Repository navigation
test(crew-log): drop three stale strict xfails on full-scan guards - #24485
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 Nothing to check. The fixes behind these three markers have merged, and with the markers gone each guard runs as an ordinary test and enforces its bound. [DESIGN-REVIEWED] 2b0ae14 Description read[DESCRIPTION-READ] 6d34cac68caf7f25a64444b0361cb7173c7dea32b22b7f746e0911b4bbbe076e 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: |
buluoray
left a comment
There was a problem hiding this comment.
Approving 2b0ae14. This fixes the red on main: #24261 (merged 15:05Z) and #24268 (merged 15:29Z) made three strict xfails in test/test_crew_log_fullscan_guard.py pass, so main's Backend Tests (3.12, 2) and Windows (2) fail on XPASS(strict) for test_a_cold_prime_copies_the_fold_state_a_fixed_number_of_times. The diff only removes those three markers and keeps every assertion. Locally at this head the file runs 5 passed, 9 xfailed, 0 failed. #24482 carries the identical diff.
bolichen97
left a comment
There was a problem hiding this comment.
Reviewed at head 2b0ae14: looks good to ship.
The diff deletes exactly three xfail(strict=True) decorators in test/test_crew_log_fullscan_guard.py and leaves every test body unchanged. The guarded fixes (#24261 and #24268) are in this branch's base, and Backend Tests shard 2 on Linux and Windows, the shards red on main, is green here. #24482 carries the identical diff; this one is preferred because it closes #24484.
Not approving yet: Backend Tests shards 7 and 8 lost their self-hosted runner mid-run (no test failure in either log), so the readiness check has not passed on this head. The failed jobs were rerun once. Approval and auto-merge remain once readiness passes.
Problem / Motivation
Goal: Turn main's Backend Tests (3.12, 2) and Windows (2) shards green again by dropping three strict xfail markers whose fixes have merged.
Closes #24484
test/test_crew_log_fullscan_guard.pymarks each full-scan guardxfail(strict=True)until its fix lands. Three fixes have landed, so those guards now pass and strict mode reports each pass asXPASS(strict), failing the shard on every PR based on main (seen on #24468):test_a_cold_prime_copies_the_fold_state_a_fixed_number_of_timestest_a_ws_connect_does_not_fold_every_open_slottest_a_ws_reconnect_with_nothing_new_reads_a_fixed_number_of_entriesWhy it matters
Every PR on current main is red on those two shards, and the Coverage Gate and PR Readiness follow, for a reason none of them caused.
Not a goal
What changed (motivation → approach → change)
The three
@pytest.mark.xfail(strict=True, ...)decorators above are deleted; each test body is unchanged, so the guard now runs as an ordinary test and fails if the full scan comes back. Same shape as #24467, which dropped the first-history-page marker.Backwards compatibility
Compatible: test-only change.
Removes nothing: three xfail markers on tests whose fixes merged; the tests themselves stay and now enforce their guard.
Tests
Proof: not provable -- test-only change
On current main (279c84f),
pytest test/test_crew_log_fullscan_guard.py:XPASS(strict)on the two WS guards), 3 passed, 9 xfailed; CI also showed the cold-prime guard asXPASS(strict)-n 0, and once with-n 4(xdist, a different order)flake8,isort --check-only,black --check: clean.test_flake_pattern_ratchet.py: 60 passed.Determinism: the changed tests count entries read and state copies through the file's
meterfixture; they use no clock, sleep, port or randomness, and the diff adds no new test code.Manual verification
N/A — test-only change, covered above.
Related Issues
Closes #24484. Same class as #24467.
Checklist
feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)Contribution License Agreement
Contributed under the project's Apache-2.0 license.