Skip to content

Retry flaky Z2MIntegrationTests once on Full CI#103

Open
tashda wants to merge 1 commit into
mainfrom
fix/ci-full-flaky-retry
Open

Retry flaky Z2MIntegrationTests once on Full CI#103
tashda wants to merge 1 commit into
mainfrom
fix/ci-full-flaky-retry

Conversation

@tashda
Copy link
Copy Markdown
Owner

@tashda tashda commented May 6, 2026

Summary

  • Nightly Full CI has been red most of the last two weeks. The April 25–May 4 streak was a real test/bridge contract bug fixed in v1.6.0 (6a25a9d). Last night (May 6) failed on a different, genuinely flaky test: testReceivesBridgeDevices timed out at client.connect(url:) for 18.9 s. The mock bridge replays ~173 retained messages on every new WS client, and back-to-back integration tests occasionally race the previous client's teardown on the shared GitHub runner.
  • Add -retry-tests-on-failure -test-iterations 2 to the unit+integration xcodebuild step so a one-off flake gets a single retry. Genuine regressions still fail twice in a row and surface red.
  • Workflow-only change, no app code touched.

Test plan

  • Manually dispatch CI (Full) from the Actions tab on main after merge and confirm the unit+integration job is green.
  • Watch tomorrow's scheduled 03:00 UTC run.

🤖 Generated with Claude Code

Pass -retry-tests-on-failure -test-iterations 2 to the unit+integration
xcodebuild step. Real regressions still fail twice; one-off WS connect
timeouts to the mock bridge (races between back-to-back tests and the
bridge replaying ~173 retained messages on each new client) absorb a
single retry.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
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