Path to Stage 4

[acutmore]

Stage 4

• Two compatible implementations which pass the Test262 acceptance tests
  • Firefox
    • commit
    • tracking issue for impl
    • tracking issue for shipping unflagged
  • es-shims: allKeyed / allSettledKeyed
  • Boa
  • Chrome
  • Safari
  • node
• Significant in-the-field experience with shipping implementations, such as that provided by two independent VMs
• committee approval
• ecma262 PR approved

test262.fyi

Stage 3

• committee approval
• merge test262 tests

Stage 2.7

• committee approval
• spec editor signoff
  • @bakkot
• spec reviewer signoff
  • @ljharb
  • @jasnell

Stage 2

• committee approval
• spec reviewers selected
  • @ljharb
  • @jasnell
• spec text written

Stage 1

• committee approval

[ljharb]

LGTM! Just some minor notes:

• CreateKeyedPromiseCombinatorResultObject presumably shouldn't go under "properties of the promise constructor" - perhaps it should go in https://tc39.es/ecma262/#sec-promise-abstract-operations ?
• allSettledKeyed is a terrible name, but we do need the variant and i understand the convention being applied
• PerformPromiseAllKeyed and PerformPromiseAllSettledKeyed are identical except for 6.b. Is there a refactoring that could be applied to avoid duplicating the other 8ish steps?

[ajvincent]

Congratulations on stage 2.7. I wish I could participate more, but it seems to be moving along well now.

[acutmore]

Thanks @ajvincent! Still 1.3 to go 😃

[linusg]

Implementations are starting to appear:

• https://phabricator.services.mozilla.com/D284151
• feat(promise): implement Promise.allKeyed and Promise.allSettledKeyed boa-dev/boa#4996

[linusg]

Deno implemented this in their standard library (though I'm not sure why given that availability of the Promise builtins is imminent):

• feat(async/unstable): add allKeyed and allSettledKeyed denoland/std#6959
• feat(async): stabilize allKeyed denoland/std#7058

[ctcpip]

all of the test262 PRs in the test plan are merged. can we check the box now or is something still outstanding?

edit: since that's a condition of stage 3, it seems safe to check it 😄

[ajvincent]

For tracking shipping this proposal in Mozilla: https://bugzilla.mozilla.org/show_bug.cgi?id=2057270

UPDATE: this ticket now says it will ship with Mozilla Firefox 155, with a projected release date of September 1, 2026.

[ctcpip]

ECMA-262 PR

[ctcpip]

Chromium: https://issues.chromium.org/issues/540020490

[ajvincent]

FWIW, I implemented this in engine262 a couple months ago, and I'm waiting for a large check-in there before I bring in the tests, update to the spec, and bring it out from behind the feature flag.

Granted, engine262 isn't widespread, but it's something.

[linusg]

Implemented in Kiesel: https://codeberg.org/kiesel-js/kiesel/pulls/224

[ajvincent]

TypeScript 7.1: microsoft/TypeScript#63701

Looks like this is not a polyfill, though, just support in the built-in types.

[ajvincent]

Pull request for MDN documentation: mdn/content#45203