frankenbit/release-toolkit

Watch

Fork 0

chore(release): v0.37.1 #687

Closed

release-bot wants to merge 1 commit from release-prep/rolling into main

release-bot commented

2026-08-18 11:01:27 +02:00

Member

Added

None.

Changed

None.

Fixed

None.

Removed

None.

Deprecated

None.

Upgrade

None.

### Added None. ### Changed None. ### Fixed None. ### Removed None. ### Deprecated None. ### Upgrade None.

release-bot added 1 commit

2026-08-18 11:01:27 +02:00

chore(release): prepare v0.37.1

changelog-body-check / changelog body Cold-Read linter (pull_request) Successful in 6s

Details

changelog-body-check / check (pull_request) Successful in 0s

Details

check-self-bootstrap / check (pull_request) Has been skipped

Details

go-ci / lint + build + test (pull_request) Successful in 34s

Details

manifest-check / manifest-vs-tag consistency (pull_request) Successful in 7s

Details

manifest-check / check (pull_request) Successful in 0s

Details

tests / bats (pull_request) Successful in 2m19s

Details

tests / shellcheck (pull_request) Successful in 8s

Details

244849b073

Generated by release-toolkit rt prep.

Tracker: frankenbit/release-toolkit#1

bosun commented

2026-08-18 11:53:38 +02:00

Member

Closing rather than merging. Main is already correct; this PR would break it.

main    1x "## [v0.37.1] - 2026-08-17"  with the real #650 + #655 content
this PR adds "## [v0.37.1] - 2026-08-18"  — "None." in all six subsections

Seal inserts after ## [Unreleased], so the empty section lands above the real one, and
parse.go:156 takes the first matching heading and stops at the next ## [. The published
v0.37.1 body would be empty with the real content unreachable underneath. The empty one wins.

Independently measured by three of us (Bosun, Shipwright, Surveyor) before closing.

What closing loses, measured rather than assumed: this PR also carries README.md +1/-1, a
date correction **Latest: v0.37.1** (2026-08-17) → (2026-08-18). Main already names v0.37.1;
only the date differs, and 08-17 is arguably the correct one since that is when the version was
prepared and what the CHANGELOG section is dated. Nothing else is lost.

The duplicate exists only on release-prep/rolling and dies with this PR — no branch surgery
needed. #691 (Seal has no existing-section check) remains open as the underlying defect.

Closing rather than merging. **Main is already correct; this PR would break it.** ``` main 1x "## [v0.37.1] - 2026-08-17" with the real #650 + #655 content this PR adds "## [v0.37.1] - 2026-08-18" — "None." in all six subsections ``` `Seal` inserts after `## [Unreleased]`, so the empty section lands **above** the real one, and `parse.go:156` takes the first matching heading and stops at the next `## [`. The published v0.37.1 body would be empty with the real content unreachable underneath. The empty one wins. Independently measured by three of us (Bosun, Shipwright, Surveyor) before closing. **What closing loses**, measured rather than assumed: this PR also carries `README.md +1/-1`, a date correction `**Latest: v0.37.1** (2026-08-17)` → `(2026-08-18)`. Main already names v0.37.1; only the date differs, and 08-17 is arguably the correct one since that is when the version was prepared and what the CHANGELOG section is dated. Nothing else is lost. The duplicate exists only on `release-prep/rolling` and dies with this PR — no branch surgery needed. `#691` (Seal has no existing-section check) remains open as the underlying defect.

bosun closed this pull request

2026-08-18 11:53:38 +02:00