Repository navigation
fix(sheets): reject append responses without update metadata - #1142
Conversation
|
🦞👀 Pull request received. I will update this pull request when review starts. ClawSweeper review completeClawSweeper finished reviewing this revision. The review result is being finalized. |
|
Codex review: needs real behavior proof before merge. Reviewed September 22, 2026, 1:21 AM ET / 05:21 UTC (Revision 2). ClawSweeper reviewWhat this changesThe PR makes Sheets append reject missing response metadata without crashing or retrying, adds regression coverage, and documents recovery guidance. Merge readiness⛔ Blocked before merge - 2 items remain This remains a useful, focused fix that main and v0.40.0 still need. No actionable correctness defect was found, and the changelog follow-up is resolved; the previously requested runtime transcript remains outstanding. Priority: P2 Review scores
Verification
How this fits togetherThe Sheets append command sends rows through Google's Sheets client, then uses response metadata for optional validation copying and result output. This change guards that post-response processing. flowchart TD
A[Rows and target range] --> B[Sheets append command]
B --> C[Google Sheets client]
C --> D[Append response]
D --> E{Update metadata present?}
E -->|No| F[Return error without retry]
E -->|Yes| G[Optional validation copy and result output]
Before merge
Agent review detailsSecurityNone. Review metrics
Technical reviewBest possible solution: Retain the focused post-response guard, unchanged successful output, and explicit guidance against blindly retrying an uncertain write. Do we have a high-confidence way to reproduce the issue? Yes, from source: HTTP 200 with {} leaves updates absent and reaches an unconditional dereference on main. The contributor reports a matching failing regression; this review did not execute it. Is this the best way to solve the issue? Yes. The guard follows the existing table-append convention, preserves the validation-copy diagnostic, and avoids retrying a potentially completed write. AGENTS.md: found and applied where relevant. Codex review notes: model internal, reasoning medium; reviewed against b2411fa43e33. LabelsLabel justifications:
EvidenceWhat I checked:
Likely related people:
Rank-up movesOptional improvements that raise the rating; they are not merge blockers.
Rating scale
Overall follows the weaker of proof and patch quality. Workflow
HistoryReview history (1 earlier review cycle)
|
Values append printed resp.Updates after a successful API call. A missing updates object panicked. Guard like table-append. Signed-off-by: Sebastien Tardif <sebtardif@ncf.ca>
90b1ca7 to
d9732a8
Compare
|
Landed as 3c4a710. The old code panicked on a successful response with missing metadata. The expanded regression passes for absent/null metadata in human and JSON modes, with empty stdout and exactly one append. |
Sheets values append could panic after a successful HTTP response with missing update metadata. Reject that incomplete response before dereferencing it, while retaining the existing validation-copy diagnostic.
Regression coverage drives the Google Sheets client and parsed append command with absent and null metadata in human and JSON modes. It verifies an ordinary non-retryable error, no success output, and exactly one append request. The automation guide explains why callers should inspect the spreadsheet before retrying an uncertain write.
Validation: the original implementation panicked on the regression fixture; the expanded cases and validation-copy regression pass on AWS Crabbox. No live spreadsheet was changed.
Thanks @SebTardif for the fix.
The final
make cigate passes on AWS Crabbox, and independent Codex review found no actionable P0–P2 findings.