Repository navigation
test: record the locale files one per line in the codex STT message test - #37689
Conversation
UX Review (Opus 5.5) — ⏭️ skippedRevision |
First Principles Review (Opus 5.5) — ⏭️ skippedRevision |
Design Review (Opus 5.5) — ✅ PASSDesign-level review of Design-Verdict: PASS Test-only change. An exact per-line record replaces the count, so concurrent locale additions now conflict or fail loudly instead of merging silently. [DESIGN-REVIEWED] 6cf326c Description read[DESCRIPTION-READ] 76d9c2c204bcc61ddf10ad832d7581a39691317094aaaab540596b297a40f682 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 Review detailsCandidate is a doc reference, not runtime defect; drop it. No findings. [OPUS-REVIEWED] 6cf326c 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: |
Code change approver recommendation: ApproveWhy: the decider estimates a 4% chance that this change introduces a defect, within the 12% limit for approving on its own, and every check passes. This PR changes one test file and no product code. It swaps An automated recommendation, not a review: a maintainer still decides. It approves on its own only when the estimated chance of a defect is at most 12% and the change has at most 400 changed lines. Decider v3c (217fe607) · commit |
KiroCrewCaptain
left a comment
There was a problem hiding this comment.
I reviewed head 6cf326c. The change swaps the hard-coded locale count in the codex STT message test for an exact list of catalogue file names, one per line. The 13 names match the locale directory on current main (excluding en.json), in the order the test's sort produces. I merged this head with current main and ran test/test_codex_stt.py: 86 passed. The new assertion is stricter than the old count, there is no product code change, and I found nothing that blocks.
Problem / Motivation
Goal:
test/test_codex_stt.py::test_extra_missing_message_names_no_install_commandstays green onmainwhen two pull requests that each add a locale both merge.The test pins the number of locale catalogues with
assert len(files) == 13. Two PRs that each add a locale both rewrite it to14, git merges the identical edit cleanly, andmainhas 15 files. It is one of the sites the K21 shared-total guard in PR #37660 counts onmain.Why it matters
Locales are added rarely, but when two land close together this red
mainfails every open PR until someone re-pins the count.Not a goal
What changed (motivation → approach → change)
assert len(files) == 13_LOCALE_FILES, one catalogue name per line, andassert [p.name for p in files] == list(_LOCALE_FILES)The rule is the one in
docs/system-specs/common/testing-conventions.md§ Merge-safe ratchets and counts (PR #37660).Backwards compatibility
Compatible: test file only.
Removes nothing: the list has 13 entries, the old count.
Tests
Two-branch merge simulation (scratch worktrees; A adds
aa.json, B addszz.json, each updating the record the way the test asks, thengit merge):git mergeorigin/mainace367bassert 15 == 14Negative control on this branch (a locale file added, record not updated): FAILED.
test/test_codex_stt.py: 86 passed; black, isort and flake8 clean.Proof: not provable -- test-only change
Manual verification
N/A — test only; the simulation above is the behavioral check.
Related Issues
Part of #37656
Checklist