duplicate-bodies is red on master since #6070: bless the four fixture copies into the ledger

[gHashTag]

Context

The duplicate-bodies job of the Duplicate Body Ratchet workflow (it runs python3 tools/dupe_scan.py) has been red on master since PR #6070 landed (0886f014c, 2026-10-04T13:09Z) and is still red on 2e852cf82 (run 37206113836). Every pull request opened since then inherits the red, and every review has to tell it apart from a red the pull request caused. That is review time spent on a red nobody's change is responsible for.

The log on master:

duplicate bodies: 598 of 4880 in 173 group(s) across 1193 spec(s)
::error::a body written under local_branch is now copied [2] time(s) and was in no ledger
::error::a body written under nested_tail is now copied [2] time(s) and was in no ledger
::error::a body written under ordinary_if is now copied [2] time(s) and was in no ledger
::error::a body written under wide_tail is now copied [2] time(s) and was in no ledger

The four groups sit in two compiler fixtures, specs/compiler/rust_tail_returns.t27 and specs/compiler/verilog_branch_context.t27, which bootstrap/tests/rust_tail_returns.rs and bootstrap/tests/verilog_branch_context.rs compile and execute. A fixture is self-contained by design: each one tests one backend on its own, so a short body that happens to equal another fixture's body is deliberate, not a copy to share. The scanner's own message names the fix for a deliberate copy: move the ledger (python3 tools/dupe_scan.py --bless). PR #6070 did not.

Reference run, 2026-10-04T14:40Z, on master 2e852cf82 in a scratch copy: --bless then a plain scan. The ledger diff is exactly four added lines, and the scan turns green:

blessed: 162 group(s), 598 body(ies) into tools/duplicate_bodies_baseline.txt
duplicate bodies: 598 of 4880 in 173 group(s) across 1193 spec(s)
ok: no new duplicate body, and none of the known groups grew.

+local_branch 2
+nested_tail 2
+ordinary_if 2
+wide_tail 2

What to write

1. Run python3 tools/dupe_scan.py --bless from the repository root. It rewrites tools/duplicate_bodies_baseline.txt and nothing else.
2. Check the diff of that file before committing: it MUST add exactly the four lines above (each in its sorted place) and remove none. If it adds or removes anything else, master has moved since this issue was written: do not commit; say in the pull request which extra lines appeared, because each one is a new copy somebody has to judge.
3. Do not edit either fixture spec. They are compiled by Rust tests that a bee cannot run here, and the fix is the ledger.

User Scenarios & Testing

• Given master at 2e852cf82 or later with no other new copy, when python3 tools/dupe_scan.py runs on the pull request's head, then it prints ok: no new duplicate body, and none of the known groups grew. and exits 0.
• Given the pull request's diff, when it is read, then it touches only tools/duplicate_bodies_baseline.txt and only adds the four lines named above.

Requirements

• FR-001: the change MUST be produced by python3 tools/dupe_scan.py --bless, not typed by hand (the ledger's sort and format belong to the tool).
• FR-002: the diff MUST touch only tools/duplicate_bodies_baseline.txt and MUST only add local_branch 2, nested_tail 2, ordinary_if 2, wide_tail 2.
• FR-003: the fixture specs MUST NOT change.

Acceptance criteria

Each line is a check the Queen's runner executes itself (commands chosen from what it admits; it does not run python3 or git). Computed by a reference run on 2e852cf82: all five fail on master today (0, 0, 0, 0, 161) and pass after --bless.

• grep -c -x 'local_branch 2' tools/duplicate_bodies_baseline.txt prints 1
• grep -c -x 'nested_tail 2' tools/duplicate_bodies_baseline.txt prints 1
• grep -c -x 'ordinary_if 2' tools/duplicate_bodies_baseline.txt prints 1
• grep -c -x 'wide_tail 2' tools/duplicate_bodies_baseline.txt prints 1
• cat tools/duplicate_bodies_baseline.txt | wc -l prints 165

Checked by the pull request's own CI, not by the runner: python3 tools/dupe_scan.py ends with ok: no new duplicate body, and none of the known groups grew. and exits 0 (the duplicate-bodies check is green), and the diff touches only tools/duplicate_bodies_baseline.txt.

Boundary

• tools/duplicate_bodies_baseline.txt

Refs #5776 (review-speed plan row B33). Epic: #5906.

[gHashTag]

Criteria rewritten. The first version numbered them (1.), and the Queen reads only - bullets under a criteria heading, so her contract for the first bee was the three FR lines -- prose, nothing her runner could execute (criteria_with_source in tools/queen/criteria_backfill.py, the twin pinned to her parser). The five new bullets are commands her runner admits; each was run against master's ledger (fails: 0, 0, 0, 0, 161) and the blessed one (passes). The first attempt was escalated at 14:44:21Z.