coreui / coreui-pro / 35891915920

95%

main: 94%

push

github

refactor(Calendar): type the date and time family's options and state (#1398)

`CalendarConfig` published `selectionType` as `string` and the Calendar
constructor and `setConfig` took any object, so `new Calendar(el, {
selectionType: 'weeks' })` compiled; `dayFormat`, `monthFormat` and
`yearFormat` also advertised a function that `DefaultType` rejects. The
options now carry the value sets the calendar handles (`SelectionTypes`,
the Intl option values, `DisabledDate | DisabledDate[] | null`), the
constructor and `setConfig` take `Partial<CalendarConfig>`, and the
`render*Cell` options take the `meta` the calendar passes, typed as the
exported `CalendarCellMeta` and `CalendarDayCellMeta`. The view is
`ViewTypes`, which drops the casts on its reads. `getOrCreateInstance`
is inherited and still takes any config.

The config types of Calendar, DatePicker, DateRangePicker, TimePicker,
DateInput, TimeInput and DateRangeInput are exported, as fourteen other
plugins already do; DateInput and TimeInput get named types, and
DateRangeInput types `disabledDates`, `format` and `type` like DateInput.
The pickers' `calendarOptions` and `inputOptions` take those types. The
picker types describe the pickers' own options, not the ones they pass on
to their field and calendar by name.

SectionInput's protected state and key handlers are typed, the
typed-digit and typed-letter results share one `SectionEntry` type,
`setConfig` takes `ComponentConfig | null`, and the `as any` casts go:
`aria-value*` attributes get `String(...)` and `aria-selected` gets
`'true'`. Popup's `isShown`, `isMobile`, `hide`, `toggle` and
`_updatePosition` return what they return, `placement` is Floating UI's
`Placement`, and the focus trap is typed.

The type test against the emitted declarations imports the seven config
types and rejects a wrong value for each narrowed option, through the
constructor and `setConfig` for `selectionType`.

6001 of 6619 branches covered (90.66%)

Branch coverage included in aggregate %.

16 of 16 new or added lines in 5 files covered. (100.0%)

11103 of 11488 relevant lines covered (96.65%)

597.72 hits per line