Sept 2017

[hzoo]

Bocoup, Boston, MA (Sept 26-28)
https://github.com/tc39/agendas/blob/master/2017/09.md

Checkbox means we completed it (some Stage 1 we may want to wait to start given instability of the proposal)
https://tc39.github.io/process-document/ for a primer what each Stage means

New Proposals

• ✍️ Pipeline Operator Stage 1 (next: need a solution for method extraction, possibly just ::a)
• Extensible Numeric Literals Stage 1 (also had objections, will probably wait)
• First Class Protocols Stage 1
• Subsume JSON to Stage 1
• Partial application to Stage 1 (but has strong objections on syntax/semantics, will probably wait)
• Nullary Coalescing to Stage 1 (need explainer for checking "nully" or "undefined"

Moving Foward

• ✍️ Throw Expressions from Stage 1 to Stage 2
• ✍️ Private Methods (and accessors) from Stage 2 to Stage 3
• ✍️ import.meta from Stage 2 to Stage 3

Needs Consensus (Stage 4)

• Early errors for RegExp literals
• ✍️ export-ns-from from Stage 2 (needs tests)

Needs Spec Work

• Optional Chaining stays at Stage 1 again (issues with short-circuiting only for ?. and use of ?.[ and ?.( in general, ?. syntax is desired but may not align with ?.[ and ?.( but ?&. and other syntax changes are not preferred)
• Object Shorthand Improvements stays at Stage 0
• Builtins.typeOf() and Builtins.is() stays at Stage 0
• Numeric Separator stays at Stage 2

Other (Builtins, etc)

• Template literal memory leak fix (needs consensus PR)
• Intl.NumberFormat.prototype.formatToParts to Stage 4
• Intl.Segmenter to Stage 3
• Intl.RelativeTimeFormat to Stage 3
• Atomics.waitAsync to Stage 2
• Intl.PluralRules to Stage 4
• Array.prototype.flatMap and Array.prototype.flatten from Stage 1 to Stage 2
• Intl.Locale to Stage 1
• String.prototype.matchAll to Stage 2
• Needs consensus: Make LocalTZA take 't' and 'isUTC' and drop DSTA(t). tc39/ecma262#778, Normative: Create global properties for sloppy function hosting as vars tc39/ecma262#888, Access the .next method once, at the beginning, of the iteration protocol tc39/ecma262#988

[jridgewell]

Pipeline operator to Stage 1: https://github.com/tc39/proposal-pipeline-operator

a |> b |> c

// - - -
c(b(a))

https://twitter.com/left_pad/status/912704437893976066

[hzoo]

Extensible literals to Stage 1: https://github.com/littledan/proposal-extensible-numeric-literals

1236536253453n		BigInt (special case)
4525i			Imaginary numbers
235435.461m		IEEE 754-2008 64-bit decimal
300px			CSS Typed OM

[mathiasbynens]

• Intl.NumberFormat.prototype.formatToParts to Stage 4
• Intl.Segmenter to Stage 3

[jridgewell]

@mathiasbynens: I don't think there's anything for us to implement with those. 😉

[mathiasbynens]

@jridgewell Given the feedback during the meeting, I’d say the same goes for the extensible literals proposal and (to a lesser extent) the pipeline operator.

[hzoo]

Could at least get the syntax in for pipeline, would be interesting trying to do multiple transforms via options (and could test in the REPL for people, assuming we get those options in)

[jridgewell]

What I mean is the Intl changes are purely runtime. We'd need to implement the parsing and syntax for literals and pipeline.

[jridgewell]

First Class Protocols to stage 1: https://github.com/michaelficarra/proposal-first-class-protocols

protocol ProtocolName {
  // declare a symbol which must be implemented
  thisMustBeImplemented;

  // and some methods that you get for free by implementing this protocol
  youGetThisMethodForFree(...parameters) {
      methodBody;
  }
}

class ClassName implements ProtocolName {
  [ProtocolName.thisMustBeImplemented]() {
    // this is the implementation for this class
  }
}

let instance = new ClassName;
instance[ProtocolName.youGetThisMethodForFree]();

[michaelficarra]

@jridgewell Sweet.js implementation with pretty comprehensive test suite located here: https://github.com/disnet/sweet-interfaces

Polyfill implementing the Protocol constructor here: https://github.com/michaelficarra/proposal-first-class-protocols-polyfill. It's correct, but will only work in newer engines and sticks itself on the global for you. PRs to make it good would be very welcome.

[hzoo]

tc39/ecma262#984

Early errors for RegExp literals: /a{5,3}/, /[z-a]/

Would this have been covered with test262 tests anyway?

[hzoo]

export-ns-from: export * as ns from 'mod' - #16

to Stage 4 (needs consensus)

[jridgewell]

Subsume JSON to stage 1:

http://babeljs.io/repl#?babili=false&browsers=ie%2010&build=&builtIns=true&code_lz=ESCgJgoZEoCYg&debug=false&circleciRepo=&evaluate=true&lineWrap=true&presets=es2015-loose%2Creact%2Cstage-3&prettier=false&targets=&version=7.0.0-alpha.19

[jridgewell]

Throw expressions to stage 2: https://github.com/tc39/proposal-throw-expressions

a || throw new Error('test');

[jridgewell]

Optional Chaining remains at Stage 1 😢.

Still need to update to remove super?.prop and a?.b = 1.

[jridgewell]

Private methods and accessors to stage 3: https://github.com/tc39/proposal-private-methods
Note: static privates footgun

https://twitter.com/littledan/status/913224214760566784

[hzoo]

Nullary Coalescing from Stage 0 to Stage 1 #14

a ?? b

// concerns about cross-cutting concerns with ?. syntax

[jridgewell]

Partial application to Stage 1: https://github.com/rbuckton/proposal-partial-application

[michaelficarra]

@jridgewell With strong objections to the syntax that is currently associated with the proposal.

[ljharb]

For partial application, please note that there is a not-insignificant number of members that believe that the proposal will not be able to be made viable.

[hzoo]

import.meta to Stage 3

[littledan]

Note that extensible numeric literals had significant objections to syntax and semantics as well.

[mgtitimoli]

Hi everyone, could you please elaborate a bit more what is happening with partial application proposal?

I would love to see it moving forward but based on your comments, I guess it is not gonna be an easy process.

Thanks.

[hzoo]

You can check the individual issues @mgtitimoli #32 - and the notes will be posted in a few weeks

[brettz9]

Array.prototype.flatten has been changed to Array.prototype.flat per https://github.com/tc39/proposal-flatMap and is now a stage 3 per https://github.com/tc39/proposals .