Repository navigation
fix(ci): daily-update failed on hook-completeness test for fresh newcomers - #14
Conversation
…ify (regression test)
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: d742c58384
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| fresh = dict(block["items"][0], name="brand-new/never-seen", repo_id=987654321, | ||
| description="A brand new repo that no hook or translation has seen yet.") |
There was a problem hiding this comment.
If the AI search returns no qualifying repositories or all its queries fail, tools/newcomers.py::discover_category legitimately produces items=[], but this new test raises IndexError while constructing its fixture. The daily-update workflow runs verification after discovery and before committing, so an empty AI board now prevents updates to every other board from being committed. I reproduced the failure by supplying an empty AI items list to this test. Construct a self-contained newcomer fixture and explicitly set the test block's success state instead of copying its first live item.
Useful? React with 👍 / 👎.
Cause
Run 41122642510 (schedule, Beijing 2026-10-31 04:22, head 64e0eae) failed in
python3 tools/verify.py:FAIL: test_every_ranked_new_repo_has_a_hook_or_is_listed_as_pending (test_hooks.LengthRuleTests)/AssertionError: Lists differ: ['ccch1mneyyy/dsh-Go', ...] != []/First list contains 22 additional elements.PR #13 added a test that asserted every ranked new repo already has a hook. The scheduled run discovers new repos first and verifies right after; CI has no model key (
CURSOR_API_KEY not set: translation will be skipped.), so fresh repos never have hooks, the test failed and nothing was committed. Earlier runs (before #13, 18:31Z) passed.Fix (tests only, no data/translation/GEO/⚠ changes)
tests/test_hooks.py: removed the live-data completeness assertion (its own docstring said "informational"). Added regression tests: a brand-new newcomer without hook/translation renders in all five languages with the original text, is listed byhook_gaps; and the old assertion cannot come back. Gaps print as a note.docs/project-governance/CHANGELOG.md: entry.Verification
python3 tools/verify.py-> 201 tests OK, All verification commands passed.Not merged. Needs a merge to main to take effect, so no manual workflow trigger.