2023-05-02
[20:04:56.0373] <shu>
 * harsh isn't the right word, i mean curt

[14:21:13.0195] <Michael Ficarra>
the OAuth spec runs ads (through https://www.ethicalads.io/)

[14:21:34.0066] <Michael Ficarra>
maybe if we ran ads, we could afford basic committee necessities without begging Ecma

[14:22:35.0113] <shu>
we can also, you know, find other IPR umbrellas outside of Ecma

[14:23:18.0428] <Michael Ficarra>
also we should add a view counter to the draft spec to compare against Istvan's download numbers that he always touts

[14:23:23.0582] <Michael Ficarra>
I am serious about that one

[14:24:30.0379] <shu>
ooo

[14:24:36.0628] <shu>
please let's do a 90s-style visitor counter

[14:25:57.0149] <Michael Ficarra>
I was thinking it would be invisible, but sure

[14:26:43.0832] <shu>
welcome to our ecmascript specification Homepage!!

[14:26:48.0153] <jmdyck>
couldn't you just scrape server logs?

[14:30:31.0463] <Michael Ficarra>
it's hosted on github pages

[14:30:36.0816] <Michael Ficarra>
we don't control that server

[14:31:24.0214] <jmdyck>
ah

[14:32:04.0422] <shu>
GH pages _do_ have analytics though don't they?

[14:33:01.0244] <shu>
seems like there's only a sliding window of 2 weeks https://github.com/tc39/ecma262/graphs/traffic

[14:33:01.0834] <Chris de Almeida (OoO until 05-14)>
sadly not

[14:33:18.0356] <shu>
actually i don't know what Visitors means there

[14:33:23.0023] <shu>
i guess it means repo visitors not GH pages visitors

[14:33:28.0315] <Chris de Almeida (OoO until 05-14)>
💯 

[14:33:55.0074] <Chris de Almeida (OoO until 05-14)>
there's a GH issue about this in the 262 repo 

[14:35:30.0105] <Chris de Almeida (OoO until 05-14)>
https://github.com/tc39/ecma262/issues/1167

[14:37:59.0965] <Michael Ficarra>
wow, daily repo views in the hundreds?

[14:38:04.0597] <Michael Ficarra>
that is not what I would've expected

[14:38:54.0373] <shu>
be hitting that F5 hoping for new features

[14:39:23.0483] <shu>
wait, is F5 Networks named after refreshing pages?

[14:43:38.0660] <Michael Ficarra>
no, it is named after the Fujita scale

[14:45:45.0803] <Michael Ficarra>
"move fast and break things" before it was cool

[14:48:05.0562] <shu>
interesting


2023-05-03
[12:08:42.0424] <shu>
hm what is this thing: Error: getaddrinfo ENOTFOUND ci.tc39.es

from https://github.com/tc39/ecma262/actions/runs/4875296114/jobs/8697312438

[12:11:00.0228] <jmdyck>
ci.tc39.es is an unknown hostname?

[12:39:02.0885] <bakkot>
yeah looks like all preview builds are down on all PRs

[12:39:17.0819] <bakkot>
either the DNS settings got broken or the host got broken

[12:40:41.0577] <jmdyck>
If it were just the host, that wouldn't make DNS forget the domain, would it?

[12:43:57.0259] <bakkot>
yeah I suppose

[16:03:01.0380] <ljharb>
oh shit, it was the dns yeah

[16:03:03.0709] <ljharb>
i'll fix that


2023-05-04
[14:51:27.0362] <shu>
Michael Ficarra: there are test262 iterator helper tests testing Iterator.prototype[@@iterator]

[14:51:32.0499] <shu>
where is that specified in the draft?

[14:51:34.0071] <shu>
i don't see it

[14:55:05.0878] <shu>
oh is this supposed to be %IteratorPrototype%[@@iterator]?

