littledan: afaik node already decided to reflag

interesting, I guess they don't need a signal from us?

they reflagged because they had no signal, and the future of the syntax was uncertain

and blocking that from going public is something i am against because i have reasons for my opinions even if tc39 disagrees

Are there any annex features that don't run on all major browsers?

This would not be annex b -- annex b must be implemented by all browsers

ljharb: I do not think we want implementations to think of themselves as being allowed to go nuts with their own syntax, regardless of whether the spec says they are allowed to

it is fine if committee wants the lesson V8 to learn from this is "conformance as a proxy measure for shipping reality becomes even more tenuous"

odd sidebar, since I'm curious: is it a possibility to eventually move Date to this "legacy" section?

wonder the real world use rate in Web of Import assertions

wonder the real world use rate in Web of Import assertions

I think yulia might have some data there?

node's ability to ever ship breaking changes is great, I love that; I also appreciate that it happens quite rarely these days

ljharb: I would prefer the committee not give too much editorial direction here. I want to retain discretion to best communicate what we understand the committee's opinion to be here.

Annex B is evidence that we don't have spec purity. I'd rather that we don't decrease purity.

ljharb: Where would I find the draft standard for patent opt out review?

yeah, i think tip-of-tree right now will get tagged

Willian Martins your name isn't in the delegates list for the notes — is your acronym WMS? https://github.com/tc39/notes/blob/main/delegates.txt

anyone know why IEEE decimal distinguishes between 1.2 and 1.20?

actually it appears to be using document.lastModified

i will be unequivocal about one thing, which is what i was planning to bring in the queue: V8 is not going to be okay with a new value type

well, waldemar brought up === overloading

assuming an object w/o overloading is being proposed, what are delegate opinions as to why that would not be acceptable

the timebox was just not big enough, we had a ton more to talk about wrt decimal

I think one of Mark's concerns about string append is character classes. If Regexp.escape is safe for new RegExp("[" + Regexp.escape(input) + "]"), is that still a concern?

how it could be safe?

anyone know why IEEE decimal distinguishes between 1.2 and 1.20?

You can't match a ` in a TTL built RegExp

I don't understand why re` or `re`\\ wouldn't work for this

re`\\\`

is it actually Clean?

/\/ doesn't work either though

re`\` is invalid syntax

And re`\\` only matches s = "\\\\"

shu: ever heard of gradual typing? checkmate

No, seriously, given the diversity of meanings of gradual typing (where people unironically use the above form), we deliberately omitted it from this presentation

people actually do this? what?

is there seriously no [NLTH] in the proposed grammar?

I don't understand DE's claim that all the things we've been adding to the language since ES6 are "unifying" the language.

is there seriously no [NLTH] in the proposed grammar?

littledan, danielrosenwasser : Is the big-picture vision that you talked about presented in text anywhere? I think that's the sort of motivation and explanation that I'm most missing here, because I don't see how type annotations can get us from here to there.

littledan, danielrosenwasser : Is the big-picture vision that you talked about presented in text anywhere? I think that's the sort of motivation and explanation that I'm most missing here, because I don't see how type annotations can get us from here to there.

I think the framing is partially about transitions, it's about "can we find tasteful alternatives to express the same concepts"

I would rather see syntax here... like:

await const {
  shape = getShape(),
  color = getColor(),
  mass = getMass(),
};

it's neat but doesn't translate naturally to JS

rbuckton: i don't define non-divergence by "most code don't diverge", but by "no code diverges"

i'm sure we could but that sounds like a different proposal maybe?

Sorry, Shu, I'm trying to understand, what are you getting at? E.g., for the <T>value syntax, what harm does it do if TS has a mode that accepts that, and JS never will, but it might support a large other subset of TS/Flow?

is it the same problem?

I think an all analogue is fine, the others seem to only make sense on unnamed collections

i understand that the whole point of this proposal is to bring TS syntax "into the fold". the harm is that if the promise is actually "we will work to converge TS and JS syntax in the future", that sounds like ecosystem divergence to me and i do not have the risk appetite for that

and i am saying, "for the most part" isn't enough for my risk appetite