Skip to content

The Queen's accept is the merge: dispatch.t27 says so, and its harness measures it - #5452

Open
gHashTag wants to merge 1 commit into
masterfrom
feat/dispatch-publish-rule
Open

gHashTag wants to merge 1 commit into
masterfrom
feat/dispatch-publish-rule

Conversation

@gHashTag

@gHashTag gHashTag commented Oct 1, 2026

Copy link
Copy Markdown
Owner

Closes #5451

specs/queen/dispatch.t27 stopped at review_state accept and said nothing about what reaches master. The owner's rule of 2026-10-01 (her accept IS the merge, and nothing else is) lived only in tools/queen/publish.py (#5422), so the spec's own harness never measured it.

What changes

  • PUBLISH_RULE, and finding f65 [publish]: of 5 verdicts × 6 head relations × 2 parent-commit kinds, the publisher arms exactly the accepts at the judged head, or under its own docs/now-only commit over that head.
  • tools/trinity_queen_dispatch.py gets a sixth evidence kind, publish. It loads the publisher's own accepted_at from this checkout and drives all 60 shapes on every check. There is no record to go stale.
  • Negative controls in --self-check: three planted publishers must each fail.
    • one arms any accept, whatever head was judged;
    • one arms any verdict at the judged head;
    • one never asks what the commit over the judged head touched.
  • Replay re-run: 2488/2488. The seal is re-saved for this spec only; I measured it with the restored seal checker, and it holds.

Evidence (local, t27c built from master's bootstrap/)

python3 tools/trinity_queen_dispatch.py run          -> 2488 passed, 0 failed
python3 tools/trinity_queen_dispatch.py check        -> check: ok
python3 tools/trinity_queen_dispatch.py --self-check -> ok (incl. the 4 new lines)

Why now: an audit of every merged queen-* PR against her verdict on its issue.

  • 515 of 554 merged bee PRs carry no accept from her: none 399, sendBack 85, escalate 19, wait 12. One caveat: the board holds her latest verdict, not the one she gave at merge time.
  • 131 seals are stale: the spec changed after sealing #5183 merged with her verdict escalate and replaced tools/check_seal_coverage.py (688 lines) with one comment line.

🤖 Generated with Claude Code

…s measures it (#5451)

specs/queen/dispatch.t27 stopped at review_state accept and said nothing about
what reaches master, so the owner's rule of 2026-10-01 lived only in
tools/queen/publish.py, unmeasured by the spec's own harness.

- PUBLISH_RULE, and f65 [publish]: of 5 verdicts x 6 head relations x 2
  parent-commit kinds, the publisher arms exactly the accepts at the judged
  head (or under its own docs/now-only commit over it).
- trinity_queen_dispatch.py: a sixth evidence kind, `publish`, which loads the
  publisher's own accepted_at and drives all 60 shapes on every `check`. Three
  planted publishers must each fail --self-check: one that arms any accept, one
  that arms any verdict at the judged head, one that never asks what the
  commit over the judged head touched.
- Replay re-run (2488/2488); the seal is re-saved for this spec only.

Why now: of 554 merged bee PRs, 515 carry no accept from her. #5183 merged
with her verdict `escalate` and replaced tools/check_seal_coverage.py with one
comment line.

Closes #5451

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@github-actions

github-actions Bot commented Oct 1, 2026

Copy link
Copy Markdown
Contributor

PR Dashboard

Generated at: 2026-10-01 16:30:26 UTC

Summary

Status Count
Total Open PRs 50
PRs with Failing Checks 46
PRs with All Checks Green 4
READY 3
FAILING 46
PENDING 0
NO CHECKS YET 0

These columns do not partition: 3 + 46 + 0 + 0 = 49, and there are 50 open PRs. A PR is being counted twice or not at all.

Seal Status

  • ⚠️ STALE -- sha256(compiler.rs)=b7d5cc5c4cf1 != manifest seal=87e5cbd3ad94.
    The committed NMSE numbers were certified against an older compiler.rs.
    Run scripts/reseal-check.sh locally for the two-step reseal command (advisory; not a merge gate).

@github-actions

github-actions Bot commented Oct 1, 2026

Copy link
Copy Markdown
Contributor

📓 NotebookLM Notebook linked to this PR

This notebook contains session context, decisions, and artifacts for this work.

This was referenced Oct 1, 2026
This was referenced Oct 6, 2026

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

dispatch.t27 says nothing about what reaches master; the accept-is-the-merge rule lives only in code

2 participants