2023-12-01
[16:14:09.0970] <akaster>
> <@annevk:matrix.org> akaster: https://docs.google.com/presentation/d/1hi4gH7pJPHsg_hnIj77XN_ce54HIaNUnBLenVwohFVo/edit#slide=id.g5641ecbac9_0_131 might help for the high-level idea.

To supplement this I've started writing down my understanding of how this is all supposed to fit together and how we're going to map the spec to Ladybird here, beginning with the JS concepts: https://github.com/SerenityOS/serenity/pull/22107/files

[19:34:54.0364] <ntim>
zcorpan: https://github.com/web-platform-tests/wpt/pull/43463/ 

[19:36:04.0366] <ntim>
Note that Firefox wrongly supports unordered list styles on ordered lists and vice-versa

[21:44:47.0778] <Domenic>
The history of people trying to spec ol/ul/li styling using CSS is very fraught... I think Firefox got furthest but IIRC it still had lots of edge cases. Things like start and reverse really stress the system.

[22:29:24.0768] <annevk>
akaster: looks pretty solid, the one piece I'm missing is an explicit callout that the reason we have agents (and not just realms and realm clusters) is because of deficiencies in early web platform design where Window objects could manipulate each other directly through various <iframe> and popup APIs

[22:30:00.0312] <annevk>
akaster: although maybe you're planning on covering that in the TODO section, heh

[09:14:11.0645] <smaug>
jarhar: are the current ideas for https://github.com/whatwg/html/issues/9799 written down somewhere, or perhaps some simple examples of what it would look like?

[09:14:23.0325] <smaug>
I definitely do like the idea of reusing existing elements

[10:40:58.0673] <Eric Portis (he/him)>
zcorpan: I'm trying to write about sizes=auto, and confusing myself. If you hadn't tried to do anything with "concrete size ignoring natural dimensions" and had just used current clientWidth, would it have been racy/cyclical? It feels like "obviously yes" but in the cases I'm working through, the intrinsic dimensions stay locked to whatever the initial extrinsic dimensions were, because of the density-correction mechanism.

[10:42:13.0876] <Eric Portis (he/him)>
 * zcorpan: I'm trying to write about sizes=auto, and confusing myself. If you hadn't tried to do anything with "concrete size ignoring natural dimensions" and had just used current clientWidth, would it have been racy/cyclical? It feels like "obviously yes" but in the cases I'm working through, the intrinsic dimensions stay locked to whatever the initial extrinsic dimensions were (or 0x0 if there were no extrinsic dimensions), because of the density-correction mechanism.

[12:30:53.0501] <Dominic Farolino>
The link element's disabled IDL attribute reflects the content attribute of the same name, whereas the style's disabled IDL attribute has no such content attribute. What principles in general decide whether we should reflect or not?

[12:32:35.0535] <akaster>
I actually didn't know that was the reason why, but it definitely makes sense. Agents feel like a bit of a strange abstraction over realms 

[12:44:48.0101] <Eric Portis (he/him)>
@zcorpan: https://gist.github.com/eeeps/a469258ba387d3c8300f12248ed52032

[13:07:25.0588] <Eric Portis (he/him)>
 * zcorpan: https://gist.github.com/eeeps/a469258ba387d3c8300f12248ed52032


2023-12-02
[22:53:49.0788] <annevk>
akaster: yeah, they only exist because of that and in new "agents" such as workers agents and realms are 1:1

[22:54:26.0620] <annevk>
akaster: having said that, I just remembered TC39 is working on ShadowRealm, which I guess brings the property of 1:N to all agents


2023-12-03
[16:57:57.0565] <Domenic>
> <@domfarolino:matrix.org> The link element's disabled IDL attribute reflects the content attribute of the same name, whereas the style's disabled IDL attribute has no such content attribute. What principles in general decide whether we should reflect or not?

Every content attribute needs to have a reflecting IDL attribute. The other way around is not relevant; there are lots of APIs which are not related to the DOM.

[16:58:16.0480] <Domenic>
The inconsistency between link and style is unfortunate, but the whole disabled feature in general has never felt very well thought out to me.

[17:21:13.0454] <emilio>
Yeah HTMLLinkElement.disabled was a compat requirement (an IEism iirc?)

[17:22:02.0379] <emilio>
https://github.com/whatwg/html/issues/3840 has a lot of that background

[17:22:45.0559] <emilio>
Firefox was consistent before ^ ftr

[05:40:25.0458] <Dominic Farolino>
> there are lots of APIs which are not related to the DOM.

So "related to the DOM" means having a content attribute then?

[05:41:46.0429] <Dominic Farolino>
If someone introduces an IDL attribute, but also wants to trigger the underlying behavior without script, then they'd have to have it reflect a content attribute with the same name. Is that a good enough "justification" / reason for doing that.

[05:42:17.0394] <Dominic Farolino>
It seems like "wanting to do things that you can do with an IDL attribute, but without JS" is the big motivation for wanting to reflect an IDL attribute?


2023-12-04
[19:36:04.0271] <Domenic>
> <@domfarolino:matrix.org> > there are lots of APIs which are not related to the DOM.
> 
> So "related to the DOM" means having a content attribute then?

By "related to the DOM" I mean "related to the structure of nodes and attributes, i.e. the DOM tree"

[19:36:46.0580] <Domenic>
> <@domfarolino:matrix.org> If someone introduces an IDL attribute, but also wants to trigger the underlying behavior without script, then they'd have to have it reflect a content attribute with the same name. Is that a good enough "justification" / reason for doing that.

I would work the other way around. Content attributes first; if they exist you automatically get an IDL attribute. If you don't need a content attribute then you are just doing normal JS API design.

