Repository navigation
ci: Read closing Issues from the PR body - #54
Conversation
The reviewer took a pull request's closing Issues from GraphQL closingIssuesReferences. GitHub stopped deriving those from closing keywords, so PRs whose body says "Closes #N" were blocked on pr-linkage. The closing Issues are now the ones the body names with a closing keyword, read in one GraphQL request; Development links are no longer consulted. Closes #53 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
There was a problem hiding this comment.
❌ OpenAI PR Review: FAIL
Conclusion: Not ready. 2 readiness blockers and 1 actionable finding must be resolved before merge.
Review checks
| Check | Result |
|---|---|
| PR format | ❌ FAIL |
| Issue design | ✅ PASS |
| Code & plan conformance | ❌ FAIL |
Scope: b64aaeb86e · 8dac71eefa..b64aaeb86e · full · 1 diff chunk
Usage: 1m 9s · 445,688 tokens · 86.3% cache hit · credits unavailable
Blockers
pr-linkagePull request must natively close at least one same-repository Issue.code-review.github/scripts/pr-readiness/closing-issues.mjs:70: Ignore closing keywords in indented Markdown code blocks
Findings
1 inline finding published.
Summary
The complete diff review found one actionable closing-Issue parsing defect. No linked Issue context or discussion was supplied, and the chunk review reported no plan-conformance blockers.
Review metadata
- Commit:
b64aaeb86e - Range:
8dac71eefa..b64aaeb86e - Mode:
full - Diff chunks:
1 - Model:
gpt-6-sol - Reasoning effort:
low - Session:
repo-1309321116-pr-54-v2 - Generation:
48787219649b9184ccab9e27a8e923d9da370954c5a8cbcfef3425e477bf29ca - Evidence:
8f8c6a2a4cd419fdc9eeb822b6f8da8a022a0c9d108099c4c75a0da919e77350
Totals
- Input:
442,043 - Cached input:
381,692 - Cache write:
59,263 - Output:
3,645 - Reasoning:
770 - Total:
445,688 - Estimated credits: unavailable for this model
Token and cache details
| Stage | Mode | Target | Time | Input | Cached | Hit | Output | Total | Credits |
|---|---|---|---|---|---|---|---|---|---|
pr |
deterministic |
format rules |
0s | 0 | 0 | 0.0% | 0 | 0 | N/A |
pr |
full |
pr |
10s | 43,976 | 28,215 | 64.2% | 535 | 44,511 | N/A |
code |
full |
chunk 1/1 |
49s | 312,231 | 270,435 | 86.6% | 2,559 | 314,790 | N/A |
code |
full |
aggregate |
10s | 85,836 | 83,042 | 96.7% | 551 | 86,387 | N/A |
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
There was a problem hiding this comment.
❌ OpenAI PR Review: FAIL
Conclusion: Not ready. 2 readiness blockers and 1 actionable finding must be resolved before merge.
Review checks
| Check | Result |
|---|---|
| PR format | ❌ FAIL |
| Issue design | ✅ PASS |
| Code & plan conformance | ❌ FAIL |
Scope: acb90efe4b · b64aaeb86e..acb90efe4b · incremental · 1 diff chunk
Usage: 40s · 288,565 tokens · 66.2% cache hit · credits unavailable
Blockers
pr-linkagePull request must natively close at least one same-repository Issue.code-review.github/scripts/pr-readiness/closing-issues.mjs:29: Exclude indented code nested under list items
Findings
1 inline finding published.
Summary
The incremental patch fixes the previously reported top-level indented-code case. One actionable case remains: indented code nested under a list item can still be treated as closing-Issue linkage.
Review metadata
- Commit:
acb90efe4b - Range:
b64aaeb86e..acb90efe4b - Mode:
incremental - Diff chunks:
1 - Model:
gpt-6-sol - Reasoning effort:
low - Session:
repo-1309321116-pr-54-v2 - Generation:
0e2ac92ba5975e9b62bf713407ea29d2f70004004bbc3fe0d788014c953869cb - Evidence:
67f9f9fe4f0db83e57408515a9326c77a146479b9470841a62e28e74d3ecc7b8
Totals
- Input:
286,352 - Cached input:
189,611 - Cache write:
96,333 - Output:
2,213 - Reasoning:
1,026 - Total:
288,565 - Estimated credits: unavailable for this model
Token and cache details
| Stage | Mode | Target | Time | Input | Cached | Hit | Output | Total | Credits |
|---|---|---|---|---|---|---|---|---|---|
pr |
deterministic |
format rules |
0s | 0 | 0 | 0.0% | 0 | 0 | N/A |
pr |
incremental |
pr |
12s | 89,574 | 44,351 | 49.5% | 536 | 90,110 | N/A |
code |
incremental |
chunk 1/1 |
20s | 95,308 | 46,634 | 48.9% | 1,210 | 96,518 | N/A |
code |
incremental |
aggregate |
8s | 101,470 | 98,626 | 97.2% | 467 | 101,937 | N/A |
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
There was a problem hiding this comment.
❌ OpenAI PR Review: FAIL
Conclusion: Not ready. 2 readiness blockers and 1 actionable finding must be resolved before merge.
Review checks
| Check | Result |
|---|---|
| PR format | ❌ FAIL |
| Issue design | ✅ PASS |
| Code & plan conformance | ❌ FAIL |
Scope: 2c7eea2ecd · acb90efe4b..2c7eea2ecd · incremental · 1 diff chunk
Usage: 42s · 339,487 tokens · 82.4% cache hit · credits unavailable
Blockers
pr-linkagePull request must natively close at least one same-repository Issue.code-review.github/scripts/pr-readiness/closing-issues.mjs:52: Calculate list padding from the marker’s actual column
Findings
1 inline finding published.
