Add [[PluralRules]] internal slot like Intl.RelativeTimeFormat

[longlho]

I'm confused by some of the results so looking for guidance. E.g:

new Intl.NumberFormat('en', { style: 'currency', currencyDisplay: 'name', currency: 'GBP'}).format(1)
// 1.00 British pounds

I would expect it to be 1 British pound especially since new Intl.PluralRules('en').select('1.00') yields one.

The other confusion is around scientific notation and whether the plural rules should use the whole number or the mantissa.

Thanks!

[sffc]

Related: #64, #80

When scientific or compact notation is used, the whole number should be considered in Intl.PluralRules. For example, "1K" should get the plural form of 1000. A few locales have special rules for such numbers, which is being addressed in ICU-13836.

The behavior "1.00 British pounds" is correct. You get that out of Intl.PluralRules if you pass in the correct option:

new Intl.PluralRules('en', { minimumFractionDigits: 2 }).select(1)
// "other"

I'll need to check the spec for what the behavior is supposed to be when you pass the currency options bag to Intl.PluralRules.

[longlho]

Thanks that fixes it! What's your thought on introducing an internal [[PluralRules]] in NumberFormat (similar to Intl.RelativeTimeFormat here). That would clear out the options bag for PL. Right now the spec is fairly ambiguous when it comes to plural selection in unit/currency/compact.

[sffc]

That's a good idea. This would be a larger change than I'm interested in doing with Unified NumberFormat currently, because we already have significant changes from Stage 3. I'll label this as "future" and we can do a follow-on Editorial PR to clean this up and make the behavior less ambiguous.

[longlho]

Cool sgtm!

[longlho]

Another question:

new Intl.NumberFormat('en', {
  style: 'currency', 
  currency: 'ZWD', 
  currencyDisplay: 'name', 
  notation: 'scientific'
}).format(10000)
// '1E4 Zimbabwean dollar (1980–2008)'

instead of +1E4 Zimbabwean dollars (1980–2008) (seems like the mantissa dictates the plural form in this case?)

[sffc]

That's an ICU bug being addressed in https://unicode-org.atlassian.net/browse/ICU-13836 (the bug says "compact notation", but it also applies to scientific notation).

Here is the PR to fix it: unicode-org/icu#938

[sffc]

Upstream issue for potential remaining work items on this issue:

tc39/ecma402#399