Normative: Stage4 PR for proposal-intl-enumeration

[FrankYFTang]

Plan to suggest moving to Stage 4 in Nov 2022
https://github.com/tc39/proposal-intl-enumeration

It is already implemented in Chrome m99, Safari 15.4 and Firefox 93.

https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Intl/supportedValuesOf#browser_compatibility

[ljharb]

a RangeError, from an unknown key? i'd expect a TypeError here.

[FrankYFTang]

Why is that? the type of the key is String, we also throw RangeError in https://tc39.es/ecma402/#sec-getoption right?

[FrankYFTang]

other places in ECMA402 we also throw RangeError
https://tc39.es/ecma402/#sec-canonicalcodefordisplaynames
https://tc39.es/ecma402/#sec-canonicalizelocalelist
"a. If collation does not match the Unicode Locale Identifier type nonterminal, throw a RangeError exception."
in https://tc39.es/ecma402/#sec-initializecollator
"a. If calendar does not match the Unicode Locale Identifier type nonterminal, throw a RangeError exception.
a. If numberingSystem does not match the Unicode Locale Identifier type nonterminal, throw a RangeError exception.
b. If the result of ! IsValidTimeZoneName(timeZone) is false, then
"
in https://tc39.es/ecma402/#sec-initializedatetimeformat

"a. If calendar does not match the Unicode Locale Identifier type nonterminal, throw a RangeError exception.
a. If collation does not match the Unicode Locale Identifier type nonterminal, throw a RangeError exception.
a. If numberingSystem does not match the Unicode Locale Identifier type nonterminal, throw a RangeError exception.
"
in https://tc39.es/ecma402/#sec-Intl.Locale
also
https://tc39.es/ecma402/#sec-apply-options-to-tag
https://tc39.es/ecma402/#sec-initializenumberformat

"

[ljharb]

That seems very strange - RangeErrors are for when something is out of a range - enums aren't a "range", conceptually or otherwise.

[gibson042]

RangeError (emphasis mine)

Indicates a value that is not in the set or range of allowable values.

It's also used this way in ECMA-262:

• NumberToBigInt
• String.fromCodePoint
• String.prototype.normalize (an especially good analog)
• InitializeTypedArrayFromArrayBuffer

[ljharb]

fromCodePoint is an actual ascii range; the AOs are ranges as well; but you’re right that normalize is a good analog. It still seems super weird to me, since an enum simply isn’t a range of any kind.

[sffc]

Ping to the editors to review this PR.

[gibson042]

Looks good after resolving some editorial nits, mostly relating to #735 or #736 or recent ECMA-262 changes.

[FrankYFTang]

PTAL and merge if ready?

[gibson042]

A couple more phrase replacements.

[FrankYFTang]

PTAL and merge

[gibson042]

This still looks good to me, modulo the linting fix. @ryzokuken, could you look it over as well?

[justingrant]

Suggested change

	AvailableCanonicalTimeZones (
	AvailableCanonicalTimeZoneIdentifiers (

With Temporal approaching Stage 4, "TimeZone" will soon be ambiguous: does it mean string identifier or a Temporal.TimeZone object?

Therefore @gibson042 and I are planning to rename all identifier-related AOs with an "Identifier" suffix. Given that this is a new AO, should we rename it here too to save churn later?

[FrankYFTang]

how about do those after merge this PR. Temporal is not going to be ES2023 but this will be

[justingrant]

OK, that's fine with me. Not a big deal either way.

[ryzokuken]

LGTM overall, I am happy with throwing a RangeError here, since it describes the exact nature of the failure better than a TypeError would, as the latter may result in user confusion.