Skip to content

fix(queen): re-pin queen/dispatch to fix/queen-worker-provider-and-prompt-size - #6377

Merged
gHashTag merged 1 commit into
masterfrom
claude/queen-dispatch-repin
Oct 5, 2026
Merged

fix(queen): re-pin queen/dispatch to fix/queen-worker-provider-and-prompt-size#6377
gHashTag merged 1 commit into
masterfrom
claude/queen-dispatch-repin

Conversation

@gHashTag

@gHashTag gHashTag commented Oct 5, 2026

Copy link
Copy Markdown
Owner

specs/queen/dispatch.t27 and tools/trinity_queen_dispatch.py pinned gHashTag/BrowserOS feat/queen-supervisor @ c25e1b02, but that branch has been deleted. This PR re-pins them to fix/queen-worker-provider-and-prompt-size @ 879799a8bfb286e2ba8dbe840f0bcb81792d19ef, the branch production deploys, and re-measures.

Closes #6368

Re-measured at the new pin (before at c25e1b02 -> after at 879799a8)

  • ts (bun 1.3.12): 2112 claim rows, board columns, constants and round gate are identical.
  • queend (rebuilt, Swift 6.3.3): choose over 120 orders, capacity, review over 120 and retry over 85 sequences are identical. queen-core source is byte-identical between the pins.
  • pg: contention (32 contenders over 20 rounds), writers, stale criteria release, CI take-back and ceiling are identical.
    • Expiry fences went from 10/11/12/13 to 1/2/3/4. Only the absolute values changed, because the database was fresh; each step is still +1.
    • The boardQuery hash changed because the board SELECT now reads judged_head. The round and ensure query hashes are unchanged.
    • server_version went from 16.15 to 17.7 (see deviations).
  • live: on 2026-10-01 the board reported working, 7 of 20 workers and 930 public cards. At the 2026-10-05 05:59Z snapshot it reported working, 18 of 70 workers and 1218 cards (spec f64 updated).
  • Replay: run passed 2488 of 2488 cases. check reported ok, --self-check caught every planted fault, t27c check reported Typecheck OK, and test-report gave 6/6 tests. The seal is refreshed.

Not re-run, or deviating

  • lotus was not re-run. It is pinned to the trinity revision, which this PR does not move, and Zig 0.15.2 is not on the host (it has 0.16.0). The record is kept as is and marked in the spec as measured in the c25e1b02 run.
  • pg ran on PostgreSQL 17.7, not 16. PG 16 and Docker were unavailable. The record keeps the server version it saw, and the spec and tool docstring say so.

Behavior that changed between the pins (read from the source; the replay does not measure it)

  • Runner lanes: a runner's task holds its issue in state queued and does not take a container worker slot. CLAIM_RULE and AGENTS_RULE are updated.
  • Runner-only pass: once the container is full, a round offers up to RUNNER_PASS_LIMIT = 20 more tasks to idle runners. CAPACITY_RULE is updated.
  • Reaping:
    • A claimed bee is reaped after 10 minutes without a renewal.
    • A runner's task is reaped by its offer (10 min), its lease (15 min) or its cap (360 min), via reapSilentRunners.
    • A bee the container runs is still reaped by age alone.
    • TASK_LEASE_RULE and f58 are updated.
  • Public cards now carry verdict and judgedHead for a finished row. PUBLIC_FIELDS is updated.
  • SOURCE_LINES are remapped to the new line numbers.

PRIORITY_RULE is unchanged. Production still applies no label priority: gHashTag/BrowserOS#534 is open and unmerged.

🤖 Generated with Claude Code

@gHashTag @claude
fix(queen): re-pin queen/dispatch to fix/queen-worker-provider-and-pr…
301d65e 
…ompt-size

feat/queen-supervisor (c25e1b02) no longer exists. Pin the branch production
deploys, fix/queen-worker-provider-and-prompt-size at 879799a8, and re-measure.

ts, queend, pg and live re-ran at the new pin; ts/queend/pg results equal the
old run (pg on PostgreSQL 17.7, 16 not on the host). lotus kept from the old
run (trinity pin unchanged, Zig 0.15.2 absent). Spec text updated for runner
lanes, the runner-only pass, renewal reaping of claimed and runner bees, and
verdict/judgedHead on public cards. PRIORITY_RULE unchanged (BrowserOS#534
unmerged). Replay 2488/2488, check ok, self-check catches every fault.

Closes #6368

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@github-actions github-actions Bot mentioned this pull request Oct 5, 2026
Merged
@github-actions

github-actions Bot commented Oct 5, 2026

Copy link
Copy Markdown
Contributor

PR Dashboard

Generated at: 2026-10-05 06:04:37 UTC

Summary

Status Count
Total Open PRs 47
PRs with Failing Checks 34
PRs with All Checks Green 13
READY 11
FAILING 34
PENDING 0
NO CHECKS YET 0

These columns do not partition: 11 + 34 + 0 + 0 = 45, and there are 47 open PRs. A PR is being counted twice or not at all.

Seal Status

  • ⚠️ STALE -- sha256(compiler.rs)=8597b6ded596 != 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).

@gHashTag
gHashTag merged commit fdcbc21 into master Oct 5, 2026
23 of 25 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

queen/dispatch: SUPERVISOR_BRANCH names a deleted branch; production runs fix/queen-worker-provider-and-prompt-size

1 participant

@gHashTag