Repository navigation
ci: register the nine unregistered Python tests - #13712
teamleaderleo wants to merge 1 commit into
Conversation
`validate_test_execution_registry.py` fails on main: nine tests exist with no registry entry, which fails workflow-guard-tests/preflight, then Guard status, then ci-status on every open pull request. The tests arrived with their own PRs; none of them added the registry entry the guard requires. Seven already run in ci-guards.yml and are registered as linux-guard. Two run in other workflows -- test_ci_actionlint_covers_every_workflow.py in testbox-broker-guard.yml and test_ci_selective_layer_wiring.py in ci-artifact-transport.yml -- and the registry has no lane for that, so they enter the legacy inventory lane the same way the other tests those two workflows run already do. tests/test_ios_screenshot_capture_guard.py ran nowhere at all. It is a static guard on the fastlane screenshot lane and passes on Linux with no iOS tooling, so wire it into ci-guards.yml beside the other release-ios guards rather than inventory a test that never executes. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
|
All contributors have signed the CLA ✍️ ✅ |
|
Navigate logical layers of code changes, visualize relationships, and explore their blast radius. No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Repository: manaflow-ai/cmux/.coderabbit.yaml Review profile: ASSERTIVE Plan: Advanced Run ID: 📒 Files selected for processing (2)
Included review availability: Your plan provides up to 10 included reviews per hour; 1 remains after this review. 📝 WalkthroughWalkthroughThe CI workflow now runs an iOS screenshot capture guard for ChangesCI guard coverage
Priority: ➖ Normal Estimated code review effort: 1 (Trivial) | ~5 minutes Change: Bug fix Merge Risk: ⚪ Minimal · up to The change expands CI coverage with an iOS guard and nine registry entries, with no current evidence of production impact or merge-blocking risk. 🚥 Pre-merge checks | ✅ 25✅ Passed checks (25 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
|
This duplicates #13710, opened five minutes earlier, and #13710 is the better of the two to keep. Both register the same nine paths. The difference is the lane they give two of them:
The reason they cannot be #13710 is also the base of a stack: #13714 ( Suggest closing this one and landing #13710, then #13714. For context on urgency: the registry guard landed today at 12:34 in #13615, and nine tests have reached |
|
Closing as stale — every change here is already on Verified against
So the diff would be a no-op plus a duplicate guard step. What is still true and worth keeping. Two things from the analysis here outlive the patch:
I have carried both points onto #13745. Nothing here is lost; reopen if I have misread the registry state. |
validate_test_execution_registry.pyfails onmain. Nine tests exist with no registry entry, which failsworkflow-guard-tests / preflight, thenguards / Guard status, thenci-status— on every open pull request, regardless of what the PR changes.Reproduced on unmodified
main(1847ef0621):Each test arrived with its own PR and none added the entry the guard requires.
How each was classified
I checked where every one actually runs rather than assigning a lane by shape.
linux-guardci-guards.yml, which is what that lane requireslegacytestbox-broker-guard.ymlandci-artifact-transport.yml; the registry models no lane for those, and the other tests those workflows run are alreadylegacytests/test_ios_screenshot_capture_guard.pyran nowhere — no workflow invoked it. It is a static guard over the fastlane screenshot lane's config and passes on Linux with no iOS tooling, so this wires it intoci-guards.ymlbeside the otherrelease-iosguards instead of inventorying a test that never executes. That is a step, not just a registry line, so it is visible in the diff.Validation
tests/test_ios_screenshot_capture_guard.pyandtests/test_ci_workflow_guards_are_wired.pyboth pass locally. Fulltests/test_ci_*sweep shows no new failures againstmain(test_ci_change_areas.pyandtest_ci_sparkle_build_monotonic.shfail identically on unmodifiedmainhere — they shell out toghand to macOS-only tooling).The two
legacyentries are inventory, not an endorsement: modelling a real lane for tests owned bytestbox-broker-guard.ymlandci-artifact-transport.ymlis worth doing separately.🤖 Generated with Claude Code
Need help on this PR? Tag
@codesmith-botwith what you need. Autofix is disabled.Summary by cubic
Registers the nine Python tests missing from the test execution registry so the
workflow-guard-tests / preflightguard and theci-statuschecks it feeds stop failing on every open PR.ci-guards.ymland are registered aslinux-guard.testbox-broker-guard.ymlandci-artifact-transport.ymland are registered aslegacy, matching the other tests those workflows run.tests/test_ios_screenshot_capture_guard.pyran nowhere; it's now wired intoci-guards.ymlbeside the otherrelease-iosguards instead of being inventoried as a test that never executes.Written for commit 16dcac4. Summary will update on new commits.
Summary by CodeRabbit
Tests
Chores