Bazel hack #1
Open
Bazel hack #1
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…envoyproxy#432) do_ci.sh runs tests with heapcheck enabled. This can discover heap leaks but requires pprof (in google-perftools) to perform the symbolization and printing of allocation site stack traces. With this PR, a failed run will produce output such as: Have memory regions w/o callers: might report false leaks Leak check _main_ detected leaks of 16 bytes in 1 objects The 1 largest leaks: Using local file /source/build/test/envoy-test. Leak of 16 bytes in 1 objects allocated from: @ 15d6b6f Network::DnsImplTest_Cancel_Test::TestBody @ 1e91e66 testing::internal::HandleSehExceptionsInMethodIfSupported @ 1e8cba8 testing::internal::HandleExceptionsInMethodIfSupported @ 1e73af5 testing::Test::Run @ 1e74381 testing::TestInfo::Run @ 1e74a12 testing::TestCase::Run @ 1e7b1bd testing::internal::UnitTestImpl::RunAllTests @ 1e92f66 testing::internal::HandleSehExceptionsInMethodIfSupported @ 1e8d86a testing::internal::HandleExceptionsInMethodIfSupported @ 1e79e47 testing::UnitTest::Run @ 181925a RUN_ALL_TESTS @ 181844d main
The MOCK_METHOD* macros from gtest are not including the override attribute and it's triggering warnings. This change includes `-Wno-inconsistent-missing-override` to that target.
Some of these types were assuming that `size_t == uint64_t` or that `uint64_t == unsigned long` but that's not universally true, some BSDs such as Darwin define uint64_t as unsigned long long which warns on literals that are defined as `UL`. Additionally there were also some assumptions made on the pch as to forward declarations. So I explicitly added some needed ones.
This change allows copy elision on unreferenced temporary objects
This change adds three main things: 1. Pushing a lyft/envoy:latest image to dockerhub every time master passes tests. 2. Check that the docker examples work on every PR. 3. Add SHA versioned lyft/envoy-build images.
* Some extra developer conveniences for Docker image.
* Refactor do_ci.sh to split the cmake and environment setup from the
build and test. This makes it convenient for devs to invoke arbitrary
make targets in the Docker image, e.g.
docker run -t -i -v $PWD:/source lyft/envoy-build:latest /bin/bash -c \
"cd /source && . ./ci/build_setup.sh && make fix_format"
* Add EXTRA_TEST_ARGS env var to pass addition flags to envoy-test. This
is useful, for example, when applying a --gtest_filter to restrict the
test run to only failing tests.
* Add a do_ci.sh fix_format target.
* Add gdb, strace to the image and a RUN_TEST_UNDER environment variable
that can be used to have envoy-test run under these programs.
* UNIT_TEST_ONLY env var that will skip the non-envoy-test aspects of
the test process.
Example command for debugging only tests matching *Dns*:
docker run -t -i -v $PWD:/source lyft/envoy-build:latest /bin/bash -c \
"cd /source && UNIT_TEST_ONLY=1 RUN_TEST_UNDER='gdb --args' ./ci/do_ci.sh debug \
--gtest_filter='*Dns*'
This is a large refactor that moves away from representing addresses internally as strings of "URL" form into proper objects. With this change in place, it should be very straightforward to add IPv6 support, as most places in the code no longer need to understand anything about the underlying address. Additionally, adding NT support in the future should also be easier. There are still some unfortunate legacy aspects, such as the fact that when loading from configuration we still parse tcp:// as well as unix:// style addresses. In the future this should likely become ip:// as well as pipe:// but we will probably need to continue to support both. Note that although this change is huge in number of lines, most of it is very straightforward, and certain portions of the code have become substantially simpler such as listener handling and client connection handling. Fixes envoyproxy#390
Previously we would incorrectly keep adding hosts and essentially leak memory. fixes envoyproxy#433
envoyproxy#389 did not actually cache them between filters which was the point of the previous change.
…iterating locally. (envoyproxy#455) It's somewhat distracting to be forced to fix_format while debugging.
envoyproxy#456) c-ares needs level triggering (envoyproxy#143).
If we get additional CDS updates while we are still initializing, we would not previously initialize new secondary clusters. This would cause init to never complete in certain cases. This commit also adds more init logging to help with debugging similar issues.
…y#563) See code comments for more info.
envoyproxy#558) Baby steps towards flow control. This patch plumbs the TCP connection buffer limits to the listener and enforces it on the read buffer only. Later patches will also plumb to the upstream cluster connection and cover the write buffer (which requires the watermark API).
…roxy#575) This is in support of Bazelifying the build (envoyproxy#415).
…xy#569) This commit cleans up a few leftovers from envoyproxy#559 .
…. (envoyproxy#571) As with fd58242, but on the upstream cluster side.
Previously we did not correctly handle "expect: 100-continue" in the http/2 codec. Because of protocol differences, we handle 100-continue directly in the codec. This change implements support in the http/2 codec similar to http/1.1 where we immediately send a continue response. This commit also adds partial support for the client codec for 1xx responses. The code will raise decodeHeaders(...) twice in the 1xx response case. This is good enough for testing, but will probably crash higher layers if a client actually responded this way. Right now we effectively trust clients so this can be handled in a follow up if necessary. fixes envoyproxy#561
* Minimal Bazel infrastructure (envoyproxy#415). This patch introduces the minimal Bazel infrastructure required to build and execute utility_test, as a step towards conversion of the cmake build system to Bazel (envoyproxy#415). The main contributions are: * Machinery for declaring Envoy C++ library and test targets (envoy_cc_library, envoy_cc_test). * External dependency management that works both inside the Docker CI environment (using prebuilts) and with developer-local dependency fetch/build (based on envoyproxy#416). * Handling of the implicit dependencies that are today sourced via prebuilts. * Simple example of building and running a unit test with only its dependencies built. With the cmake system, we would need to build all of Envoy and all tests to run utility_test. E.g. blaze test //test/common/common:utility_test This is not intended to be used by anyone other than those working with the Bazel conversion at this point. The plan is to add "ci/do_ci.sh bazel.debug" as a Travis target to ensure we don't bit rot.
This bit rot crept in during in b8b1563.
We don't want debug code in the coverage stats.
…y#584) This prevents strange behavior from build/CMakeCache.txt.
…d and connect. (envoyproxy#581) Add getUnusedPort in test/test_common/network_utility.*, which for now just selects a port at random in the ephemeral range. Fix bug in Ipv6Instance ctor where it specified AF_INET (i.e. IPv4) as the address family. Doh! Add testSocketBindAndConnect, and apply to both Ipv4Instance and Ipv6Instance.
…o be used in IpList (or its successor), as that only supports IPv4 ranges today. (envoyproxy#576)
…oxy#592) Previously, we would get all buffer slices on the stack and stack overflow if the chain was very large. This change limits the number of slices that we write during each iteration. There are a number of potential improvements possible in terms of collapsing small moves, enforcing fairness in terms of number of iterations, etc. This commit also removes a bunch of RawSlice to evbuffer_iovec conversions for performance reasons. NOTE: The added test repros the crash before the fix. fixes envoyproxy#585
Also added a "set -e" to configgen.sh to ensure we catch this in presubmit in the future. As a bonus, plumbed in arbitrary address binding to admin server config, since we had some configs that had been updated to this already and it seems a generally useful thing to have.
This breaks statsd. I also took the opportunity to plumb through Stats::Scope properly in a bunch of places which will be needed for LDS.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.