ci(benchmarks): restore lifeops-quality bench + CI gate deleted by #11271 - #11530
Conversation
…1271 (#11376) Re-lands packages/benchmarks/lifeops-quality (20 files) and .github/workflows/lifeops-quality-bench.yml byte-identical from the pre-clobber parent 5b714c7^. All three workflow lanes verified green locally against current develop: unit 27/27, triage gate 6/6, timeliness gate 1/1 (both DST windows). The 4 orphaned benchmark tests from the same #11419 row are excluded: already restored (and partly evolved) by merged #11490. Refs #11376 #11419 #11271 [cloud-security]
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
There was a problem hiding this comment.
Your trial has ended. Reactivate Greptile to resume code reviews.
|
Important Review skippedAuto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
✨ 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 |
lalalune
left a comment
lalalune
left a comment
There was a problem hiding this comment.
Verified the restore/syntax/path side locally. The restored lifeops-quality package and workflow are byte-identical to the pre-clobber 5b714c7^ paths, actionlint .github/workflows/lifeops-quality-bench.yml passes, PyYAML safe-load passes, git diff --check origin/develop...HEAD passes, and the workflow's referenced repo paths/action inputs exist in this PR worktree. I did not rerun the full benchmark suite locally in the detached worktree because it has no installed workspace deps; the PR's pending lifeops-quality-bench Actions run remains the execution proof gate.
Restores the UI Fixture E2E workflow (125 lines, byte-identical to the pre-#11271 state at 5b714c7^) that gated the ten packages/ui fixture e2e runners (agent-surface bridge, bottombar, ambient background, background/launcher/connectors pages, view lifecycle, ftu-home, orchestrator-accounts, fused-wake) on packages/ui changes. It was deleted wholesale by the #11271 squash and is the last of the five launch-critical workflow files named in the #11419 tracker without a restore (test.yml / scenario-pr.yml / kokoro-real-smoke.yml restored by 19b40d1; lifeops-quality-bench.yml restored by #11530). All ten 'test:*-e2e' scripts the workflow invokes verified present in packages/ui/package.json on current develop. Refs #11419 Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
There was a problem hiding this comment.
Your trial has ended. Reactivate Greptile to resume code reviews.
|
Claude encountered an error —— View job I'll analyze this and get back to you. |
LifeOps Multi-Tier BenchmarkSuite:
|
2 participants
What
Restores the
packages/benchmarks/lifeops-qualitybenchmark package (20 files) and its CI gate.github/workflows/lifeops-quality-bench.yml, deleted by the #11271 stale-base squash (5b714c74e6). All files re-landed byte-identical from the pre-clobber parent5b714c74e6^.Slice tracked by #11376 (claimed there before mutating) / umbrella #11419. #11444 explicitly deferred the workflow restore to this slice because the bench package was still absent.
Restored paths
packages/benchmarks/lifeops-quality/— entire package:package.json,baseline.json,budgets.json,vitest.unit.config.ts,vitest.gate.config.ts,results/.gitignore, docs (README.md,AGENTS.md,CLAUDE.md),triage/(corpus, fixtures, metrics + tests + gate) andtimeliness/(corpus, oracle + tests + gate).github/workflows/lifeops-quality-bench.yml— the PR/nightly benchmark gateNot included: the 4 orphaned benchmark tests from the same #11419 row (
interrupt-bench/tests/honest-scoring.test.ts,three-agent-dialogue/__tests__/verification.test.ts,orchestrator_lifecycle/{events.py,tests/test_events.py}) — already restored by merged #11490; two of them evolved past the pre-clobber blob there (controlActionsupport), so re-landing them from5b714c74e6^would regress #11490. Verified by blob-hash before excluding.CI workflow: included — validated, not assumed
Since the .yml re-adds a PR-triggered gate, I verified it end-to-end before including it:
python3 yaml.safe_load).plugins/plugin-inbox/src/inbox/,plugins/plugin-personal-assistant/src/lifeops/,plugins/plugin-personal-assistant/test/helpers/,plugins/plugin-scheduling/src/,packages/core/src/services/triggerScheduling.ts,.github/actions/setup-bun-workspace(all 5 inputs it passes still exist in the action),packages/scripts/run-turbo.mjs.actions/checkout/setup-nodeSHAs used 100+ times across.github/workflows;upload-artifact@v7used elsewhere).bun run test(unit lane): 27/27 passedbench:triage(real classifier gate): 6/6 passedbench:timeliness(real PGlite scheduler-tick gate over both DST windows): 1/1 passed (both DST windows replayed, ~58s locally)package.jsoninternal consistency: both vitest configs it references are restored; name@elizaos/lifeops-quality-benchis unique in the repo; the package was never a root-workspace member pre-clobber either (rootpackage.jsonbenchmark workspace globs are identical pre/post), so no root manifest change is needed — CI resolves vitest from the hoisted root install exactly as it did before refactor(cloud): fold rejectDelivered into the fenced generic refund() + concurrent refund-race tests (#11167) #11271.Note: several of the bench's cross-repo dependencies (
triage-classifier.ts,lifeops/repository.ts, PAtest/helpers/runtime.ts,vitest.src-integration.config.ts) are still at the clobbered blob on develop (the PA source restore is the remaining open part of #11376) — the gates above pass against that current state, so the gate is green now and will keep guarding the PA slice as it gets restored.Refs #11376, #11419, #11271
[cloud-security]