* make error independent
* Separate grpc_error into its own library
This is forward work to move Closure, ExecCtx into their own libraries
in order to make use of them in the activity code for resource quota
wakeups.
* Automated change: Fix sanity tests
* fixes
Co-authored-by: ctiller <ctiller@users.noreply.github.com>
* 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>
* Add Python 3.10.0rc1 binary wheels
* Drop Python 3.5 artifacts
* Document the drop of 3.5
* Fix the wrong pip pointer
* Update manylinux2014 to a newer version, remove 3.5 distribtest
* Update manylinux aarch64 to see if the absl error go away
* Use the preferred alias
* Allow different wheel library to produce different tag order
* Remove unused shell var and log produced wheels
* Use copy instead of move
* Make bash happy about the wildcard
* Upgrade the debian image to use 3.5+ Python
* Polish the comments for the Dockerfiles
* Implement type safety for TLS
This is mostly free when compiler support is available, but requires
careful templating when implemented using pthread.
Significantly slimmed the tls.h interface; it now only defines the "TLS
keyword" for each supported compiler, delegating enforcement of correct
usage (i.e. must be static) to the compiler itself.
Implemented implicit conversion for the pthread wrapper so it can be
used (mostly) the same as native support. Notable exception to this is
that static_cast<void*> is needed when printing a pointer stored in TLS
as %p.
* Use GPR_THREAD_LOCAL macros consistently
* Bump Envoy and related submodules to latest
* Update googleapis
* Update scripts for newer submodules
* Update udpa
* Add opentelemetry
* Python changes for opentelemetry
* Update udpa in check_submodules
* Add opentelemetry to check submodules
* Regenerate upb files
* Add new proto dependency to upb-gen and Bazel
* Regenerate project
* Add isort_code.sh to sanity tests
* Run tools/distrib/isort_code.sh
* Fine tune the import order for relative imports
* Make pylint and project generation happy
* Fix a few corner cases
* Use --check instead of --diff
* The import order impacts test result somehow
* Make isort print diff and check output at the same time
* Let tools/run_tests/python_utils be firstparty library
* Run isort against latest HEAD
* fix: If client request iterators raise exceptions, _consume_request_iterator should not dump its stack to stderr
fixes: #25339
* fix: use format_exc
* fix: linter
* fix: lint round 2
This code adds an iomgr implementation that's backed by an EventEngine. This uses the EventEngine API alone, and separate work will introduce an EventEngine prototype to plug into it.
See also drfloob#1: @nicolasnoble has a pull request against this branch, implementing the libuv-based EventEngine. One goal here is to implement the iomgr code such that it can be merged independently without affecting normal builds.
This implementation can be built using bazel build --cxxopt='-DGRPC_USE_EVENT_ENGINE' :all
Some shortcuts are being taken to get a working, testable version of the engine. EventEngines are not pluggable, for example.
* support user provided audience in gdc and jwt
* fix 1st round of comments
* fix ruby and php to use new GDC API
* fix python clang issue
* address 2nd round of comments
* fix string_view issue
* remove length param in string_view constructor
* Handle gevent exception in gevent poller
Currently the gevent poller ignores exceptions raised by
`gevent.wait()`, which causes greenlets to be unkilable while waiting.
This change handles exceptions raised while waiting in the gevent
poller, cancels the gRPC call and propagates the error back to the
application.
Co-authored-by: Kostis Lolos <klolos@arrikto.com>
* Fix imports in header files
* Lint gevent tests
* Set grpc event type to GRPC_QUEUE_SHUTDOWN upon cancel error
To prevent `grpc_completion_queue_next()` to be called indefinitely when
the queue is shut down.
* Remove unnecessary `except *`
* Improve gevent tests
* Format code
* Remove unnecessary import
Co-authored-by: Kostis Lolos <klolos@arrikto.com>
* Add grpcio-csds pacakge
* Remove unused file
* Fix the proto import path issue
* Update the CSDS package and xds-protos for PY2
* Make tests happy
* Fix Bazel proto dependency
* Add Python2 tests for CSDS
* Expose code and details from context on the server side
This makes them accessible to server-side interceptors.
* Complete synchronous implementation
* Add experimental API notes
* Complete async implementation
* Revert bogus copyright change
* Fix metadata tests for py2.7
* Add new test to tests.json
* Check error codes in tests
* Bump autogenerated Python version
* pass XdsClient to LB policies via channel args
* add channel arg for overriding bootstrap config on a per-channel basis
* change tests to use new channel arg approach -- currently failing for server-side tests
* use the same channel args approach on the server side
* clang-format
* fix CircuitBreakingMultipleChannelsShareCallCounter test
* fix XdsEnabledServerTest test
* add TODO
* clang-format
* generate_projects
* fix clang-tidy
* fix build
* attempt to fix python
* rename ChannelData to ClientChannel
* make ClientChannel class definition public
* move retry code to its own filter
* move LB call factory method to ClientChannel class
* move dynamic termination filter out of ClientChannel class
* update comments
* remove retry parsing from client channel service config parser
* fix clang-tidy
* fix service_config_test
* clang-format
This caused problems internally. The interface deletion needs a cherrypick, but I will do an overnight test first, before I reintroduce. CC @nicolasnoble
* Implement the xDS Config Dump as CSDS in Core
* Revemp the logic of caching and constructing the CSDS response
* Unref created error
* Fix the clang tidy complains
* Resolve comments about symbol preload and Json ctor
* Improve readability
* Merge with new ADS logic && support latest CSDS
* Refactor the version/client_status logic
* Add support for error_state
* Add support for does_not_exist
* Add node information
* Fix the ProtoBuf segfault
* More test cases
* Refactor to use bytes as cache instead of JSON
* Apply clang_tidy's suggestion
* Resolve reviewer's requests
* Tiny stylish fixes && make sanity test happy
* WIP: still working on the matchers
* Update a bunch of matchers
* Improve readability a bit
* Resolve reviewer's suggestions
* Resolve reviewer's comments
* Make Unpack fail fast
* Revert "Revert "xDS status notifier (#25321)" (#25702)"
This reverts commit 3c9f3972e3.
* Remove connection from map when OnClose is not registered
* Reviewer comments
* Refactored with dynamic filters
* Error-tolerance tuned
* Fix leak of grpc_error and data race of canceller
* Adopt the latest xDS HTTP filter framework
* Fix fault injection tests' conflict with router filter
* Test alternative setup (override, no-override) without copy
* Refactor file strcutures of fault injection filter
* Rewrite the Json parsing/assembling logic again
* Added logic for aborting streaming RPC && resolve comments
* Revert "Revert "xDS HTTP filter support for gRPC client" (#25555)"
This reverts commit 785a128a83.
* xds_end2end_test: avoid flakes from lingering shutdown from previous test
* Download the working msys64 and use it to compile 27
* Remove the cygwin detection override logic
* Only install Python interpreters in jobs running on winserver2016
* Fix the batch script substring check
* Increase the timeout for related jobs
* TIL that there is an allow list for Kokoro env vars
Craig Tiller