Polyfill: Fix daysInYear and inLeapYear in Islamic calendars

[justingrant]

Fixes #2740. There were two separate problems with the islamic calendar implementation in the Temporal polyfill, both fixed in this PR:

• Previous implementation assumed that when the 12th month had 30 days, then it was a leap year. But the islamic-umalqura calendar appears to break this rule. New implementation calculates the number of days in the year; if 355 then it's a leap year.
• Previous implementation had a bug where it was using ISO date instead of calendar date when calculating daysInYear.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (af85ca8) 96.38% compared to head (59f8cae) 96.73%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2743      +/-   ##
==========================================
+ Coverage   96.38%   96.73%   +0.35%     
==========================================
  Files          21       21              
  Lines       12444    12444              
  Branches     2250     2255       +5     
==========================================
+ Hits        11994    12038      +44     
+ Misses        395      349      -46     
- Partials       55       57       +2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[MohsenAlyafei]

Great

[ptomato]

Seems correct to me. I will try to review the tests soon. BTW, it seems like the new tests are not covering dayOfYear and that's partly why codecov is complaining, but I don't understand why it's also complaining about inLeapYear since the tests are definitely covering that.

[MohsenAlyafei]

@justingrant as pointed out by @ptomato, the dayOfYear is not yet fixed.

The day of the year ordinal integer is the difference in days between the date in question and the start of the year of that date + 1.

[justingrant]

BTW, it seems like the new tests are not covering dayOfYear and that's partly why codecov is complaining,

@justingrant as pointed out by @ptomato, the dayOfYear is not yet fixed.

Oh! I'll take a look.

[justingrant]

I added tests for dayOfYear (and the trivial tests for monthsOfYear too) but I am not able to repro any problems. All the tests pass.

@MohsenAlyafei could you provide an example test case where the code in this PR doesn't correctly calculate dayOfYear for a particular calendar?

Note that the Temporal polyfill uses Intl.DateTimeFormat for its source of calendar data (which in turn gets its calendar calculations from ICU) so if there's a problem with how Intl.DateTimeFormat (and hence ICU) is calculating dates, then there's not much we can do here... the fix will have to happen upstream.

[ptomato]

dayOfYear seems to work correctly here:

d = Temporal.PlainDate.from({year:1445, month:1, day:1, calendar:'islamic-umalqura'})
  // => Temporal.PlainDate <2023-07-19[u-ca=islamic-umalqura]>
d.dayOfYear
  // => 1
d.subtract({days:1}).dayOfYear
  // => 354