Next batch of editorial issues #2934
Commits on Sep 11, 2024
-
Tests: Fix setup code in test file
If running this test file without first importing the proposal polyfill, it would fail due to the new intrinsics not being defined. We never noticed this because we usually run datemath.mjs before this file.
-
Polyfill: Perform eraYear checks only if calendar has eras
In the calendar-specific code, in isoToCalendarDate(), we should not be computing the era year if the calendar doesn't have eras.
-
Polyfill: Ignore eraYear in calendars without eras
In 'hebrew', 'chinese', and 'dangi' calendars, there are no eras. Any eraYear property given in calendar fields must be ignored, as per CalendarResolveFields in the spec text. Adjust the helpers for Hebrew and Chinese/Dangi calendars accordingly; the eraYear was being used in place of the year in some cases. Closes: #2870
-
Editorial: Use calendar addition to calculate weeks start
This was an overzealous replacement of CalendarDateAdd with BalanceISODate spotted by Anba. It doesn't produce any observably different results, because in all supported calendars weeks are always 7 days; but it's still conceptually incorrect. Go back to using CalendarDateAdd here. Closes: #2881
-
Editorial: Fix typo in loop comparison operator
Otherwise the loop would not run. Closes: #2890
-
Editorial: Simplify early return condition in BubbleRelativeDuration
As pointed out by Anba, this AO does nothing if smallestUnit equals largestUnit, not just if smallestUnit equals "year". We can simplify the early return condition. See: #2890
-
Editorial: Remove dead clause in BubbleRelativeDuration
smallestUnit is a date unit. In the loop, unitIndex runs from the index before smallestUnit and is decremented, meaning that unit can only ever be "year", "month", or "week". Thanks to Anba for spotting this. See: #2890
-
Editorial: Use ToIntegerWithTruncation in GetRoundingIncrementOption
Same thing, just shorter and clearer. Closes: #2923
-
Editorial: Remove unthrowable exception from duration rounding
It's not possible for the duration to not be valid here, as _duration_ is a valid duration, and the duration being tested is _duration_ truncated to days, which can only ever be less than _duration_. Spotted by Anba. Closes: #2880
-
Polyfill: Remove eras from buddhist, ethioaa, indian, islamic*, persi…
…an calendars Dates in these calendars display an era when they are formatted in English using Intl.DateTimeFormat, but not when they are formatted for the region where the calendar is used. That likely indicates they shouldn't have eras. Remove `constantEra` from the calendar helpers, and introduce a new helper that calculates dates relative to a fixed year zero. Closes: #2899
-
Polyfill: Implement era aliases
As of the current state of the Intl Era and Month Codes proposal, the eras table in https://tc39.es/proposal-intl-era-monthcode/#table-eras has a canonical name for each era, and a number of aliases which should also be accepted. Legacy ICU names such as "era0" should not be accepted, although they still need to be handled when using Intl.DateTimeFormat to convert dates. Closes: #2901
-
Polyfill: Save a few property copies in calendar helpers
The makeHelper___() functions already return a new object, so we don't need to Object.assign() them to {}. -
