Repository navigation
fix(workflow): refuse callback-erased legacy loop publications - #4862
Conversation
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
|
Warning Review limit reachedYou've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository. Next included review available in 56 minutes. View limit detailsLimit details: You’ve used the included review currently available. Review configuration: ⚙️ Run configuration
📒 Files selected for processing (3)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
There was a problem hiding this comment.
Your trial has ended. Reactivate Greptile to resume code reviews.
|
Score: 80/100. Replaces the fragile two-key loop-output heuristic with a fail-closed provenance rule, and the regression matrix is thorough.
No blocking issues found. Approval is conditional on green CI. Generated by Claude Code |
📦 Client bundle boundary
A server module in a client graph aborts hydration in the browser. New leaks fail CI; known leaks are tracked in |
|
Independent review of exact head 51f1607: correctness + spec + quality, 95/100, no blocking findings. Reviewed durable erasure matrix for both callbacks, all two-key/all-key combinations, empty object, definition drift, and side-effect counts; input/step/wait provenance and retained composite corroboration preserve genuine recovery. Tests inspected; author ran the affected DAG file (402 steps PASS), changed-file lint/format/typecheck. No score threshold is used as a merge gate. |
|
@codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 51f1607f68
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
Review disposition: the fail-closed compatibility scope is intentional for historical object rows with no input/step/wait provenance and no corroborated composite tree. Existing positive regressions retain legitimate step output, integer/array parallel output, nested composite publication, and genuine wait identity. The empty-arm branch impersonation is a separate definition-drift BUG, filed as veryfront/veryfront-issue-inbox#2600; #2472 explicitly scopes removed/step/parallel replacement. No remaining blocking finding; exact-head required CI remains the merge gate. |
Codecov Report❌ Patch coverage is
📢 Thoughts on this report? Let us know! |
|
|
@codex review Re-review exact head 51f1607 after the prior P2 disposition: the empty-arm branch replacement bug is filed as veryfront/veryfront-issue-inbox#2600, linked in the reply, and its review thread is resolved. This PR satisfies #2472 removed/step/parallel drift; source tests, format, lint, typecheck, integration, binary tests, and Sonar are now green. Independent exact-head review found no blocking finding. |
|
Codex Review: Didn't find any major issues. Breezy! Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |



Refs veryfront/veryfront-issue-inbox#2472
A loop completion callback can erase every result metadata key, so JSON-persisted output no longer satisfies the old two-key heuristic. Refuse ambiguous object publications without input/step/wait provenance; recover parallel and branch publications only when retained child states corroborate them. Context snapshots continue to preserve exact publications.
Expanded durable regression matrix covers both callbacks, all two-key combinations, all-three erasure, completely empty output, and removed/step/parallel replacement (including raw empty parallel). Corrected the prior single-key parallel fixture to agree with its retained integer child result; historical wait fixtures retain genuine wait identity. Documented the fail-closed compatibility rule.
Validation: red on origin/main (24 erasure cases failed); green affected DAG file (402 steps); changed-file format/lint/typecheck and semantic-disposition audit pass. Testing-front-door audit passes with unrelated baseline improvements reported. Local full verify skipped under owner instruction to leave full suites to CI; required remote checks and exact-head review remain mandatory.
Filed bug: veryfront/veryfront-issue-inbox#2600 tracks empty-arm branch impersonation, outside #2472 removed/step/parallel replacement.
Follow-ups (not filed): none.