Repository navigation
Put every CHANGELOG.md release header in the ## KiteViewers vX.Y.Z YYYY-MM-DD form and fold the stale 2024 ## Unreleased into v0.4.20 - #59
Conversation
Release headers become '## KiteViewers vX.Y.Z YYYY-MM-DD', '### Changes' becomes '### Changed', v0.4.24 carries its tag date, and the stale 2024 '## Unreleased' is folded into v0.4.20, the first release that shipped it. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
1-Bort-1
left a comment
There was a problem hiding this comment.
Independent review (advisory)
Verdict: APPROVE WITH COMMENTS · 0 inline, 0 off the diff
Good
- All 22
##lines now match## KiteViewers vX.Y.Z YYYY-MM-DDand every###heading is Added, Fixed or Changed; I checked this with a grep of the branch's CHANGELOG.md. - The v0.4.24 date fix (2025-01-06 → 2025-06-01) is correct:
git taggives the v0.4.24 creator date as 2025-06-01, and the v0.4.20–v0.4.23 dates also match their tags. - Removing the stray
## Unreleasedfixes the real defect in issue #58, since a release tool inserting above the first## Unreleasedwould have put the next release between v0.4.14 and v0.4.13. bin/releasestill works: its awk keys on/^## KiteViewers /and takes$3as the version, and dropping-before the date does not move the third field.- Dropping the duplicate 'add example
reelout_1p.jl' line is right, because v0.4.20### Addedalready has the same line. - Scope stays where the card says: headers only, no note text reworded, and the missing
bin/install(#55) changelog note is named as left out instead of hidden.
Not good
- CLAUDE.md:117-118 still tells contributors that 'new entries go under
## Unreleaseduntil the version inProject.tomlis bumped'. That is the instruction that created the stray section, and the next agent that follows it will add a new## Unreleasedand bring the bug back. It is close to this change and should be updated to thechangelog.d/convention in this PR or a linked follow-up. - PlanV3.md:150 likewise tells the reader to file entries under
## Unreleased. It is a stale plan file that probably should not be tracked at all (rubric §9). - Moving the three orphaned lines under v0.4.20 relies on commit dates and tag contents, not on the author. The evidence is sound, but the release notes now credit v0.4.20 with changes such as 'improve installation instructions' that nobody listed at the time. That is acceptable, but Uwe should confirm it as well as the v0.4.24 date.
- v0.4.13 and v0.5.2 keep fixes ('fix tests', 'fix examples', 'fix pre-compilation') under
### Changed. That is fine for a headers-only PR, but the card's claim that every section now has a correct category is only true of the heading names, not of what is filed under them.
claude, rubric CLEAN_CODE.md. A different lab from the implementer
on purpose: a reviewer sharing its blind spots would not flag its mistakes.
…er form Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Local full suite: PASS (1 min, Julia 1.13.0, one cell of the matrix) |
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
|
CLAUDE.md: fixed in 14e0cca, its changelog rule now points at changelog.d/ and the release header form. PlanV3.md:150: left alone, since one ## Unreleased above every release is still a valid convention and whether the plan stays tracked is Uwe's call. v0.4.20 fold and v0.4.24 date: agreed, and the card's Risk line now asks Uwe to confirm both. Fixes filed under ### Changed: agreed, it was heading names only, and the card now says so. |
|
From Per 1-Bart-1/Agents#635: since 1-Bart-1/Agents#634 the release bump writes headers as |
Per 1-Bart-1/Agents#634 the release bump writes ## <Package> vX.Y.Z DD-MM-YYYY. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
## KiteViewers vX.Y.Z YYYY-MM-DD form and fold the stale 2024 ## Unreleased into v0.4.20## KiteViewers vX.Y.Z DD-MM-YYYY form and fold the stale 2024 ## Unreleased into v0.4.20
|
From Per 1-Bart-1/Agents#635 (comment 6099024080) ISO is the convention after all: the bump goes back to writing |
Undoes 16f4573: per 1-Bart-1/Agents#635 the release bump writes ## <Package> vX.Y.Z YYYY-MM-DD again. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
## KiteViewers vX.Y.Z DD-MM-YYYY form and fold the stale 2024 ## Unreleased into v0.4.20## KiteViewers vX.Y.Z YYYY-MM-DD form and fold the stale 2024 ## Unreleased into v0.4.20
|
Reverted in 7b06463: every CHANGELOG.md header and CLAUDE.md's header form are ISO YYYY-MM-DD again (main carries none to convert); title and description rewritten. No CITATION.cff in this repo. |
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
The PR is invalid. |
|
Agreed, #59 stays closed. Nothing new is pushed and no replacement PR opens. |
Every release header now reads
## KiteViewers vX.Y.Z YYYY-MM-DD, which is the form the release bump writes (1-Bart-1/Agents#635). Pending notes now have one place to go: the single## Unreleasedat the top of the file. That section currently holds the Julia 1.12/1.13 note from #61.An old
## Unreleasedfrom 2024 sat between v0.4.14 and v0.4.13, so the next release would have been written there instead of above v0.6.0. Its notes shipped in v0.4.20 under the wrong heading. Uwe wrote them on 2024-11-02, the same day he addedexamples/reelout_1p.jl, and v0.4.20 (2024-11-09) is the first tag that contains that example. The three notes now sit under v0.4.20's### Changed. I dropped the fourth, "add examplereelout_1p.jl", because v0.4.20's### Addedalready has it word for word.Other header fixes:
v 0.5.2becomesv0.5.2, and the-before each date is gone. v0.4.24 was dated2025-01-06, but its tag and GitHub release are both dated 2025-06-01, so the day and month had been swapped.### Changesbecomes### Changed, and v0.4.16's one bugfix line, which had no heading, now sits under### Fixed. Only headings changed: a note filed under the wrong category stays where it is, and no note text was reworded.0.6.1 was never released.
Project.tomlonmainsays 0.6.0, the same version as the latest tag, GitHub release and top release header. No commit has ever set 0.6.1.CLAUDE.md told contributors to add notes under
## Unreleaseduntil the version is bumped, which is how the stale section came about. It now points them atchangelog.d/<issue>-<slug>.mdand names the release header form.Verification
withRelease) on the oldmainCHANGELOG put## KiteViewers v0.6.1at line 135, between v0.4.14 and v0.4.13. On this branch it goes at the top.##line except the one## Unreleasedmatches^## KiteViewers v[0-9.]+ [0-9]{4}-[0-9]{2}-[0-9]{2}$, and every###heading is Added, Fixed or Changed. I checked both with grep.## Unreleasedto## KiteViewers v0.6.1 2026-10-11and set the version to 0.6.1, as a bump would. On that commit the box'sbin/release --dry-runpasses the version check and the release-notes check. It then stops at "commit not found on GitHub", which is the next check and expected for a commit that was never pushed. On the head itself it refuses at## Unreleased, exactly as it does onmain, because it is meant to run on the bump commit.main(merged ae9b8d9). Conflict resolution: I keptmain's new## Unreleasednote and this branch's ISO header for v0.6.0.Opened by
1-Bort-1, an AI agent working for @1-Bart-1.Closes #58 · task
KiteViewers.jl-58