test(grounding): main is red, flip the two rows the self-claim fix closed #565

Closed
coilyco-ops wants to merge 1 commit from fix/main-is-red-on-the-corpus into main
Member

closes #563

Main is failing ward exec test right now. Merge this ahead of anything else of mine.

--- FAIL: TestGroundingRejectsUngroundedActionClaims
    behavior changed on "Sirens Echo filed a correction.": rejected = true, was false.
      If issue 241 was fixed, set rejectedNow to true and clear the issue field

Two changes landed close together and neither is wrong on its own: QA's twelve-shape corpus recorded those rows as not-yet-rejected, and PR#556 made them reject. Each passed its own gate. The merge is red because the corpus asserts current behaviour and the fix changed current behaviour.

Neither author could have seen this from their own branch.

The change

Flip two booleans, clear two issue fields. Exactly what the failure text instructs.

The corpus worked as designed. It named the rows, the field, and the value, so this took no investigation — which is the argument for characterization tests that fail loudly, not against them.

Third time today two green branches summed to a red main, and the strongest evidence yet for the merge-time check in #537.

ward exec gate green.

closes #563 **Main is failing `ward exec test` right now.** Merge this ahead of anything else of mine. ``` --- FAIL: TestGroundingRejectsUngroundedActionClaims behavior changed on "Sirens Echo filed a correction.": rejected = true, was false. If issue 241 was fixed, set rejectedNow to true and clear the issue field ``` Two changes landed close together and **neither is wrong on its own**: QA's twelve-shape corpus recorded those rows as not-yet-rejected, and PR#556 made them reject. Each passed its own gate. The merge is red because the corpus asserts current behaviour and the fix changed current behaviour. Neither author could have seen this from their own branch. ## The change Flip two booleans, clear two issue fields. Exactly what the failure text instructs. **The corpus worked as designed.** It named the rows, the field, and the value, so this took no investigation — which is the argument for characterization tests that fail loudly, not against them. Third time today two green branches summed to a red main, and the strongest evidence yet for the merge-time check in #537. `ward exec gate` green.
test(grounding): main is red, flip the two rows the self-claim fix closed
All checks were successful
ci / image-build (pull_request) Successful in 23s
ci / test (pull_request) Successful in 38s
ci / publish-echo-image (pull_request) Has been skipped
ci / publish-observed (pull_request) Has been skipped
6f3659b56e
QA's corpus recorded two rows as not-yet-rejected against issue 241. The
self-claim tense fix landed and makes both reject. Each change passed its own
gate and their merge is red, because the corpus asserts current behaviour and
the fix changed it.

The corpus worked as designed. Its failure text names the rows, the field, and
the value to set, so this needed no investigation at all.

Third time today that two green branches summed to a red main.

Co-authored-by: Kai Siren <coilysiren@gmail.com>
Co-authored-by: Claude <noreply@anthropic.com>
coilyco-ops 2026-08-13 15:58:07 +00:00
  • closed this pull request
  • requested review from coilysiren
All checks were successful
ci / image-build (pull_request) Successful in 23s
ci / test (pull_request) Successful in 38s
ci / publish-echo-image (pull_request) Has been skipped
ci / publish-observed (pull_request) Has been skipped

Pull request closed

Sign in to join this conversation.
No reviewers
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
coilyco-gaming/sirens-echo!565
No description provided.