fix(changelog): move pr-review attribution entry back under Unreleased - #87
Conversation
There was a problem hiding this comment.
Now I have enough information for Step 4c-sel (CLASSIFY). Let me apply the selector-mode logic:
Changed files: CHANGELOG.md (1 file)
Step 4b-i candidates (judgment rules with glob matching CHANGELOG.md):
changelog/conventional-prefix-required— owner: agent-auditor, trigger:CHANGELOG.mdgit-commit/feature-branch-no-tag— trigger:@commits(always candidate)git-commit/imperative-mood— trigger:@commits(always candidate)git-workflow/never-direct-commit-to-master— trigger:@commits(always candidate)
Step 4c-sel CLASSIFY:
| Rule | Decision | Reason |
|---|---|---|
changelog/conventional-prefix-required |
skipped | applies_when checks unreleased bullets lack fix:/feat: prefix — the new bullet is fix:, so the condition does not hold |
git-commit/feature-branch-no-tag |
skipped | applies_when checks for git tag on feature branch — not applicable to this diff |
git-commit/imperative-mood |
skipped | applies_when checks commit message subjects — not a commit diff |
git-workflow/never-direct-commit-to-master |
skipped | applies_when checks direct commits to master — not a commit operation |
Applicable set: ∅ — selector clean
Mechanical findings from pre-computed output: 0
Step 4d-sel: skipped — applicable set empty + no mechanical findings
Step 5: Consolidated Report
Must Fix (Critical): None.
Should Fix (Important): None.
Nice to Have (Optional): None.
Selector-mode traceability:
- Candidates: 4 rules matched by Step 4b-i glob filter
- Applicable: 0 rules selected by Step 4c-sel
- Skipped:
changelog/conventional-prefix-required→ bullet hasfix:prefix, condition not metgit-commit/feature-branch-no-tag→ not a tag operationgit-commit/imperative-mood→ not a commit messagegit-workflow/never-direct-commit-to-master→ not a direct commit
Notes:
- The change moves the
fix: pr-reviewentry from## v0.35.5back to## Unreleased, restoring the correct section placement after the prior PR incorrectly placed it under the versioned heading. LICENSEfile is present (Step 3a verified).- precommit skipped (selector mode) — CI covers lint+test.
Step 6: No test coverage gaps to address.
Step 7: Manual review of the diff confirms the change is correct — the fix: bullet is correctly placed under ## Unreleased and uses the proper conventional prefix. No issues found.
{
"verdict": "approve",
"summary": "Trivial structural fix that correctly restores the `fix: pr-review` changelog entry to `## Unreleased` from `## v0.35.5`. All candidates classified as non-applicable; mechanical funnel clean.",
"comments": [],
"concerns_addressed": []
}
The
fix: pr-review — every finding must now lead with a bold file referencebullet folded into the released## v0.35.5section, which does not contain the code change.What happened
The releaser cut
v0.35.5ata9682ccfor the go-testing-guide docs entry that was already sitting in## Unreleased. That happened in the window between pushing #86 and merging it, so the rename## Unreleased→## v0.35.5swept in an entry whose commit landed afterwards:Verified against the tag itself rather than the working tree:
So
v0.35.5claimed a fix it does not ship. The pre-push fold guard was run on #86 and was clean at that moment — the fold happened after, which is precisely the overlap the sequencing rule warns about.The fix
Move the entry back under
## Unreleased; the released section is restored to exactly what its tag contains, not edited into agreement with reality. After this change:The attribution fix then ships in the next release, which will genuinely contain it.
CHANGELOG only — no code touched.