ecma402
ecma402 copied to clipboard
Editorial: Move NumberFormat objects to DateTimeFormat constructor an…
…d add new internal slots
fix #563
@ryzokuken done. Thanks a lot for the review!
CC @FrankYFTang -- can you review this proposed change from @longlho and comment on whether it would impact implementations like V8?
@sffc I think @anba mentioned that it wouldn't affect ICU-based implementations.
CC @FrankYFTang -- can you review this proposed change from @longlho and comment on whether it would impact implementations like V8?
I believe this is editorial and has no implementation impact to v8.
@gibson042 @leobalter can we get one editor +1 and merge this? :D
@anba could you also take a look at this?
@anba care for another look? thanks!
Ping @anba PTAL.
@ryzokuken any update on this? ping @anba as well
@longlho I'd been waiting for @anba to give their approval before merging but I suppose I could go through their comments later today and see if their concerns have been addressed and merge if they are busy right now.
thanks a lot for the fixes @gibson042 !
gentle ping @anba :)
ping @ryzokuken
@ryzokuken do you know if this is still mergeable? happy to rebase if that's the case
Sorry for the delay, @longlho. It looks all good from my side, but looks like Andre still couldn't review it. If you rebase it, perhaps we could merge it and file an issue to address their concerns in a follow-up?
@anba @longlho @ryzokuken Who needs to take the next action on this PR?
Hmm, this change may no longer be possible, because the current Temporal draft spec changes Intl.DateTimeFormat to be able to format Temporal objects with different resolved patterns. For example when formatting Temporal.PlainTime and Temporal.PlainDateTime with a single Intl.DateTimeFormat, the resolved pattern for Temporal.PlainTime can have a different amount of fractionalSecondDigits when compared to the resolved pattern for Temporal.PlainDateTime. Therefore it's not possible to store a single [[FractionalSecondNumberFormat]] internal slot.
Oh I see, yeah, I don't think we should do this. The spec does not forbid implementations from doing internal optimizations like this so long as the observable behavior is the same.
I'll close the PR since it has been idle for a long time and it's not clear if it is a good PR based on this and @anba's comment.