[15:06:25.0135] <bakkot>
yeah those are two ways of writing the same thing

[15:10:50.0663] <shu>
yeah okay i think these tests are wrong

[15:11:21.0136] <shu>
%IteratorPrototype%[@@iterator] always returns `this`, and there are tests that assume Iterator.prototype[Symbol.iterator] returns a function

[15:11:30.0410] <shu>
(i guess it's assuming it returns `Iterator`)

[15:17:28.0858] <Michael Ficarra>
ugh, those are left over from when I took over the PR from Rick

[15:17:35.0155] <Michael Ficarra>
I will look into them later

[15:17:48.0333] <Michael Ficarra>
in hindsight, I really should have started from scratch

[15:18:05.0247] <Michael Ficarra>
there's essentially nothing that's going to be left over from the original PR

[15:18:17.0133] <Michael Ficarra>
basically just the property descriptor tests


2023-05-08
[15:21:04.0766] <shu>
Michael Ficarra: FYI: the biggest thing the V8 implementation got wrong was not tracking if an iterator helper was completed. since that's one of the things implicitly done by generators, and we implemented them as direct iterators

[16:22:00.0570] <Michael Ficarra>
shu: did tests help catch it?


2023-05-09
[06:47:47.0258] <shu>
Michael Ficarra: indeed they did! thank you

[08:28:55.0655] <Michael Ficarra>
that's great!

[08:29:16.0950] <Michael Ficarra>
there's not much left to do, I could finish them this week, but probably won't since we have a couple stage3/4 reviews to do

[09:53:09.0183] <Rob Palmer>
Hello Editors!

[09:57:11.0234] <Rob Palmer>
I forwarded you a request from Istvan to get the formal TC39 vote on ES2023 onto the Reflector and agenda for next week's meeting.  This is for Ecma rule compliance.

Would someone like to take that?

(I need to catch a train now.  If no one picks that up I will create it in 4 hours time.)

[12:00:10.0847] <Michael Ficarra>
Rob Palmer: I feel that's more of a chair responsibility. If you are able, please create the notices.


2023-05-12
[17:25:45.0874] <Michael Ficarra>
shu: it was a pleasure reviewing your waitAsync PR

[17:26:17.0185] <Michael Ficarra>
it's really nice when the PR author already knows our conventions, at both the micro and macro levels

[07:51:56.0870] <shu>
i give at least 50% of the credit to jmdyck

[08:53:03.0092] <jmdyck>
aw shucks


2023-05-17
[08:53:45.0839] <Michael Ficarra>
we really need to do something about the `NewTarget` special form

[08:54:35.0503] <Michael Ficarra>
I just want to get rid of it entirely and use `GetNewTarget()` and `lexicalEnvironmentRecord.[[NewTarget]]`

[08:57:29.0670] <bakkot>
nah it's fine

[08:57:37.0084] <bakkot>
it's no worse than "the this value"

[09:03:12.0932] <Michael Ficarra>
there's no need for the special form though

[09:03:26.0698] <Michael Ficarra>
we already have (and use) `GetNewTarget()`

[09:04:08.0684] <Michael Ficarra>
we just use hand-wavey prose to set the `[[NewTarget]]` slot, and it makes it sound like they're two distinct concepts

[09:05:18.0770] <bakkot>
hm

[09:06:10.0511] <bakkot>
I feel like it's at least somewhat reasonable to write the "I am using new.target in a builtin" and "I am defining the semantics of new.target" forms in a different way

[09:07:04.0184] <Michael Ficarra>
> When invoked with [[Construct]], the this value is uninitialized, the [[Construct]] argumentsList provides the named parameters, and the [[Construct]] newTarget parameter provides the NewTarget value.

[09:07:26.0517] <Michael Ficarra>
you can't tell me that that's a good way to say `Set lexicalEnvironmentRecord.[[NewTarget]] to _newTarget_`

[09:07:56.0775] <Michael Ficarra>
same here:

[09:07:57.0618] <Michael Ficarra>
> The this value is uninitialized, argumentsList provides the named parameters, and newTarget provides the NewTarget value.

[09:12:17.0274] <Michael Ficarra>
 * you can't tell me that that's a good way to say `Set _lexicalEnvironmentRecord_.[[NewTarget]] to _newTarget_.`

[09:13:54.0532] <bakkot>
> <@michaelficarra:matrix.org> you can't tell me that that's a good way to say `Set _lexicalEnvironmentRecord_.[[NewTarget]] to _newTarget_.`

that's because it's not doing that

[09:14:07.0719] <bakkot>
the [[NewTarget]] slot is only used for user code

[09:14:40.0438] <Michael Ficarra>
so they *are* two separate concepts then?

