Repository navigation
Conversation
Stage-2 reproduction for repoprompt#1452. Await a fixed diagnostic publication before unsubscribing and assert approval identity remains correct. Intentionally retain the original failing exact-count assertion; no production or semantic-oracle fix is included.
Preserve the ordered approval identity prefix and exact decisions while checking every trailing snapshot remains fully cleared. Require the controlled diagnostic marker within that suffix. Fixes repoprompt#1452.
Require non-default diagnostics on trailing cleared snapshots so an extra clear cannot masquerade as a legitimate diagnostic publication. Keep the ordered identity prefix, marker and exact decisions unchanged.
bcgalvin
marked this pull request as ready for review
October 23, 2026 01:04
baron
approved these changes
Oct 23, 2026
baron
left a comment
baron
left a comment
Collaborator
There was a problem hiding this comment.
Reviewed the PR diff; no actionable defects found.
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fixes #1452
Summary
The two-window approval test required exactly five state snapshots. The controller posts denial diagnostics from an unawaited task, and diagnostic updates can add a full snapshot before unsubscription.
Check the exact five-snapshot approval prefix, then require fully cleared approval fields and non-default diagnostics on every trailing snapshot. Await a diagnostic marker before unsubscribing so the trailing case is always exercised. In this fixture, an extra clear before diagnostic publication would still fail.
Test-only; production publication is unchanged. The distinct-generation and decision assertions remain. The fixture has no process principal, so nil-path checks cover clearing, not non-nil propagation. The controlled red demonstrates the brittle count assertion, not the historical publisher.
Validation
Candidate:
1585ce45b858b24ebafd6166ccf5249288ecf573.ab84d47eplus the awaited marker (recorded later as0560ab84): 1 test, 1 failure at the old equality, observing[nil, 1, nil, 2, nil, nil]instead of five entries.RepoPromptTests.MCPConnectionAdmissionTests/testTwoWindowServiceResponsesCannotResolveSuccessorGeneration: 1 test, 0 failures. Executed source is byte-identical to the candidate.pr-readyfailed: 6,598 tests executed, 8 skipped, 1 failure inAgentTranscriptPaneSnapshotPublicationTests/testTargetReplacementDropsOutgoingRowsFromPublishedSnapshotatAgentTranscriptRestorationPresentationTests.swift:2628(XCTAssertFalse failed - precondition: outgoing rows), the exact signature tracked in Flaky test: testTargetReplacementDropsOutgoingRowsFromPublishedSnapshot fails when run more than once in a test process #1428. This branch does not touch that test or its fixture; its relationship to this change remains unproven. The MCP approval test rejects an extra cleared-state snapshot #1452 test passed in that run. No isolated retry was run.