Implement Intl.DateTimeFormat

[nekevss]

Closes #4438

This pull request aims to implement Intl.DateTimeFormat.

Posting as a draft early as I lost my initial work when my laptop died a few weeks back 🙃 So better safe then sorry currently.

There's still plenty of work to be done, but feel free to take a look early 😄

[github-actions]

Test262 conformance changes

Test result	main count	PR count	difference
Total	52,598	52,598	0
Passed	49,385	49,421	+36
Ignored	2,134	2,134	0
Failed	1,079	1,043	-36
Panics	0	0	0
Conformance	93.89%	93.96%	+0.07%

Fixed tests (36):

test/intl402/language-tags-invalid.js (previously Failed)
test/intl402/language-tags-with-underscore.js (previously Failed)
test/intl402/DateTimeFormat/subclassing.js (previously Failed)
test/intl402/DateTimeFormat/constructor-options-dayPeriod-invalid.js (previously Failed)
test/intl402/DateTimeFormat/constructor-options-throwing-getters-dayPeriod.js (previously Failed)
test/intl402/DateTimeFormat/constructor-options-throwing-getters-timedate-style.js (previously Failed)
test/intl402/DateTimeFormat/constructor-options-order-dayPeriod.js (previously Failed)
test/intl402/DateTimeFormat/constructor-options-calendar-invalid.js (previously Failed)
test/intl402/DateTimeFormat/timezone-legacy-non-iana.js (previously Failed)
test/intl402/DateTimeFormat/constructor-invalid-offset-timezone.js (previously Failed)
test/intl402/DateTimeFormat/constructor-options-timeStyle-invalid.js (previously Failed)
test/intl402/DateTimeFormat/constructor-options-throwing-getters-fractionalSecondDigits.js (previously Failed)
test/intl402/DateTimeFormat/constructor-options-numberingSystem-invalid.js (previously Failed)
test/intl402/DateTimeFormat/constructor-options-dateStyle-invalid.js (previously Failed)
test/intl402/DateTimeFormat/constructor-options-fractionalSecondDigits-invalid.js (previously Failed)
test/intl402/DateTimeFormat/constructor-options-timeZoneName-invalid.js (previously Failed)
test/intl402/DateTimeFormat/constructor-options-order-fractionalSecondDigits.js (previously Failed)
test/intl402/DateTimeFormat/timezone-invalid.js (previously Failed)
test/intl402/DateTimeFormat/constructor-calendar-numberingSystem-order.js (previously Failed)
test/intl402/DateTimeFormat/prototype/this-value-not-datetimeformat.js (previously Failed)
test/intl402/DateTimeFormat/prototype/format/related-year-zh.js (previously Failed)
test/intl402/DateTimeFormat/prototype/format/format-function-name.js (previously Failed)
test/intl402/DateTimeFormat/prototype/format/throws-value-non-finite.js (previously Failed)
test/intl402/DateTimeFormat/prototype/format/length.js (previously Failed)
test/intl402/DateTimeFormat/prototype/format/format-function-builtin.js (previously Failed)
test/intl402/DateTimeFormat/prototype/format/taint-Object-prototype.js (previously Failed)
test/intl402/DateTimeFormat/prototype/format/builtin.js (previously Failed)
test/intl402/DateTimeFormat/prototype/format/name.js (previously Failed)
test/intl402/DateTimeFormat/prototype/format/prop-desc.js (previously Failed)
test/intl402/DateTimeFormat/prototype/format/format-function-property-order.js (previously Failed)
test/intl402/DateTimeFormat/prototype/format/date-constructor-not-called.js (previously Failed)
test/intl402/DateTimeFormat/prototype/format/offset-timezone-gmt-same.js (previously Failed)
test/intl402/DateTimeFormat/prototype/format/format-function-length.js (previously Failed)
test/intl402/DateTimeFormat/prototype/format/no-instanceof.js (previously Failed)
test/intl402/DateTimeFormat/prototype/format/time-clip-near-time-boundaries.js (previously Failed)
test/intl402/DateTimeFormat/prototype/format/time-clip-to-integer.js (previously Failed)

[nekevss]

This is ready for review.

Just to note, this is primarily just the format functionality. The rest will need to be added in follow up PRs, but hopefully this keeps the general progress a bit more incremental.

EDIT: It also does not currently support Temporal. So that will need to be added in follow ups as well.

[sffc]

Nice work!

[sffc]

Thought: This code is a bit more verbose than I was expecting; locale preferences are supposed to make code like this easier to write. If you have ideas for new functions we could add to icu_locale to make this easier, please suggest.

[nekevss]

This has been pretty standard for resolution on all Intl builtins. There may be a way to make this resolution easier to work through, but ideally this portion of the code would at least live in a potential intermediary crate.

[jedel1043]

Yeah, this is an unfortunate side effect of ECMA402 requiring that all options and extensions must be tested against the data to verify that they're "supported" in some sense, or they should be excluded from the final locale.

I guess we could add on icu_locale a method to do "DataProvider-aware resolution", but it seems better to have that in an ICU4X-ECMA402 crate, since it's a very niche use case.

[hansl]

No notes. Maybe just a test for the parts that aren't supported by ICU4X (like H24). Otherwise LGTM.