fix: avoid duplicate media completion fallback#77754
Conversation
|
Codex review: needs real behavior proof before merge. Summary Reproducibility: yes. by source inspection. On current main, a direct announce response shaped like Real behavior proof Next step before merge Security Review findings
Review detailsBest possible solution: Land the narrow pending-status guard with regression coverage after adding or explicitly folding the user-facing duplicate-media-fallback note into the existing Unreleased media changelog entries. Do we have a high-confidence way to reproduce the issue? Yes by source inspection. On current main, a direct announce response shaped like Is this the best way to solve the issue? Yes for the code path. Recognizing Full review comments:
Overall correctness: patch is correct Acceptance criteria:
What I checked:
Likely related people:
Codex review notes: model gpt-5.5, reasoning high; reviewed against 9c4a335007d7. Re-review progress:
|
byungskers
left a comment
There was a problem hiding this comment.
Clean fix for the race condition between media generation and completion fallback. Extracting into a dedicated predicate makes the intent very clear.
The test case for pending video generation is a good addition — it locks in the behavior that we shouldn't fall back while an announce-agent run is still in flight.
LGTM.
byungskers
left a comment
There was a problem hiding this comment.
Clean fix for the race condition between media generation and completion fallback. Extracting isGatewayAgentRunPending into a dedicated predicate makes the intent very clear.
The test case for pending video generation is a good addition — it locks in the behavior that we shouldn't fall back while an announce-agent run is still in flight.
LGTM.
|
Landed via rebase onto main.
Changelog repair is included in the follow-up refactor diff. |
Summary
Test