Observation-class. No runtime behaviour, no failing assertion, nothing a user hits — two prose comments in packages/core/src/utils/** cite formatPercent as the live example of the divide-by-100 route, and that stopped being true when objectui#4590 (PR #4595) landed.
Found while implementing #4590. Not fixed there: packages/core/src/utils/** was outside that card's declared surface — it CONSUMES style: 'percentPoints' rather than editing its home — so this is recorded rather than folded in.
The two sites
packages/core/src/utils/dataset-format.ts, in the formatMeasure percent comment:
Intl's style: 'percent' wants a fraction — so routing through that (the way formatPercent does) would mean dividing by 100 for Intl to multiply straight back
packages/core/src/utils/__tests__/dataset-format.percent-convention.test.ts:40, in the file header:
The obvious way to adopt the locale's convention is Intl's style: 'percent' — the way formatPercent does it — but that style wants a FRACTION
Both were accurate when written. formatPercent now renders through style: 'percentPoints' with no division, so each parenthetical points at a function that no longer does the thing it is being cited for.
Why it is worth a line rather than nothing
The surrounding argument is still correct and still valuable — it is the measured case for why the percentage-points route was chosen, and the tie/extreme pins under it are what keep that route honest. Only the "who does it the other way" example has expired. A reader checking the citation finds the opposite of what it claims, which is the kind of thing that makes a correct comment get distrusted wholesale.
The narrower fix is to change the tense and drop the example: the divide-by-100 route is a real alternative worth arguing against on its own merits, and after #4590 the repo has no caller taking it — which is itself the better sentence.
Not in scope here
The 27,581 / 1,200,013 figure in those comments came from #4576's grid. #4590 re-measured a slightly different grid (0.005 steps to 2,000, precisions 0/1/2 = 1,200,003 forms) and got 27,577. Neither number is wrong; they are different grids. Worth stating the grid alongside the count if these lines are touched anyway, so the two are not read as a discrepancy.
Generated by Claude Code
Observation-class. No runtime behaviour, no failing assertion, nothing a user hits — two prose comments in
packages/core/src/utils/**citeformatPercentas the live example of the divide-by-100 route, and that stopped being true when objectui#4590 (PR #4595) landed.Found while implementing #4590. Not fixed there:
packages/core/src/utils/**was outside that card's declared surface — it CONSUMESstyle: 'percentPoints'rather than editing its home — so this is recorded rather than folded in.The two sites
packages/core/src/utils/dataset-format.ts, in theformatMeasurepercent comment:packages/core/src/utils/__tests__/dataset-format.percent-convention.test.ts:40, in the file header:Both were accurate when written.
formatPercentnow renders throughstyle: 'percentPoints'with no division, so each parenthetical points at a function that no longer does the thing it is being cited for.Why it is worth a line rather than nothing
The surrounding argument is still correct and still valuable — it is the measured case for why the percentage-points route was chosen, and the tie/extreme pins under it are what keep that route honest. Only the "who does it the other way" example has expired. A reader checking the citation finds the opposite of what it claims, which is the kind of thing that makes a correct comment get distrusted wholesale.
The narrower fix is to change the tense and drop the example: the divide-by-100 route is a real alternative worth arguing against on its own merits, and after #4590 the repo has no caller taking it — which is itself the better sentence.
Not in scope here
The 27,581 / 1,200,013 figure in those comments came from #4576's grid. #4590 re-measured a slightly different grid (0.005 steps to 2,000, precisions 0/1/2 = 1,200,003 forms) and got 27,577. Neither number is wrong; they are different grids. Worth stating the grid alongside the count if these lines are touched anyway, so the two are not read as a discrepancy.
Generated by Claude Code