Reviewers for Stage 3

[leobalter]

Tracking issue for spec reviews for advancement to Stage 3.

From notes: JHD, RGN, BSH, DE MAH

• Richard Gibson @gibson042
• Bradford Smith @brad4d
• Mathieu Hofman @mhofman
• Jordan Harband @ljharb
• (Editor) Michael Ficarra @michaelficarra
• (Editor) Kevin Gibbons @bakkot
• (Editor) Shu-yu Guo @syg

[ljharb]

I'm still not a fan of the term "identity" here - I believe tc39/ecma262#2405 specifically intends to only use that term with spec values and not ecmascript values.

Otherwise, the only nit I have is that HasIdentity could say "is Object or Symbol," and be reduced to two lines (none of these should be stage 3 blockers, but I'd strongly advocate for changing the word to something less confusing before merging into the main spec)

[littledan]

LGTM for Stage 3. I think we can iterate on the names of abstract algorithms after Stage 3.

[brad4d]

LGTM for Stage 3 modulo concerns from @ljharb
Thanks for pointing those out BTW.

[leobalter]

Thanks! I have no strong preference over the abstraction name. HasIdentity was called this way to reflect general functionality but I also don't intend to expand this proposal to other types.

One review item I received in private:

we can't have this be a slippery slope and add numbers and booleans to weak collections

I don't have intention or goals to add other primitive values to weak collections and this remains out of scope for this proposal. Someone else might wanna discuss about Records and Tuples but it's not the case for this proposal.

[leobalter]

Thanks! @ljharb @brad4d #18 should address most of it. I'm now using the most simple name for the abstraction and I hope we are good to go with it.

[acutmore]

Almost 1 year later, we are once again in a position where we require spec review as we prepare to go for stage 3.

As per #19 the full set of proposed changes is being done as a PR due to the proposal's need for many small changes (e.g. updating liveness to reference both objects and symbols).

acutmore/ecma262#1

The key difference from last year is that 'registered symbols' (those created with Symbol.for) are not considered suitable weak keys/entries/targets.

cc: @mhofman @gibson042 @brad4d @ljharb @michaelficarra @bakkot @syg