prompt: retire two already-delivered PyAutoBrain bug prompts - #155
Merged
Conversation
Both prompts in draft/bug/pyautobrain/ were picked up for development and found already fixed on main. No code was written; these are ledger backfills crediting the PRs that actually delivered them. hygiene-gone-ref-counter — PyAutoBrain#205 (7ad1e43) fixed the [gone] counter 2026-08-07; PyAutoBrain#200 (5cb1c73) derived the repo sets from the body map 2026-08-05. #200 was already recorded as hygiene-coverage-drift; #205 had no record, so the ledger carried a gap. version-drift-stamp-consistency — PyAutoBrain#144 (e26ab17) reframed the check to stamp consistency 2026-07-19. No record existed. The prompt's Status still read "pending review/merge" three weeks after that branch merged, which is what made it look open. Both verified before retiring rather than assumed: the [gone] control was run on a purpose-built fixture with a genuinely gone upstream (old form 0, shipped form 1), and the version_drift fix was confirmed on main by merge-base ancestry plus reading the shipped script and the /wake_up wording. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016VnfCyAwgQdm1V3iJBtiga
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Both remaining prompts in
draft/bug/pyautobrain/were picked up for development and found already fixed on main. No code was written — this is a ledger backfill crediting the PRs that actually delivered them, and it clears two stale drafts that would otherwise send a future session to re-do finished work.hygiene-gone-ref-counter→complete/2026/08/Delivered by two PRs, neither of which the prompt knew about:
[gone]counter could never be non-zero) — PyAutoBrain#205,7ad1e43, 2026-08-07. Switched tofor-each-ref '%(upstream:track)'and added a regression test on a fixture with a genuinely gone upstream.5cb1c73, 2026-08-05. Repo sets now derived fromrepos.yamlwith arepos_sync.pycoverage check. This is exactly the resolution the prompt asked someone to decide on.#200 was already recorded as
hygiene-coverage-drift; #205 had no record, so the ledger carried a gap.version-drift-stamp-consistency→complete/2026/07/Delivered by PyAutoBrain#144,
e26ab17, 2026-07-19 — option 2 of the three the prompt offered: flag stamps out of step with the sibling consensus rather than asserting equality to the release tag, plus both documentation updates the prompt required (version_drift.shheader,/wake_upstep 5). No record existed.The prompt's
Status:still read "fix-implemented-on-branch — pending review/merge" three weeks after that branch merged. That stale status is what made it look open.Verified, not assumed
The prompt for the
[gone]counter specified its own control — "A green run on a tree with zero[gone]refs proves nothing — pick the tree first." That control was run: a throwaway repo with a real bare remote, a pushed-then-deleted branch and afetch --prune, so the checkout genuinely carried a[gone]ref. On that same tree the old form counts 0 and the shipped form counts 1. Also confirmed no repo-name array literal survives inhygiene.sh, including inrun_tidyandenumerate_condemn_candidates, the two the prompt named as sharing the limitation.For version_drift:
merge-base --is-ancestoragainst main, plus reading the shipped script and the/wake_upwording to confirm both doc requirements landed.A trap worth keeping, recorded in the first record
Establishing "already fixed" required deepening the clone. Cloud-session checkouts are shallow, and a shallow PyAutoBrain makes
git diff origin/main...<branch>fail withno merge baseand makesgit log origin/main..<branch>list commits that are on main. Both look like evidence of divergence and are artifacts of the graft boundary. This is the second time in this session that shallowness produced a confidently wrong reading of history.lifecycle.py check: OK.lifecycle.py index --check: OK (937 records).Note: the branch was reset from
mainand force-pushed with lease — it previously held only the two commits squash-merged as8b33239in #154.🤖 Generated with Claude Code
https://claude.ai/code/session_016VnfCyAwgQdm1V3iJBtiga
Generated by Claude Code