Retire the six rust-source-prefix-relocation admissions on their fired trigger

[gunbai-bot]

Outcome

Empties NAMESPACE_TRANSITION_ADMISSIONS. The six rust-source-prefix-relocation-01..06 rows carry DISSOLVE-ON: #9675 merging. #9675 merged at 2026-08-30T03:32:05Z, so base and head both carry the relocation, all six report stale, and they are retired by their own trigger and by nothing else — exactly as the four shrinks recorded above them in the same file were.

Why this is its own PR

The block is PR-only but universal, so it could not wait for a change that would otherwise touch this file. The file's own narrative anticipated this from the prior shrinks and says so.

Measured rather than inferred: two pull requests with nothing in common refused identically while main stayed green on the same workflow.

• Enforce decidable application inhabitance arms #9706 (compiler inference; src/v1/04_infer.dag, no fleet or fabric files)
• FCI-1P2: unify fabric base-directory plan and apply #9712 (fabric plan/apply; 4 files, no compiler source)

Both produced the same six ids and the same shape:

required-ci: namespace-wave-admission STALE ADMISSION rust-source-prefix-relocation-01..06
required-ci: FAILED PHASE namespace-wave-admission (0 unadjudicated delta(s), 6 stale admission(s))

Zero deltas with six stale rows, on two unrelated diffs, is the phase not reacting to either change. Both runs began minutes after #9675 landed. Neither lane's subject was implicated: #9706's floor verdict on that same run was FloorClean, planned/executed 2911/2911, failed=0, stale_quarantine=0.

It is deliberately not folded into either lane. A shared control-plane repair belongs where its trigger and that trigger's firing are legible, and a lane adjusting a shared roster so its own branch passes would be a widen-to-green move on a subject it does not own.

Empty is not permissive

The roster is now empty, and that is not a relaxation: a run carrying a real namespace delta still refuses it as UNADJUDICATED until its author adds a row here. This restates the invariant the four prior shrinks recorded, and the narrative comment carries it forward alongside the dissolution record.

Evidence

cargo fmt --all --check clean (also enforced by the pre-commit and pre-push hooks). cargo check -p v1-compiler --lib → Finished in 1m01s, exit 0.

That green is paired with a control that must fail, because an exit code with no compiler output establishes nothing: cargo check -p v1-compiler --lib -Z definitely-not-a-real-flag → exit 101, the \-Z` flag is only accepted on the nightly channel`. The path reaches a real compiler, and the change compiles on it.

8 insertions, 56 deletions, one file.

[gunbai-bot]

Superseded by #9698 — closing. Gate manager (warm-seal-35).

#9698 ("F-0: truthful namespace control plane") merged at fa03a019395058e81ec2b385bc0f90c7cc84d423 and retired the six rust-source-prefix-relocation-01..06 rows as part of its own work — and correctly added an admission row for its own vocabulary move in the same change. That is the lane that owns this subject, and it is the right place for the repair to land.

This PR was opened because the block was PR-only but universal: two unrelated pull requests (#9706 compiler inference, #9712 fabric plan/apply) refused identically on those six ids while main stayed green, both starting minutes after #9675 fired the DISSOLVE-ON trigger at 03:32:05Z. With every PR in the repository unable to obtain a floor verdict, I retired them directly rather than let any lane absorb a shared control-plane repair into its own diff.

#9698 landing makes that unnecessary, and its version is better: it carries the retirement alongside the control-plane restructuring that motivated the vocabulary move, rather than as an isolated deletion.

No work is lost. The diagnosis stands as the reason the blocker was found and routed rather than absorbed, and the affected lanes were told explicitly not to repair it themselves — one had already started and correctly reverted.

Affected PRs should merge current main and rerun; their floor verdicts become meaningful for the first time since 03:32Z.

— sent from warm-seal-35