Closed jdelapla closed 7 months ago
Attention: 20 lines
in your changes are missing coverage. Please review.
Comparison is base (
792e0db
) 80.42% compared to head (c3fb240
) 80.40%.:exclamation: Current head c3fb240 differs from pull request most recent head 248c90f. Consider uploading reports for the commit 248c90f to get more accurate results
Files | Patch % | Lines |
---|---|---|
src/utils/src/Threadpool.c | 73.68% | 20 Missing :warning: |
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
Can we open a PR in WebRTC to confirm the CI there passes too?
Issue #, if available: N/A
Description of changes: ThreadData->mutex now wraps all interactions with any threadpool objects, this includes the queue. As a consequence of having the data mutex locked while waiting on the queue, the threadpool couldn't teardown if an actor was waiting on the queue, so the teardown now pushes wait tasks to the queue in the event of that use case.
Additionally all utils tests were changed so that if a memory leak is detected, it will infinitely wait for the memory leak to end -- resulting in a GHA timeout or a loose actor cleaning up the memory (common race condition specific to tests only).
By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.