fix(review): clear stale AR change requests after clean re-review #52

New issue

Closed

opened 2026-05-03 09:07:20 -07:00 by jwilger · 0 comments

jwilger commented

2026-05-03 09:07:20 -07:00

Owner

Problem

PR #48 could not be merged even though the latest Auto Review run reported no blocking findings and both required status checks were green.

The merge blocker was stale Forgejo review state:

main branch protection has block_on_rejected_reviews=true and block_on_official_review_requests=true.
PR #48 still listed auto-review as a requested reviewer.
An earlier official Auto Review record remained undismissed:
- review id: 751
- state: REQUEST_CHANGES
- official: true
- stale: true
- dismissed: false
The latest Auto Review record was clean, but did not clear the blocker:
- review id: 753
- state: COMMENT
- official: false

As a result, Forgejo continued to block merge despite the latest AR result and commit status being successful.

Expected behavior

When Auto Review re-runs on a newer head SHA and finds no blocking issues after a previous AR REQUEST_CHANGES, it should clear its prior blocking review state. Acceptable behavior may be one of:

submit an official APPROVED review for the latest head;
dismiss or otherwise supersede the stale official REQUEST_CHANGES review if Forgejo permits that via API;
remove auto-review from requested reviewers once the clean review is posted;
or another Forgejo-compatible flow that satisfies protected-branch review gates.

Observed evidence from PR #48

Head SHA: 869e3677e9f89ea35e937609a1f76ec4831630ba
PR mergeable: true
Aggregate commit status: success
Required status contexts were green:
- CI / Nix flake check (pull_request)
- auto_review
Branch protection required review gates still blocked merge because stale official AR REQUEST_CHANGES remained active.

Impact

Users can be unable to merge a PR after addressing all Auto Review findings unless they manually dismiss stale AR review state or clear requested reviewers. This makes clean re-reviews misleading and undermines the CI-triggered review workflow.

## Problem PR #48 could not be merged even though the latest Auto Review run reported no blocking findings and both required status checks were green. The merge blocker was stale Forgejo review state: - `main` branch protection has `block_on_rejected_reviews=true` and `block_on_official_review_requests=true`. - PR #48 still listed `auto-review` as a requested reviewer. - An earlier official Auto Review record remained undismissed: - review id: `751` - state: `REQUEST_CHANGES` - `official: true` - `stale: true` - `dismissed: false` - The latest Auto Review record was clean, but did not clear the blocker: - review id: `753` - state: `COMMENT` - `official: false` As a result, Forgejo continued to block merge despite the latest AR result and commit status being successful. ## Expected behavior When Auto Review re-runs on a newer head SHA and finds no blocking issues after a previous AR `REQUEST_CHANGES`, it should clear its prior blocking review state. Acceptable behavior may be one of: - submit an official `APPROVED` review for the latest head; - dismiss or otherwise supersede the stale official `REQUEST_CHANGES` review if Forgejo permits that via API; - remove `auto-review` from requested reviewers once the clean review is posted; - or another Forgejo-compatible flow that satisfies protected-branch review gates. ## Observed evidence from PR #48 - Head SHA: `869e3677e9f89ea35e937609a1f76ec4831630ba` - PR mergeable: `true` - Aggregate commit status: `success` - Required status contexts were green: - `CI / Nix flake check (pull_request)` - `auto_review` - Branch protection required review gates still blocked merge because stale official AR `REQUEST_CHANGES` remained active. ## Impact Users can be unable to merge a PR after addressing all Auto Review findings unless they manually dismiss stale AR review state or clear requested reviewers. This makes clean re-reviews misleading and undermines the CI-triggered review workflow.

jwilger referenced this issue from a pull request that will close it,

2026-05-03 09:19:03 -07:00

fix(review): approve clean re-reviews #53

jwilger closed this issue

2026-05-03 09:27:38 -07:00

jwilger referenced this issue from a commit

2026-05-03 09:27:40 -07:00

fix(review): approve clean re-reviews (#53)