Summary
The patch fixes the previous finding for space-padded list items. One actionable case remains for tab-padded list items.
Review metadata
- Commit:
2c7eea2ecd - Range:
acb90efe4b..2c7eea2ecd - Mode:
incremental - Diff chunks:
1 - Model:
gpt-6-sol - Reasoning effort:
low - Session:
repo-1309321116-pr-54-v2 - Generation:
aff86b530226f1ab73ba51fce330a7fb43dc1f71e9e481bfb919b17f328dfea4 - Evidence:
3933388c189157de732eec91769b6226a7c0a7700f6343760739db2401b2710e
Totals
- Input:
337,214 - Cached input:
277,745 - Cache write:
59,061 - Output:
2,273 - Reasoning:
1,119 - Total:
339,487 - Estimated credits: unavailable for this model
Token and cache details
| Stage | Mode | Target | Time | Input | Cached | Hit | Output | Total | Credits |
|---|---|---|---|---|---|---|---|---|---|
pr |
deterministic |
format rules |
0s | 0 | 0 | 0.0% | 0 | 0 | N/A |
pr |
reused |
cached evidence |
0s | 0 | 0 | 0.0% | 0 | 0 | N/A |
code |
incremental |
chunk 1/1 |
33s | 220,156 | 163,132 | 74.1% | 1,811 | 221,967 | N/A |
code |
incremental |
aggregate |
9s | 117,058 | 114,613 | 97.9% | 462 | 117,520 | N/A |
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
There was a problem hiding this comment.
❌ OpenAI PR Review: FAIL
Conclusion: Not ready. 1 readiness blocker must be resolved before merge.
Review checks
| Check | Result |
|---|---|
| PR format | ❌ FAIL |
| Issue design | ✅ PASS |
| Code & plan conformance | ✅ PASS |
Scope: 824e0e3c5f · 2c7eea2ecd..824e0e3c5f · incremental · 1 diff chunk
Usage: 16s · 250,091 tokens · 94.8% cache hit · credits unavailable
Blockers
pr-linkagePull request must natively close at least one same-repository Issue.
Summary
The incremental patch fixes the previous tab-padding finding. No actionable code findings or plan-conformance blockers remain in the aggregated review.
Review metadata
- Commit:
824e0e3c5f - Range:
2c7eea2ecd..824e0e3c5f - Mode:
incremental - Diff chunks:
1 - Model:
gpt-6-sol - Reasoning effort:
low - Session:
repo-1309321116-pr-54-v2 - Generation:
576c1fdc4b687c6358f27f1a2eca0544588864a7a5947850c9a221b71c817f61 - Evidence:
60909ea325d594b8bbad412f92ad72c9ac2792ecb76092cc6ccc5d095a19f86f
Totals
- Input:
249,308 - Cached input:
236,427 - Cache write:
12,609 - Output:
783 - Reasoning:
72 - Total:
250,091 - Estimated credits: unavailable for this model
Token and cache details
| Stage | Mode | Target | Time | Input | Cached | Hit | Output | Total | Credits |
|---|---|---|---|---|---|---|---|---|---|
pr |
deterministic |
format rules |
0s | 0 | 0 | 0.0% | 0 | 0 | N/A |
pr |
reused |
cached evidence |
0s | 0 | 0 | 0.0% | 0 | 0 | N/A |
code |
incremental |
chunk 1/1 |
9s | 122,516 | 111,635 | 91.1% | 455 | 122,971 | N/A |
code |
incremental |
aggregate |
7s | 126,792 | 124,792 | 98.4% | 328 | 127,120 | N/A |
1 participant
Closes #53
Summary
Closes #N,fixes owner/repo#N, an Issue URL, …) instead of GraphQLclosingIssuesReferences. GitHub stopped deriving those references around 2026-09-30 07:00 UTC, so PRs that sayCloses #Nwere blocked onpr-linkage(h2vivi/firmwares #1477 and #1479, GizClaw/gizos #617).pr-readiness/closing-issues.mjsparses the body, reads up to 100 referenced Issues in one aliased GraphQL request, and normalizes them. The reusable workflow's context step andverify.mjsboth use it, replacing their duplicated Issue normalization.too-many-closing-issues.analyzePullRequestand the snapshot shape are unchanged. The body was already hashed, so changing a closing reference still invalidates the snapshot..github/scripts/review-request, whosestripQuotedTextthe parser reuses.closedByPullRequestsReferences; a PR GitHub has not linked is refreshed by its next event or an@codex reviewcomment.This PR itself cannot pass the currently deployed
pr-linkagerule for the reason it fixes: GitHub does not create the native link forCloses #53.Validation
node .github/scripts/pr-review/test.mjsnode .github/scripts/review-request/test.mjsnode .github/scripts/pr-readiness/test.mjsnode .github/scripts/issue-review/test.mjsactionlint -ignore 'property "workflow_(repository|sha)" is not defined' .github/workflows/*.ymlgit diff --checkcloses #1479(a pull request),closes #999999, andcloses nobody-xyz/none#1are ignored.review-request/common.mjsimport in the resolve job, and the step is executed bypr-review/test.mjsagainst a fake client.🤖 Generated with Claude Code