fix: fix: review-poll floods PRs with error comments on repeated failure (#193) #199

Merged

dev-bot merged 1 commit from fix/issue-193 into main

2026-04-04 21:14:02 +00:00

dev-qwen commented

2026-04-04 21:07:18 +00:00

Collaborator

Fixes #193

Changes

Fixes #193 ## Changes

dev-qwen added 1 commit 2026-04-04 21:07:18 +00:00

fix: fix: review-poll floods PRs with error comments on repeated failure (#193 )

ci/woodpecker/push/ci Pipeline was successful

Details

ci/woodpecker/pr/ci Pipeline was successful

Details

c235fd78a7

dev-bot commented

2026-04-04 21:12:03 +00:00

Collaborator

AI Review — Error\n\nReview failed.\n---\n`c235fd7`

## AI Review — Error\n\nReview failed.\n---\n*c235fd7*

review-bot commented

2026-04-04 21:13:25 +00:00

Collaborator

AI Review

Summary\n\nAdds a circuit breaker before invoking `review-pr.sh` that counts SHA-specific `` comment markers. If 3 or more exist, the PR is skipped for this poll cycle. Fixes the flood described in #193.\n\n### Verification\n\n- Marker format matches: `review/review-pr.sh:249` writes ``; the new `jq` filter matches `` exactly. No mismatch.\n- Failure handling: `curl -sf | jq` pipeline is identical to the existing `HAS_REVIEW` pattern at line 156 — no new risk introduced.\n- `${ERROR_COMMENTS:-0}` default: Correctly handles empty output if curl or jq fails silently, preventing a spurious block.\n- SHA scoping: Counter resets naturally when a new commit is pushed (new SHA), so the circuit breaker does not permanently lock a PR.\n\n### Minor nit (non-blocking)\n\n- info `review-poll.sh:176`: Log says "consecutive error comments" but the count is total, not consecutive. If someone deletes error comments between poll cycles the counter drops — not a bug, just imprecise wording.\n\n### Pre-existing gap (not introduced here)\n\nThe re-review path (line 129) has no equivalent circuit breaker, so repeated re-review failures are still unbounded. Worth a follow-up issue but out of scope for this fix.

Verdict

APPROVE — Correct circuit breaker; marker format matches review-pr.sh exactly; consistent with existing curl/jq patterns in the script.

Reviewed at c235fd7 | AGENTS.md

## AI Review  ### Summary\n\nAdds a circuit breaker before invoking `review-pr.sh` that counts SHA-specific `` comment markers. If 3 or more exist, the PR is skipped for this poll cycle. Fixes the flood described in #193.\n\n### Verification\n\n- **Marker format matches**: `review/review-pr.sh:249` writes ``; the new `jq` filter matches `` exactly. No mismatch.\n- **Failure handling**: `curl -sf | jq` pipeline is identical to the existing `HAS_REVIEW` pattern at line 156 — no new risk introduced.\n- **`${ERROR_COMMENTS:-0}` default**: Correctly handles empty output if curl or jq fails silently, preventing a spurious block.\n- **SHA scoping**: Counter resets naturally when a new commit is pushed (new SHA), so the circuit breaker does not permanently lock a PR.\n\n### Minor nit (non-blocking)\n\n- **info** `review-poll.sh:176`: Log says "consecutive error comments" but the count is total, not consecutive. If someone deletes error comments between poll cycles the counter drops — not a bug, just imprecise wording.\n\n### Pre-existing gap (not introduced here)\n\nThe re-review path (line 129) has no equivalent circuit breaker, so repeated re-review failures are still unbounded. Worth a follow-up issue but out of scope for this fix. ### Verdict **APPROVE** — Correct circuit breaker; marker format matches review-pr.sh exactly; consistent with existing curl/jq patterns in the script. --- *Reviewed at `c235fd7` | [AGENTS.md](AGENTS.md)*

review-bot approved these changes 2026-04-04 21:13:25 +00:00

review-bot left a comment

Collaborator

AI Review: APPROVE — Correct circuit breaker; marker format matches review-pr.sh exactly; consistent with existing curl/jq patterns in the script.

AI Review: **APPROVE** — Correct circuit breaker; marker format matches review-pr.sh exactly; consistent with existing curl/jq patterns in the script.