Normative: Change date AOs to accept mathematical values

[Aditi-1400]

This PR fixes issue #2315, by redefining following Date operations to calculate in mathematical values instead of Number values:

• MakeDay -> ISODateToEpochDays
• MakeDate -> EpochDaysToEpochMs
• DaysInYear -> MathematicalDaysInYear
• DayFromYear -> EpochDayNumberForYear
• TimeFromYear -> EpochTimeForYear
• YearFromTime -> EpochTimeToEpochYear
• InLeapYear -> MathematicalInLeapYear
• MonthFromTime -> EpochTimeToMonthInYear
• DayWithinYear -> EpochTimeToDayInYear
• DateFromTime -> EpochTimeToDate
• WeekDay -> EpochTimeToWeekDay

These operations would be unified when tc39/ecma262#1087 is fixed

[codecov]

Codecov Report

Merging #2518 (cd3ec37) into main (1aa6b80) will increase coverage by 0.00%.
The diff coverage is n/a.

@@           Coverage Diff           @@
##             main    #2518   +/-   ##
=======================================
  Coverage   96.16%   96.16%           
=======================================
  Files          20       20           
  Lines       11285    11289    +4     
  Branches     2163     2163           
=======================================
+ Hits        10852    10856    +4     
  Misses        369      369           
  Partials       64       64           

see 1 file with indirect coverage changes

[ptomato]

Thanks!

[ptomato]

Looks good, thanks!.

I might suggest being more specific about what is meant by "(epoch) time", since it could be confusing that we mean epoch milliseconds here, whereas in the rest of Temporal we use epoch nanoseconds. But on the other hand, maybe this is fine — it's very clear already that these are MV duplicates of the ECMA-262 operations, and the names are quite long already without adding "milliseconds" to them.

[gibson042]

I hate seeing this redundancy, but I don't currently have enough time to determine which parts of it are necessary.

[ptomato]

I think the usage of EpochTimeToEpochYear (vs YearFromTime in ECMA-262) makes it necessary, until someone can go through ECMA-262 and prove that using the MV version exclusively would be unobservable. If this turns out to really be redundant, then removing it will by definition be an unobservable change.

[ptomato]

@Aditi-1400 One last thing, would you mind removing .DS_Store and rebasing this on main so that the test262 update is not necessary? (I think it updates test262 to the same commit that it already has on main)

[justingrant]

Also, please squash into one commit please!

[ptomato]

For future reference, this PR achieved consensus at the March 2023 TC39 meeting.