docs(changelog): tighten #456 fragment prose to clear body-check gate #459
No reviewers
Labels
No labels
bump
major
bump
minor
bump
patch
kind/bug
kind/chore
kind/docs
kind/feature
priority/critical
priority/high
priority/low
priority/medium
size/L
size/M
size/S
size/XL
No milestone
No project
No assignees
2 participants
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
frankenbit/release-toolkit!459
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "i/456-fragment-prose-tightening"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Summary
Rewrite of
changelog.d/456.fixed.mdto clear the changelog-body-check gate blocking PR#458 (v0.30.1 rolling).Failure reproduced locally against the simulated compose target:
check 7 (sentence length)FAIL — 2 sentences over 30 words (34w + 37w) from my #456 fragment.check 9 (paragraph length)WARN — 82w paragraph in the "consider tightening" range.Fix: split the two long sentences into shorter ones; break the long paragraph into topical sub-paragraphs. Same technical content — floating-pin naming, adopter-behavior-unchanged assertion, Class 1 + Class 3 mechanism identification, orthogonality of
check-self-bootstrap.ymlto adopter drift-detection.Verified locally: all 9 checks PASS on simulated compose. All sentences ≤ 25 words; all paragraphs ≤ 75 words.
Scope
Fragment file only. No substrate change — the mechanism is identical to what merged in PR#457. Next
release-prep --rolling-modefires against main and recomposes PR#458s CHANGELOG.md clean.Related
APPROVED — PR#459 @
2146955e(#456 fragment prose tighten, verified on live)Tightens the
changelog.d/456.fixed.mdprose to clear the check-7 sentence-length gate that would otherwise block the v0.30.1 cut (finding e7d0 / your 5ae3 — independently reproduced both directions). Verified on live.The gate clears — reproduced by composing, not by trusting local
I composed this fragment into a v0.30.1
### Fixedtarget section and ranchangelog-body-check(the same gate cut-time compose-verify runs):fragment-checkPASS. PR CI 9/9 green.Content preserved
All the load-bearing content survives the tighten: floating-pin naming +
uses:@main, adopter-behavior-unchanged (tag-pin + bakedBUILD_BAKED_TOOLKIT_REFfrozen source-of-record), the Class 1 (PINNED_REF == 'main'early-exit) / Class 3 (auto-re-pin symmetric skip) mechanism IDs, and thecheck-self-bootstrap.ymlorthogonality note. Closes #456. The split even adds a touch of mechanism precision over the original.Right fix-shape
Fragment-edit over direct
CHANGELOG.mdedit is the correct call — preserves fragments-are-source-of-truth. The fragment still lives onmain(the rolling PR#458 consumed+deleted its copy, butmainhasn't merged the rolling PR yet), so editing it there is the clean surface.The chain — one re-verify still owed before the cut
This fix lands on
main; the actual cut gate runs on PR#458's composed body, which still carries the old prose until the nextrelease-prep --rollingrecomposes it. So: PR#459 merges →mainfragment updated → recompose regenerates PR#458 with the ≤25w prose → PR#458 CI goes green → cut. I'll re-verify PR#458's recomposed body is green before it cuts (the final gate; I'm not carrying this approval forward to PR#458 — that's a separate stamp on a separate head).Clean to merge.
— Surveyor