Normative: Canonicalize value of extension in ApplyUnicodeExtensionToTag #846
Conversation
FrankYFTang
changed the title
|
Thanks! I tested this change over in Serenity's LibJS and it looks good to me: > const loc = new Intl.Locale('en', {calendar: 'islamicc'});
> loc.toString()
"en-u-ca-islamic-civil"
> loc.calendar
"islamic-civil"We also now pass |
| @@ -136,7 +136,8 @@ <h1> | |||
| 1. Let _optionsValue_ be _options_.[[<_key_>]]. | |||
| 1. If _optionsValue_ is not *undefined*, then | |||
| 1. Assert: Type(_optionsValue_) is String. | |||
| 1. Set _value_ to _optionsValue_. | |||
| 1. Let _optionsUValue_ be the ASCII-lowercase of _optionsValue_. | |||
| 1. Set _value_ to the String value resulting from canonicalizing _optionsUValue_ as a value of key _key_ per <a href="https://unicode.org/reports/tr35/#processing-localeids">Unicode Technical Standard #35 Part 1 Core, Annex C LocaleId Canonicalization Section 5 Canonicalizing Syntax, Processing LocaleIds</a>. | |||
There was a problem hiding this comment.
ResolveLocale also explicitly converts key to ASCII lower-case before this step. Should we do the same here? (key is already in ASCII lower-case, so it's not strictly required to perform any case conversions, neither here nor in ResolveLocale.)
There was a problem hiding this comment.
In here, Inside 14.1.3 ApplyUnicodeExtensionToTag ( tag, options, relevantExtensionKeys )
https://tc39.es/ecma402/#sec-apply-unicode-extension-to-tag
5. For each element key of relevantExtensionKeys, do
and only place call it is in
14.1.1 Intl.Locale ( tag [ , options ] )
https://tc39.es/ecma402/#sec-Intl.Locale
29. Let r be ApplyUnicodeExtensionToTag(tag, opt, relevantExtensionKeys).
where relevantExtensionKeys is from
2. Let relevantExtensionKeys be %Locale%.[[RelevantExtensionKeys]].
and %Locale%.[[RelevantExtensionKeys]] is defined in
https://tc39.es/ecma402/#sec-intl.locale-internal-slots
"The value of the [[RelevantExtensionKeys]] internal slot is « "ca", "co", "hc", "kf", "kn", "nu" ». If %Collator%.[[RelevantExtensionKeys]] does not contain "kf", then remove "kf" from %Locale%.[[RelevantExtensionKeys]]. If %Collator%.[[RelevantExtensionKeys]] does not contain "kn", then remove "kn" from %Locale%.[[RelevantExtensionKeys]]."
I agree with you that the "converts key to ASCII lower-case" in ResolveLocale is also unnecessary. But that is another editoral issue we can change separately
|
So the rule here is, "canonicalize if we recognize something, but if there's an unrecognized value, let it continue as uninterpreted text" right? And because of this latter part, we can't convert things like first day of the week into a number in this place, right? I'm OK with this PR, but if you could summarize how these things fit together, that'd help my understanding. |
|
Ben- why do you merge this PR now. This PR is on the agenda for TC39 discussion tomorrow https://github.com/tc39/agendas/blob/main/2024/02.md we should not merge it until TC39 agree with it first. |
|
This is on the agenda for discussion on 2024-02-06 TC39 meeting https://github.com/tc39/agendas/blob/main/2024/02.md |
|
oh my goodness, sorry about that! I've pushed a revert commit for it. |
|
Has concensus from TG1 this morning around 11:25am PST (2024-02-06) Please reland this. @ben-allen |
|
Hi @ben-allen - friendly ping, can the commit be restored? Thanks! |
the calenar name should be canonicalized Spec change tc39/ecma402#846
the calenar name should be canonicalized Spec change tc39/ecma402#846
Close #707