[09:14:59.0716] <bakkot>
they could be collapsed

[09:15:12.0011] <bakkot>
but we write them in different ways

[09:16:01.0983] <bakkot>
mostly, afaict, because we don't set up environment records for built-in functions

[09:16:05.0468] <bakkot>
only for user functions

[09:16:08.0994] <bakkot>
 * only for user code

[09:16:43.0090] <bakkot>
so it is not trivial to just re-use the ambient lexical environment

[14:31:04.0185] <jmdyck>
no editor call this week, I take it

[14:57:29.0755] <ljharb>
nope, generally not during plenary week


2023-05-25
[07:50:29.0182] <Michael Ficarra>
do we want to lint for places where the chain production could be used but isn't?

[07:50:38.0587] <Michael Ficarra>
 * do we want to lint for places where the chain rule could be used but isn't?

[08:32:48.0681] <shu>
eh, i don't really like the chain rule

[08:33:08.0433] <shu>
if we had that machinery i'd rather insert some kind of default-collapsed box that has the auto-generated chain rule

[08:33:22.0270] <shu>
 * if we had that machinery i'd rather insert some kind of default-collapsed box that has an explicit, auto-generated chain rule

[14:30:20.0558] <bakkot>
ljharb: not a big deal, but https://github.com/tc39/ecma262/pull/2613 should have been one commit, not 7

[15:09:42.0660] <ljharb>
ah, did someone say that or did i miss it?

[15:09:54.0715] <ljharb>
they all kind of looked atomic so i didn't think to check

[15:12:01.0944] <bakkot>
no one said it but it's just a single edit

[15:12:37.0820] <bakkot>
like there's a "merge paragraphs" commit and then later a "split them back again" commit

[15:12:43.0046] <bakkot>
didn't need both of those

[15:13:00.0165] <bakkot>
default should be to land things as single commits most of the time

[15:23:22.0690] <Michael Ficarra>
almost all of the time

[15:50:58.0353] <ljharb>
fair enough


2023-05-30
[10:35:34.0145] <shu>
i may miss tomorrow's meeting

[10:35:37.0797] <shu>
(driving back from CO)


2023-05-31
[09:11:34.0252] <ljharb>
Michael (and shu after you’re back), if y’all have any time to review https://github.com/tc39/proposal-array-grouping/issues/54 before next plenary that’d be appreciated. there’s plenty of time til then, but I’m traveling between, so if there’s issues i want to make sure i have time to fix them. Thanks!

[09:41:17.0262] <shu>
ljharb: yep, i'm actually converting to the static method version and reviewing it as i go along

[09:41:33.0266] <ljharb>
beautiful, thanks

[13:44:32.0887] <Michael Ficarra>
only thing I see is "c. Append entry as the last element of map.[[MapData]]." should just be "Append ... to ..."

[13:46:58.0281] <Michael Ficarra>
also, I've really warmed up to the name "assort" for this and wish we would've tried that

[13:47:38.0695] <Michael Ficarra>
same for "3. Append group as the last element of groups."

[13:49:04.0165] <Michael Ficarra>
 * same for "3. Append group as the last element of groups." and "Append value as the last element of g.[[Elements]]."

[15:38:50.0179] <Michael Ficarra>
ljharb: JSON is a recursive data structure. `stringify` is a catamorphism; it "folds" the recursive structure into a flat summary (a string). `parse` is an anamorphism; it "unfolds" a summary string into a recursive structure.

[15:40:17.0737] <Michael Ficarra>
 * ljharb: the JSON-safe subset of ES values (objects, arrays, numbers, etc) are a recursive data structure. `JSON.stringify` is a catamorphism; it "folds" the recursive structure into a flat summary (a string). `JSON.parse` is an anamorphism; it "unfolds" a summary string into a recursive structure.

[16:31:38.0873] <shu>
don't you put that evil on me

[16:31:43.0844] <shu>
parse and stringify are better words