Repository navigation
test(website): name each diff-scoped gate wiring instead of counting them - #38296
Conversation
First Principles Review (Opus 5.5) — ⏭️ skippedRevision |
UX Review (Opus 5.5) — ✅ PASSUX-level review of UX-Verdict: PASS Nothing to check. The PR only changes a test file ( [UX-REVIEWED] 56acec8 Description read[DESCRIPTION-READ] cec8d7885ceda04a1a61bd825fbfb9285871f2069edc208a1a71ab72be6ab938 That sha256 names everything this verdict read: the PR description, plus the 3 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with |
Design Review (Opus 5.5) — ✅ PASSDesign-level review of Design-Verdict: PASS Nothing blocks: replacing the shared total with a sorted list of names is the merge-safe ratchet that testing-conventions already describes, and it is aimed at the main break the PR names. Notes
[DESIGN-REVIEWED] 56acec8 Description read[DESCRIPTION-READ] 9517fc7c3b13b53f62b666124b1d2cb67c55013f5aa0a800f4b79acf09defcd0 That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with |
Opus 5.5 Review — ✅ no blocking findingsReviewed Verdict parsed from the review's SHA-scoped output markers for commit False positive or not applicable? A repository writer can comment: |
GPT 6.1 Review — ✅ no blocking findingsGPT 6.1 completed its review of This comment is updated in place on each push. Review detailsNo findings. False positive or not applicable? A repository writer can comment: |
…them EMOJI_ICON_BASE_REF). Only one bumped toHaveLength(13), so the two green PRs together left main at 14 wirings and the test red on every PR. A total is a merge-unsafe ratchet: list the env names one per line, so two PRs that each add a gate edit different lines and git surfaces any overlap.
3aa84e8 to
56acec8
Compare
Problem / Motivation
Goal: Two PRs that each add a diff-scoped CI gate can no longer break main for everyone by both bumping one total in
renderVerdict.test.ts.website/src/test/renderVerdict.test.ts::CI supplies a base commit on both paths > wires every diff-scoped gate stepassertedexpect(wirings).toHaveLength(13). #38250 (METRICS_SCHEMA_BASE_REF) and #38097 (EMOJI_ICON_BASE_REF) each added a wiring; each was green on its own base, but together main had 14 and the test failed on every PR (15 PRs blocked at the 23:16 PST scan) until #38295 bumped it to 14.Why it matters
A total is a merge-unsafe ratchet: every PR that adds a gate edits the same number, so any two that merge close together leave main red for all PRs until someone notices. This is the fourth main break of this kind today.
Not a goal
base.shaassertions below are unchanged).What changed (motivation → approach → change)
Symptom: main red on every PR after two gate PRs merged → root cause: the test pins a total, not the set → change: the test now asserts the sorted list of
*_BASE_REFenv names, one per line. Two PRs that each add a gate edit different lines, so git shows the overlap instead of a silent semantic conflict, and the test still fails when a gate is added or removed without updating the list.Backwards compatibility
Compatible: test-only; the list matches the 14 wirings on main.
Tests
vitest run src/test/renderVerdict.test.ts: 61/61 on main after test: count the fourteen diff-scoped base-ref wirings on main #38295.ci.yml/fast-gate.ymlwithout updating it.Manual verification
N/A — test-only change.
Screenshots / video
N/A — no UI change.