[Mikola Lysenko (mikolalysenko)]

Problem

Every lint-ecosystems run (PR, merge queue and main push) posts 10 fake notice annotations, for example on main's latest push run (check run 117423824414 at 7d3086f):

Could not verify merge-queue CI (TimeoutExpired); running all checks.
Could not compare with the previous PR head (CalledProcessError); running all checks.
… 10 in total

Nothing failed. The reuse scripts never run in that job. These lines come from the job's python3 -B -m unittest discover -s scripts/tests step: the fallback self-tests let ci-reuse-merge-group.py and ci-reuse-pr-head.py print their real ::notice:: workflow commands to stdout, and GitHub turns those lines into annotations. A local run of the suite prints 15 of them, and GitHub keeps 10 (its per-step cap).

The harm:

• A reader can take the annotations to mean the merge-queue reuse check or the PR-head reuse check failed, which sends anyone debugging a CI-reuse decision down the wrong path.
• They fill the step's notice cap, so a real notice from that step would be dropped.

CI is otherwise healthy this run: the merge queue is empty, nothing is queued or running, main is green, and the only failed run in 3 days (38806080761) was fixed by that PR's next push. So this falls under priority 4 (make CI more robust).

Root cause

test_api_failure_timeout_and_bad_payload_* in scripts/tests/test_ci_reuse_merge_group.py and scripts/tests/test_ci_reuse_pr_head.py call reusable_run on its error path without capturing stdout.

Fix

A small assert_falls_back helper in each test file. It runs reusable_run under contextlib.redirect_stdout, asserts that the result is None (unchanged), and also asserts that the ::notice:: fallback line was printed. The fallback is now checked, not just silenced. No production script or workflow changes.

Proof

• python3 -B -m unittest discover -s scripts/tests: 347 tests, OK (6 skipped). Lines starting with :: in stdout went from 15 to 0.
• Negative check: removing ::notice:: from ci-reuse-merge-group.py makes the new assertion fail 7 subtests.

Removed or moved tests

None. Every test still runs in the same place (lint-ecosystems).

https://claude.ai/code/session_01SeHhBxEvePf8ERTaTTenMK

---

Generated by Claude Code

[Mikola Lysenko (mikolalysenko)]

bugbot run

---

Generated by Claude Code

[cursor]

✅ Bugbot reviewed your changes and found no new issues!

Comment @cursor review or bugbot run to trigger another review on this PR

^{Reviewed by Cursor Bugbot for commit 564c31a. Configure here.}

[Mikola Lysenko (mikolalysenko)]

Labeled Ready for review (burn-down agent).

• Head: 564c31a
• CI: 54/54 green (35 success, 19 skipped). No main-wide failures.
• Bugbot: reviewed 564c31a, no findings. 0 open threads.
• Mergeable with no conflicts. CHANGELOG.md is not touched.
• Reviewer focus: only the two reuse self-tests under scripts/tests/ change; check they no longer emit real ::notice annotations.

Slack announcement not sent this run (the Slack send tool was unavailable). The next run will retry.

---

Generated by Claude Code

[Mikola Lysenko (mikolalysenko)]

The merge queue removed this PR at 17:49 UTC because e2e (ubuntu-latest, e2e_vex_build, pipenv:: --ignored, 2026.8.0) failed (run 39467483928). That failure doesn't come from this PR:

• The queue group was main 14eee88 plus this PR, which changes only two Python self-tests under scripts/tests/.
• The failing test is pipenv::pipenv_pylock_projects_wire_what_pipenv_installs. It fails with uv pip install -r pylock.toml → The target environment is not supported by the lockfile.
• Before this PR, the same leg evicted Fix npm workspace members named build/dist skipped (#1627) #1628 and Fix requirements pin spellings not contesting VEX (#1602) #1632 from the head of the queue. Last night it also failed in main's nightly run.

Cause: uv 0.13.2 (released 2026-10-19) rejects the environments = [] that Pipenv 2026.8.0 writes, and the Pipenv/pip legs install the latest uv. I bisected it locally: 0.13.1 accepts the file and 0.13.2–0.13.4 reject it.

#1639 pins setup-uv to 0.13.1 and fixes this. I'm not porting the pin here, so the fix doesn't land twice. Once #1639 merges, this PR can go back in the queue unchanged, and its approval still stands.

---

Generated by Claude Code

[Mikola Lysenko (mikolalysenko)]

[final reviewer] Re-added to the merge queue (squash, auto-merge) at 564c31a13, which is the commit Wenxin Jiang (@Wenxin-Jiang) approved. The earlier eviction was the main-wide uv 0.13.2 Pipenv pylock break. main now pins setup-uv to 0.13.1 (#1605), and the same e2e … pipenv:: --ignored, 2026.8.0 leg is green on main a54753a. This is the one re-enqueue after that eviction.

---

Generated by Claude Code