ~3200 times in case it's non-deterministic).
* fix
* verbiage; longer timeout
* fix int type comparison warning
* rm illegal term
* Automated change: Fix sanity tests
* disable tests
Co-authored-by: drfloob <drfloob@users.noreply.github.com>
* Revert "Revert "Add support for systemd socket activation (#30485)" (#31617)"
This reverts commit 867dc6cae2.
* Add checks to unix tests
* Ran generate_projects.sh and fixed styling in test
* Fix variable in unit test
* Use reinterpret_cast in test
* Rebase and fix sanity failures
Previously attempting to build the Ruby C extensionwould result in
undefined symbol:
```
Undefined symbols for architecture arm64:
"_ruby_abi_version", referenced from:
-exported_symbol[s_list] command line option
ld: symbol(s) not found for architecture arm64
clang: error: linker command failed with exit code 1 (use -v to see invocation)
```
As noted in https://github.com/ruby/ruby/pull/6231,
`_ruby_abi_version` is only available for development
versions. `RUBY_PATCHLEVEL` < 0 denotes a development version, while
>= 0 denotes a released version.
This is motivated by #31227 which I also encountered recently. I'm not familiar with the call lifecycle details, but it seems clear that there is some kind of race condition between user-initiated cancellations and other completion status updates.
If a user cancels a call before being notified of its completion, I believe it should be considered cancelled unconditionally, regardless of whether the call actually completed successfully before its cancellation status could be set.
So it seems better to re-raise explicitly rather than using self._raise_for_status(), which is a no-op in this race condition and can lead to an UnboundedLocalError.