| commit | 797325f77839b1e141d4fe19d816ab6307ee8234 | [log] [tgz] |
|---|---|---|
| author | ulfjack <ulfjack@google.com> | Thu Oct 17 06:33:40 2019 -0700 |
| committer | Copybara-Service <copybara-worker@google.com> | Thu Oct 17 06:34:39 2019 -0700 |
| tree | 676d5e5f8a77185fbec0f6638ea4e12870a406c5 | |
| parent | f6c4b698ac48f743ecef1a9b14eb01380ddf4b9b [diff] |
Post test result for cancelled runs We create the equivalent of an empty FailedAttemptResult with a status code of INCOMPLETE, and post that in the case of a cancelled concurrent test action. We add a new method on TestRunnerSpawn to finalize a cancelled test. Also add test coverage to check that a TestResult with the correct status is posted for each action. Why INCOMPLETE? We have three options here: 1. use INCOMPLETE, 2. use NO_STATUS, or 3. add a new status code to BlazeTestStatus. The other existing enum values are all inappropriate: PASSED, FLAKY, TIMEOUT, FAILED, REMOTE_FAILURE, FAILED_TO_BUILD, and BLAZE_HALTED_BEFORE_TESTING About option 2: NO_STATUS would not allow us to distinguish cancelled tests from tests that were not run in the first place. INCOMPLETE is currently unused and undocumented, so it's unclear what the originally intended semantics are. Before open sourcing Bazel, Google had an INTERRUPTED status, which got translated to INCOMPLETE in the process. About option 3: This would require updating all the downstream systems to support the new status code, including the BEP, which has the exact same list of status codes as BlazeTestStatus. CancelFuture handling? It might be better to push down the cancelFuture handling to the TestAttemptContinuation - that'd allow us to collect additional data about the test attempt and simplify this code. Possibly, we could only push down interrupt handling, but we probably need to know that the interrupt is due to a cancelled run rather than a user interrupt. PiperOrigin-RevId: 275246041
{Fast, Correct} - Choose two
Build and test software of any size, quickly and reliably.
Speed up your builds and tests: Bazel rebuilds only what is necessary. With advanced local and distributed caching, optimized dependency analysis and parallel execution, you get fast and incremental builds.
One tool, multiple languages: Build and test Java, C++, Android, iOS, Go, and a wide variety of other language platforms. Bazel runs on Windows, macOS, and Linux.
Scalable: Bazel helps you scale your organization, codebase, and continuous integration solution. It handles codebases of any size, in multiple repositories or a huge monorepo.
Extensible to your needs: Easily add support for new languages and platforms with Bazel's familiar extension language. Share and re-use language rules written by the growing Bazel community.
Follow our tutorials:
See CONTRIBUTING.md
