Stage 4 tracking

[bakkot]

This proposal achieved stage 3 at the November 2022 meeting. This is a tracking issue for getting stage 4.

• committee consensus
• open PR to the spec - PR
  • editors have approved the PR
• write test262 tests - issue, PR
  • merge test262 tests
• at least two implementations
  • JavaScriptCore: issue (shipping in Safari 17)
  • V8: issue (shipping in Chrome 122)
  • SpiderMonkey: issue
  • Serenity LibJS: PR
• significant in-the-field experience

[linusg]

Initial implementation in LibJS has landed: SerenityOS/serenity#16279 (thanks @IdanHo!)

[Nico314159]

What exactly can I do as an individual to help this proposal reach stage 4?
I ask because this proposal is (no pun intended) in the intersection of the venn diagram for "would be quite useful" and "has a good shot at eventual acceptance".
There's a lot of things like operator overloading, pipe operator, Decimal, and Records/Tuples that meet only the first and not the second, but this proposal I have a lot of hope for.

[bakkot]

@Nico314159 The thing which is needed at stage 3 is shipping implementations. There's implementations in all the major browsers, but I believe Chrome and Firefox are waiting for test262 tests before they ship. I've written tests at tc39/test262#3966 but they need review from a maintainer. So, nothing individuals can do right now except contribute to maintainership of test262.

Contributing test262 tests is often a good way to help stage 3 proposals advance. (Or proposals at the new stage 2.7.) In this case they're already written though.

[Nico314159]

@Nico314159 The thing which is needed at stage 3 is shipping implementations. There's implementations in all the major browsers, but I believe Chrome and Firefox are waiting for test262 tests before they ship. I've written tests at tc39/test262#3966 but they need review from a maintainer. So, nothing individuals can do right now except contribute to maintainership of test262.

Contributing test262 tests is often a good way to help stage 3 proposals advance. (Or proposals at the new stage 2.7.) In this case they're already written though.

Out of curiosity, why is it called stage 2.7?

[ljharb]

It's later than 2, close to 3, but not quite there yet.

[Nico314159]

It's later than 2, close to 3, but not quite there yet.

But I why specifically .7? Does it have to do with Euler's number? Or is it just an arbitrary choice?

[ljharb]

Pretty arbitrary i think.

[bakkot]

It was the outcome of an extremely tedious discussion, the dozens of pages of notes for which will be published in a couple of weeks if you need a soporific.

[Nico314159]

It was the outcome of an extremely tedious discussion, the dozens of pages of notes for which will be published in a couple of weeks if you need a soporific.

I genuinely cannot tell if this is a joke or not.

[bakkot]

Alas it is not. TC39 is as susceptible to bikeshedding as any other group. And we do publish transcripts of meetings. We had a little extra time this meeting, which probably contributed to this topic running long.

[frou]

Apparently the new set methods arrived in V8 a couple of days ago, so might show up in Chrome soonish?

[bakkot]

Yup, they should be in Chrome 122, which is stable on Feb 14.

[mhkeller]

Looks like this has shipped in Chrome and Safari and perhaps will come in Firefox 126, which I imagine will ship around May.

[linusg]

Implemented in Kiesel, thanks @sno2!