Update validation, coercion for all built-in fields.

[cjtenny]

Do not merge; RFC only in this state

This change does a few other things as well:

• Implement era, eraYear for all remaining objects.
• Implement eraYear/year split behavior for 'gregory' and 'japanese' calendars (NB: choice of anchor/default era in 'japanese' should be revisited with more cultural context).

I only included the specification for the new PrepareTemporalFields abstract operation; I'm still refactoring the spec text into all the other ToTemporal*Fields, ToRelativeTemporalObject, etc abstract operations and that's messy / incomplete at the moment. However, PrepareTemporalFields closely matches how ES.ToRecord was used throughout the polyfill already, so this brings the specification and polyfill much closer together.

There are a number of questions and FIXME comments; I'm also uncertain if the language I used around abstract operation references in a table is valid specification. Bear with me :) Will finish this up over the next day and change.

[codecov]

Codecov Report

Merging #1319 (a8c6dd5) into main (f8fa242) will decrease coverage by 0.82%.
The diff coverage is 71.52%.

@@            Coverage Diff             @@
##             main    #1319      +/-   ##
==========================================
- Coverage   95.61%   94.78%   -0.83%     
==========================================
  Files          19       19              
  Lines        9341     9561     +220     
  Branches     1440     1493      +53     
==========================================
+ Hits         8931     9062     +131     
- Misses        404      489      +85     
- Partials        6       10       +4     

Flag	Coverage Δ
test262	54.41% <33.14%> (-1.13%)	⬇️
tests	90.69% <60.84%> (-1.06%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
polyfill/lib/plainyearmonth.mjs	91.66% <41.66%> (-3.09%)	⬇️
polyfill/lib/plaindatetime.mjs	93.38% <42.85%> (-2.08%)	⬇️
polyfill/lib/plainmonthday.mjs	88.82% <63.63%> (-1.82%)	⬇️
polyfill/lib/calendar.mjs	86.95% <72.11%> (-4.41%)	⬇️
polyfill/lib/plaindate.mjs	93.67% <78.57%> (-1.00%)	⬇️
polyfill/lib/zoneddatetime.mjs	97.31% <78.57%> (-0.60%)	⬇️
polyfill/lib/ecmascript.mjs	95.54% <84.88%> (-0.30%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update f8fa242...3ff286b. Read the comment docs.

[cjtenny]

(Validation in getter prototypes is also TODO, but shouldn't be anything surprising given the table)

[ptomato]

I think the removal of ToRecord is an improvement in particular!

[justingrant]

Partial review here-- will finish reviewing later today but wanted to get you initial notes ASAP.

[ptomato]

Would it be possible to separate out the validation changes from the addition of era/eraYear? I think since the latter is newer in this PR, it might take another round of review, whereas the validation changes are basically ready to go with minor comments. The changes for monthCode in #1203 also need the validation, so it would be good to get that merged sooner rather than later.

[cjtenny]

It would be harder to pull out era/eraYear because the current implementation contains era and year used in inconsistent ways; trying to revert to a base state would have to implement some validation with those, or break things. The era & eraYear portions of this change are small, so hopefully they don't take up too much review surface - I'll address the outstanding bug in the polyfill type lib separately, since it's not related to validation.

[ptomato]

A few minor comments.