ci: skip macOS shards for docs site changes#1671
Conversation
|
Codex review: needs maintainer review before merge. Reviewed June 20, 2026, 12:20 AM ET / 04:20 UTC. Summary Reproducibility: yes. for the CI behavior: current main requires macOS for docs site code, while the PR allow-lists selected docs site paths. The linked proof run demonstrates a docs/site.css-only change skipping swift-test-macos while the aggregate job passes. Review metrics: 2 noteworthy metrics.
Merge readiness Overall follows the weaker of proof and patch quality, so missing proof can cap an otherwise strong patch. Rank-up moves:
Risk before merge
Maintainer options:
Next step before merge
Security Review detailsBest possible solution: Merge after maintainers accept the expanded docs-site skip policy and exact-head required checks pass; otherwise narrow the allow-list while preserving fail-closed handling for unknown docs code and contract files. Do we have a high-confidence way to reproduce the issue? Yes for the CI behavior: current main requires macOS for docs site code, while the PR allow-lists selected docs site paths. The linked proof run demonstrates a docs/site.css-only change skipping swift-test-macos while the aggregate job passes. Is this the best way to solve the issue? Yes, if maintainers accept the CI policy change. The implementation is narrow, keeps fail-closed behavior for unknown docs code and contract files, and adds focused gate fixtures plus updated aggregate wording. AGENTS.md: found and applied where relevant. Codex review notes: model internal, reasoning high; reviewed against 0f85e05f1c3d. Label changesLabel changes:
Label justifications:
Evidence reviewedWhat I checked:
Likely related people:
What the crustacean ranks mean
Shiny media proof means a screenshot, video, or linked artifact directly shows the changed behavior. Runtime, network, CSP, and security claims still need visible diagnostics. How this review workflow works
|
|
Proof PR for the new path gate: Yuxin-Qiao#4 It changes only
The remaining Linux CLI build jobs were still pending when checked, but the macOS queue behavior is already verified: docs site-only changes no longer request macOS Swift shards. |
|
@clawsweeper re-review |
|
🦞🧹 I asked ClawSweeper to review this item again. |
|
Addressed the stale aggregate wording in
New exact-head CI is running at https://github.com/steipete/CodexBar/actions/runs/27859863170; macOS shards are currently queued as expected because this PR still changes |
Summary
Safety
docs/site.jsremains covered bynode --checkinlint-linuxdocs/custom-tool.jsstill requires macOSValidation
bash -n Scripts/ci_macos_test_gate.shbash -n Scripts/test_ci_path_gate.sh./Scripts/test_ci_path_gate.shgit diff --check./Scripts/lint.sh lint-linuxran portable checks and SwiftLint with 0 violations; local sandbox still returns nonzero when a tool tries to write a macOS plist cache outside the workspace after lintingBehavior proof
cc5d993eff5ed911c69f71688b81ab7631fe8147docs/site.csson top of this PR branchchanges,lint, both Linux CLI builds, andlint-build-testpassed;swift-test-macoscompleted as skipped