You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
[core] Two more sites carry the bare 27,581 of 1,200,013 percent figure with no grid named, so it still reads as a discrepancy against #4590's 27,577 #5608
Observation-class. No runtime behaviour, no failing assertion. Recorded while implementing objectui#4596, whose declared surface was exactly two files.
Background
Two measurements of the same phenomenon exist, both correct:
The counts differ because the call shapes differ, not because either is wrong. Note that naming "the grid" alone does not separate them — both cards describe the grid identically. The disambiguator is the surface measured.
objectui#4596 attached that qualification to the two comments inside its fence. Two further live sites carry the bare figure and were out of that fence:
packages/core/src/utils/number-display.ts:118 — the DisplayNumberFormatOptions.style doc comment: "27,581 of 1,200,013 ordinary-magnitude en-US forms move".
packages/core/src/utils/__tests__/number-display.percent-points.test.ts:96 — "Measured: 27,581 of 1,200,013 ordinary-magnitude en-US forms differ between the two routes."
Both are accurate. Neither says which grid or which call shape produced the number, so a reader who has also seen packages/fields/src/index.tsx:503 or packages/fields/CHANGELOG.md still meets two counts that look like a contradiction — which is the reading objectui#4596 set out to stop.
Why this is a line and not nothing
Same reasoning objectui#4596 gave for its own two sites: a figure a reader cannot reconcile is what makes a correct comment get distrusted wholesale. This is cheap to close and the correct form is already written in objectui#4596's PR (#5606) for the two sites it covered — the wording can be lifted.
Explicitly not in scope
The CHANGELOG.md entries in packages/core, packages/i18n and packages/fields also carry these figures. Those are historical release records and should not be edited — they were accurate as published.
Suggested size
XS, prose only. Two comment lines, zero executable change; a .changeset with empty frontmatter is the declaration form (packages/core/src/** is released source, so the presence gate does require the declaration).
Observation-class. No runtime behaviour, no failing assertion. Recorded while implementing objectui#4596, whose declared surface was exactly two files.
Background
Two measurements of the same phenomenon exist, both correct:
formatMeasure/formatDisplayNumber's call shape.formatPercent.The counts differ because the call shapes differ, not because either is wrong. Note that naming "the grid" alone does not separate them — both cards describe the grid identically. The disambiguator is the surface measured.
objectui#4596 attached that qualification to the two comments inside its fence. Two further live sites carry the bare figure and were out of that fence:
packages/core/src/utils/number-display.ts:118— theDisplayNumberFormatOptions.styledoc comment: "27,581 of 1,200,013 ordinary-magnitude en-US forms move".packages/core/src/utils/__tests__/number-display.percent-points.test.ts:96— "Measured: 27,581 of 1,200,013 ordinary-magnitude en-US forms differ between the two routes."Both are accurate. Neither says which grid or which call shape produced the number, so a reader who has also seen
packages/fields/src/index.tsx:503orpackages/fields/CHANGELOG.mdstill meets two counts that look like a contradiction — which is the reading objectui#4596 set out to stop.Why this is a line and not nothing
Same reasoning objectui#4596 gave for its own two sites: a figure a reader cannot reconcile is what makes a correct comment get distrusted wholesale. This is cheap to close and the correct form is already written in objectui#4596's PR (#5606) for the two sites it covered — the wording can be lifted.
Explicitly not in scope
The
CHANGELOG.mdentries inpackages/core,packages/i18nandpackages/fieldsalso carry these figures. Those are historical release records and should not be edited — they were accurate as published.Suggested size
XS, prose only. Two comment lines, zero executable change; a
.changesetwith empty frontmatter is the declaration form (packages/core/src/**is released source, so the presence gate does require the declaration).