Skip to content

[CI] Parity: use authoritative flaky-test summaries - #3560

Open
ethanwee1 wants to merge 1 commit into
ethanwee/parity-preview-replacementfrom
ethanwee/parity-flaky-develop
Open

[CI] Parity: use authoritative flaky-test summaries#3560
ethanwee1 wants to merge 1 commit into
ethanwee/parity-preview-replacementfrom
ethanwee/parity-flaky-develop

Conversation

@ethanwee1

@ethanwee1 ethanwee1 commented Aug 11, 2026

Copy link
Copy Markdown

Summary

Parse run_test.py's authoritative failed-then-passed summary instead of attributing flakiness to the last unrelated PASSED line.

Clean replacement for #3431 on current develop, stacked on #3554. Current develop already contains SHA-prefixed log label handling, so the empty duplicate commit was intentionally omitted.

Test plan

  • Compile detect_log_failures.py
  • Verify normal and SHA-prefixed log classification
  • Verify the focused diff only replaces flaky attribution logic
  • Re-run the documented real-log regression from the original PR

Made with Cursor

Alignment landing order

Depends on #3554 only for branch stacking. Land after the downloader resilience PRs; the flaky parser change itself is isolated.

The log-based flaky detector inferred flakiness from the last individual
"... PASSED" line seen before a "Test succeeded in new process" marker. That
heuristic attaches the flaky flag to whatever test happened to pass most
recently, which is often an unrelated test that passed normally much earlier in
the shard -- not the test that actually crashed/failed and then passed on
rerun.

Concrete example (run 29255582284, mi350, inductor/test_compiled_autograd):
test_wrap_recreates_contexts_for_repeated_runs passed normally at ~16% and was
wrongly reported as flaky, while the test that truly failed-then-passed on
rerun was FuncTorchHigherOrderOpTestsWithCompiledAutograd::test_jacfwd.

run_test.py already prints an authoritative per-shard summary naming the exact
tests:

  The following tests failed and then succeeded when run in a new process
  ['test/inductor/test_compiled_autograd.py::...::test_jacfwd']

Parse that line directly instead of guessing. Verified against the real ROCm
and CUDA logs from the run above: test_wrap is no longer flagged and test_jacfwd
is correctly reported on both platforms.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant