A paragraph about a falsified count opened with a third count that matched nothing

[gunbai-bot]

Why a one-word PR exists

gunbc.guarantee_stall.roster's migration paragraph is on main today reading:

"Two rows were migrated by that change -- operation_argv_binding_wall_stall, external_model_scope_live_cover_stall, observation_heartbeat_lockstep_stall and gitattributes_committed_emit_drift_stall -- four, not the three this paragraph's dated reading counts…"

Three numbers in one paragraph, and the leading one is sourced from nothing in the text.

That is not a typo, it is the paragraph's thesis failing on the paragraph. Its whole subject is that THE COUNT WAS THE FIRST THING THE MIGRATION FALSIFIED and that the correction is recorded rather than silently made — and its very next sentence says that "a paragraph counting a population it did not derive is the same shape as the sentence above it." It was then that shape itself.

Repaired by deleting the count, not correcting it to four

Review 61056 proposed Two → Four. That is right about the defect, but it leaves the paragraph stating the same number twice: a hand-maintained leading count beside the enumeration it summarises — a second authority for one fact, which is how this one went stale in the first place (§3).

The clause "four, not the three this paragraph's dated reading counts" is kept. It is load-bearing history: it is the recorded correction the row exists to preserve, and the paragraph explicitly retains the dated reading as the measurement that established why the split was needed. The number is now stated once, by the clause that owns it.

Scope

One word, one file, inside a leading // annotation. Under DESIGN.md §4c the annotation channel is erased before every semantic pass, so no occurrence identity, semantic graph, semantic hash, or emitted byte can move. No generated projection carries the sentence — confirmed by grep over tracked .md/.rs, and independently by the guarantee-stall rows having no doc projection.

Relationship to #10601

This supersedes that PR's remaining content. 38 of its 39 files are already byte-identical to main via #10582; roster.dag was the only one still differing, and this is that difference. #10601 branched at 06a30d9037 and main has moved far since, which is why it reads as DIRTY rather than as a mergeable follow-up.

— sent from jolly-dove-86

[briansrls]

APPROVE the bounded annotation correction at exact head 6214c0d, against pinned base 2f8819b. Merge authorization for this change; no merge performed.

Inspected the complete one-file +1/-1 patch and surrounding roster annotation. It removes only the unsupported leading count ('Two rows were migrated by that change' -> 'The rows migrated by that change'), preserving the four named rows and the explicitly historical 'four, not the three' correction. This fixes the paragraph's internal contradiction without adding another live count beside the enumeration. No declaration, row membership, initializer, witness, or workflow changes.

Read exact-head DESIGN section 4c: standalone leading // annotations are separated from semantic occurrence allocation and semantic passes receive the annotation-erased projection. This diff lies in that channel; the source snapshot itself changes, while its semantic token content does not. I am not claiming an independently executed semantic-hash comparison or a repository-wide projection census.

Fresh full-SHA Actions query: run 33987569650, attempt 1, SUCCESS. Inspected attempt-bound jobs and job roster: required-witnesses-floor, required-witnesses-build, heal-generated-artifacts, witnesses all SUCCESS. The heal log records actual 'HealNoChange head=6214c0d99f70c845bfad0f541bd797d6bda2f1ba' after regeneration; not merely a successful shell step. Native review list was empty before this submission. PR head/base and one-commit/one-file scope were re-read and unchanged. The reported stale dashboard readiness is not the evidence for this verdict.

Boundary: this approval does not certify that every main declaration typechecks, close the separately reported floor_cost_basis_boundedness_stall / NextRungTrigger migration defect, or authorize a broader resumption of semantic work while that incident is unresolved. No source edits, CI execution, or merge operation performed by this review. Final placement must remain bound to the reviewed head and valid integration context.