fix: repair dirty automerge prs#23
Conversation
|
@clawsweeper automerge |
|
🦞🦞 I added Draft PRs stay fix-only until GitHub marks them ready for review. A maintainer can pause this with |
|
Codex review: passed for ClawSweeper automerge. What this changes: The PR routes trusted automerge pass markers with dirty, behind, or conflicting merge state into the repair dispatch path, adds mergeable state to PR target snapshots, refactors the passed-repair dispatch helper, and adds regression tests. Automerge follow-up: No repair-lane job is needed from this review; the PR is already opted into automerge and has no actionable review finding for an automated repair worker. Security review: Security review cleared: No concrete security or supply-chain issue found; the diff changes router logic and tests only, with no workflow, dependency, permission, secret, or package-resolution changes. Review detailsBest possible solution: Land the narrow router change after the normal automerge gates pass, so trusted clean review markers trigger a repair pass when GitHub reports stale or conflicting merge state instead of ending as a blocked merge attempt. Do we have a high-confidence way to reproduce the issue? Yes. A high-confidence reproduction path is a trusted ClawSweeper automerge pass marker on an opted-in PR whose target has DIRTY or BEHIND merge_state_status, or CONFLICTING mergeable state; current main plans merge, while the PR's helper tests cover the repair-trigger cases. Is this the best way to solve the issue? Yes. Routing the decision in classifyAutomergePass is the narrowest maintainable fix because the target snapshot already carries GitHub merge state, exact-head validation still runs first, and autoRepairAlreadyPlanned continues to cap repeated repair dispatches. What I checked:
Likely related people:
Codex review notes: model gpt-5.5, reasoning high; reviewed against 6df47f88f85d. |
|
🦞🦞 Source: The automerge loop is complete. |
Summary
Validation