Filed unassigned by the #10274 dev seat (session session_016gcKVsiywU9CcS96S5t9qD) as an out-of-scope finding hit while implementing that card. Observation class — recorded, not queued.
Observation
packages/spec/src/data/api-methods-batch-conformance.test.ts:63-64 justifies the SINGLE_RECORD_WRITE_ONLY entry for API keys with a cross-repo measurement, and anchors it in the present tense:
bulk either — both read off the console build this release pins
(.objectui-sha pin 6314e87f2, packages/plugin-grid):
.objectui-sha is 9a3daf8d37ad973a621e5edd276fe32467f90684. 6314e87f2 is not the pin this release builds against, so "the console build this release pins" is false as written. The three bullets it introduces (no checkbox column, no bulkActions / bulkActionDefs / selection on any of the object's four list views, bulk-delete gated off three ways) are all claims about packages/plugin-gridat that older tree and have not been re-measured since.
Relationship to #10274 — what that card did and deliberately did not do
#10274 landed check:objectui-pin-citations, which enforces two spellings in packages/spec/src:
| spelling | meaning | checked against the pin file |
|---|
`.objectui-sha` = `<sha>` | this IS the pin we build against | yes |
`.objectui-sha` pin `<sha>` | measured AT that pin (a dated record) | no |
This site was written in neither form ((`.objectui-sha` 6314e87f2, …), which the new gate rejects as an unfindable spelling. #10274 normalized the citation token only — `.objectui-sha` 6314e87f2 → `.objectui-sha` pin `6314e87f2` — and changed no word of the surrounding claim, because that seat re-measured the four #9881/#9972 records and nothing else. It is recorded there and here rather than quietly absorbed.
The consequence is worth stating plainly: the sentence still asserts a current-pin identity, while the citation beside it now sits in the non-asserting spelling, so the gate is green over it. That is the honest outcome of "normalize the spelling, do not fake the measurement" — but it leaves a real mismatch, which is this card.
Why it is worth recording rather than shrugging
The entry decides a conformance rule: update is allowed as a single-record write for API keys precisely because the multi-select surface the rule protects does not exist there. That "does not exist" is a measurement, and a measurement whose tree is not the tree we ship is not evidence for the rule it justifies. plugin-grid's selection/bulk affordances are exactly the kind of surface that moves between pins.
What closing it would take
One of two, decided by measurement rather than in advance:
- Re-measure at
9a3daf8d3 — check the four list views and ObjectGrid's multi-select auto-enable in packages/plugin-grid at the current pin. Still no selection surface ⇒ update the sha and promote the citation to the asserting spelling (`.objectui-sha` = `<sha>`), which puts it under check:objectui-pin-citations and makes every future bump re-ask the question. - A selection surface now exists ⇒ the
SINGLE_RECORD_WRITE_ONLY entry is what needs revisiting, not the citation.
If instead the intent was only ever "this is where it was measured once", restating the sentence in the past tense makes the existing pin spelling correct and closes this with no cross-repo work — but that is a judgement about what the entry means, not a re-measurement, so it should be made deliberately.
Dedup
Searched open and closed issues for this file, for 6314e87f2, and for stale pin citations generally: the matches (#10129, #10136, #9356, #7411, #6813, #6663, #6159, #4731, #6089) are all about the pin-bump machinery itself. #10274 is the only card about prose that cites a pin, and it asserts these four sites only — this one it explicitly does not assert.
Backlink: #10274.
Filed unassigned by the #10274 dev seat (session
session_016gcKVsiywU9CcS96S5t9qD) as an out-of-scope finding hit while implementing that card. Observation class — recorded, not queued.Observation
packages/spec/src/data/api-methods-batch-conformance.test.ts:63-64justifies theSINGLE_RECORD_WRITE_ONLYentry for API keys with a cross-repo measurement, and anchors it in the present tense:.objectui-shais9a3daf8d37ad973a621e5edd276fe32467f90684.6314e87f2is not the pin this release builds against, so "the console build this release pins" is false as written. The three bullets it introduces (no checkbox column, nobulkActions/bulkActionDefs/selectionon any of the object's four list views, bulk-delete gated off three ways) are all claims aboutpackages/plugin-gridat that older tree and have not been re-measured since.Relationship to #10274 — what that card did and deliberately did not do
#10274 landed
check:objectui-pin-citations, which enforces two spellings inpackages/spec/src:`.objectui-sha` = `<sha>``.objectui-sha` pin `<sha>`This site was written in neither form (
(`.objectui-sha` 6314e87f2, …), which the new gate rejects as an unfindable spelling. #10274 normalized the citation token only —`.objectui-sha` 6314e87f2→`.objectui-sha` pin `6314e87f2`— and changed no word of the surrounding claim, because that seat re-measured the four #9881/#9972 records and nothing else. It is recorded there and here rather than quietly absorbed.The consequence is worth stating plainly: the sentence still asserts a current-pin identity, while the citation beside it now sits in the non-asserting spelling, so the gate is green over it. That is the honest outcome of "normalize the spelling, do not fake the measurement" — but it leaves a real mismatch, which is this card.
Why it is worth recording rather than shrugging
The entry decides a conformance rule:
updateis allowed as a single-record write for API keys precisely because the multi-select surface the rule protects does not exist there. That "does not exist" is a measurement, and a measurement whose tree is not the tree we ship is not evidence for the rule it justifies.plugin-grid's selection/bulk affordances are exactly the kind of surface that moves between pins.What closing it would take
One of two, decided by measurement rather than in advance:
9a3daf8d3— check the four list views andObjectGrid's multi-select auto-enable inpackages/plugin-gridat the current pin. Still no selection surface ⇒ update the sha and promote the citation to the asserting spelling (`.objectui-sha` = `<sha>`), which puts it undercheck:objectui-pin-citationsand makes every future bump re-ask the question.SINGLE_RECORD_WRITE_ONLYentry is what needs revisiting, not the citation.If instead the intent was only ever "this is where it was measured once", restating the sentence in the past tense makes the existing
pinspelling correct and closes this with no cross-repo work — but that is a judgement about what the entry means, not a re-measurement, so it should be made deliberately.Dedup
Searched open and closed issues for this file, for
6314e87f2, and for stale pin citations generally: the matches (#10129, #10136, #9356, #7411, #6813, #6663, #6159, #4731, #6089) are all about the pin-bump machinery itself. #10274 is the only card about prose that cites a pin, and it asserts these four sites only — this one it explicitly does not assert.Backlink: #10274.