Pin the two declined-re-anchoring figures in the ratchet header — #1341 reflowed them into worked rows on purpose, and the pin belongs beside the existing header-table pin, not in a second parser

[os-trump]

Filed by the repo:hotcrm PM seat on behalf of the #1335 dev, which deliberately left this unpinned and asked where the pin belongs rather than putting it in the wrong file. ⛔ Do not start before PR #1341 merges — it is what reflows the sentence this card pins.

What #1341 left, and why it left it

scripts/check-source-token-ratchet.mjs's header records two anchor() outputs for re-anchorings that were declined as raises:

anchor(82,489)  = 87,000   vs committed  85,000   ⇒ declined, a RAISE
anchor(133,840) = 141,000  vs committed 140,000   ⇒ declined, a RAISE

Both correct today, both unpinned. #1335's card pre-authorised leaving them unpinned with a recorded reason, because the sentence wrapped across a comment line break and a row-style regex would have been brittle.

⭐ PR #1341 removed that obstacle: it reflowed the sentence into one worked line per declined re-anchoring, in the shape of the table above it. So the reason for declining the pin is discharged, and what remains is mechanical.

⛔ Why the pin goes in test/source-token-ratchet.test.ts and nowhere else

The #1335 dev's own analysis, and it is the ruling:

The natural home already pins the header table beside this very sentence and already holds every helper the pin needs — source(), num(), ceilingOf(), and the RUN / RUN_READINGS pair that ties a row's reading to the anchoring run it came from.

⚠️ Putting it in test/docs-readme-token-figures.test.ts instead would mean copying that machinery into a second file — a second parser for one artefact, which is the defect class #1335 and #1339 exist to remove. ⛔ Do not do that. A fix that commits the defect it is closing is not a fix.

Scope — the shape, carried verbatim from the dev that declined it

One case in test/source-token-ratchet.test.ts, beside the existing header-table pin, reusing its helpers. ~12–15 lines, no new machinery, ⛔ no change to the gate:

1. A single-line regex for the reflowed rows.
2. Per row: anchor(reading) equals the stated figure.
3. Per row: the stated figure is greater than ceilingOf(label) — this is the "declined as a RAISE" claim itself, and it is the assertion that makes the row mean something rather than merely parse.
4. Per row: tie the reading to the recorded run for its date, via the existing runs() helper.

⚠️ Read the existing header-table pin first — including its comment instructing its own maintainer ⛔ never to relax the regex when the table's shape changes. Match that discipline; ⛔ do not invent a second convention.

⛔ Do not change any ceiling value, BUFFER, the README banner, or the gate itself. ⛔ Do not touch scripts/check-source-hygiene.mjs (#1314 / #1339's file).

Acceptance

⭐ Mutate, do not merely pass. Every figure named here is correct today — that is the defect, not the test.

• Change a committed ceiling so a declined re-anchoring becomes an accepted one, and confirm the "greater than" assertion goes red naming its row.
• Falsify one stated figure and confirm the anchor() assertion goes red naming its row.
• Mutations proved on disk by blob hash before any verdict is read; restore with git checkout HEAD -- PATH (⛔ the bare form restores from the index), proved by an empty git diff HEAD.
• ⚠️ A zero-matching anchor is a no-op whose green reading is void — report it void and re-run.
• ⚠️ Prove non-vacuity structurally, as test(ratchet): read the buffer and the ceilings from the gate, not from prose beside it #1341 did: assert the parsed rows equal an expected label set, so an unparseable sentence fails rather than passing empty.

Related

• Three more copies of the ratchet's ruled figures with no producer-side pin — one calls itself "reused" while being a hand copy, and one has already rotted against the ceiling #1320 moved #1335 / PR test(ratchet): read the buffer and the ceilings from the gate, not from prose beside it #1341 — reflowed the sentence and declined the pin, with the reason recorded
• The hygiene banner binds the wrong count to the marker and header checks — it says 339 + 3 where those checks read allTs = 329, and two cards this round were told to trust that line as the producer #1339 — the hygiene gate's mis-bound banner figure; ⛔ different gate, different file, do not conflate
• docs: delete three hand-maintained integers from prose, keeping the claims they propped up #1338 — the opposite ruling for figures with no producer: delete the numeral. ⚠️ These two have a producer, which is why they get pinned instead