[19:37:34.0515] <Domenic>
> <@domfarolino:matrix.org> It seems like "wanting to do things that you can do with an IDL attribute, but without JS" is the big motivation for wanting to reflect an IDL attribute?

No, the motivation is wanting to do things in JS, slightly easier and more idiomatically. Without IDL attribute reflection: `el.setAttribute("foo", "")`. With IDL attribute reflection: `el.foo = true`.

[02:34:51.0212] <zcorpan>
Thanks for the heads-up!

[02:41:00.0360] <zcorpan>
Eric Portis (he/him): I think the main issue is when the w and the image's actual width don't match. Let's say in your example "a" actual width is 4 pixels, then the layout width becomes 2px after loading the image, which would cause the next image to load, etc.

[02:43:15.0741] <zcorpan>
Eric Portis (he/him): a proposed alternative fix iirc was to stop the "loop" after the second load, but based on prior feedback about double downloads I think it's not acceptable to have double downloads even if it only happens sometimes

[02:46:35.0914] <zcorpan>
Eric Portis (he/him): the other reason is to make it look wrong when not using width/height attributes (or have dimensions with css). Otherwise people might omit dimensions and we would likely double-download in such cases without size containment

[02:48:06.0936] <zcorpan>
Eric Portis (he/him): I suspect size containment for images will be more annoying, and that's unfortunate but I hope acceptable!

[06:20:41.0214] <Dominic Farolino>
> If you don't need a content attribute then you are just doing normal JS API design.

This is the part I'm curious about. I just don't really know when you would "need" a content attribute vs just stick with JS API design on the element's IDL interface. If someone says "I want to introduce a new 'attribute' on this element", then they've two options: (1) content attribute with automatic IDL reflection, vs (2) just IDL attribute, for a JS API. I'm not clear on how to provide guidance to select between those two. I would need to ask them a question relating to "If you need/don't need a content attribute" above. So I guess I'd ask "Can this just be a JS API on the element's interface, or do you need it to also be a content attribute". But to "need a content attribute" (i.e., an IDL attribute alone is insufficient) means your design desires to trigger the attribute's behavior without script... I think.. like just from HTML.

[06:58:19.0984] <annevk>
Dominic Farolino: I think usually it would become pretty clear, pretty fast

[07:27:26.0670] <Eric Portis (he/him)>
zcorpan: Thanks! The incorrect w attribute thing makes sense; looks like you suggested a fix for exactly that here https://github.com/whatwg/html/issues/9448#issuecomment-1611905790 (damn my memory)... but... seems like containment just felt cleaner from a spec/implementation perspective. An explicit severing of any possible way intrinsic size will or could ever influence layout size.

[07:27:51.0578] <jarhar>
the current idea as i would like it is to do whats in the explainer here but reuse the select element instead of creating a selectlist element: https://open-ui.org/components/selectlist/

[07:28:25.0152] <jarhar>
oh and it seems likely that we will reuse the datalist element instead of creating a listbox element

[07:28:50.0926] <jarhar>
so basically if you provide a <button> child to a select element, that will be slotted in to the default button and replace it

[07:29:06.0391] <jarhar>
same with the datalist element replacing the native/default listbox

[08:49:54.0517] <zcorpan>
Eric Portis (he/him): right. Especially the case where the image's final width depends on the image's intrinsic height seems tricky to make stable without size containment


2023-12-05
[16:47:26.0478] <Domenic>
> <@domfarolino:matrix.org> > If you don't need a content attribute then you are just doing normal JS API design.
> 
> This is the part I'm curious about. I just don't really know when you would "need" a content attribute vs just stick with JS API design on the element's IDL interface. If someone says "I want to introduce a new 'attribute' on this element", then they've two options: (1) content attribute with automatic IDL reflection, vs (2) just IDL attribute, for a JS API. I'm not clear on how to provide guidance to select between those two. I would need to ask them a question relating to "If you need/don't need a content attribute" above. So I guess I'd ask "Can this just be a JS API on the element's interface, or do you need it to also be a content attribute". But to "need a content attribute" (i.e., an IDL attribute alone is insufficient) means your design desires to trigger the attribute's behavior without script... I think.. like just from HTML.

Yes, your last sentence is correct.

[16:48:03.0304] <Domenic>
In general HTML elements are designed as HTML first; the fact that they have JS APIs is an interesting addition in some cases.

[16:48:25.0029] <Domenic>
But if you just want a JS API then you can just design a JS API. Maybe even not tied to a HTML element at all.

[19:56:41.0058] <sirisian>
Has anyone ever suggested allowing an element to exist in two places in the dom tree? Like a <slot> that can point to another node? I have vague memories of reading about this over a decade ago before web components. Seems more feasible now though.

[23:06:17.0132] <annevk>
sirisian: we could use it for `<select>`, but it's more of a CSS feature I'd think

[09:45:22.0351] <sirisian>
@annevk That's exactly what I've always needed it for. I have thousands of drop downs and code to deduplicates such picker options. My code would be very elegant if I could just point to their options and not move stuff around manually.

[14:51:46.0047] <sirisian>
> <@annevk:matrix.org> sirisian: we could use it for `<select>`, but it's more of a CSS feature I'd think

How did you envision those working? Thinking of trying to make a very basic proposal. "display: contents; contents: #foo;" I foresee two modes. #foo tag is slotted and another mode where #foo's children are slotted. So "contents: #foo" and "contents: #foo / children;" Not sure on the syntax.

[14:51:57.0667] <sirisian>
> <@annevk:matrix.org> sirisian: we could use it for `<select>`, but it's more of a CSS feature I'd think

 * How did you envision that working? Thinking of trying to make a very basic proposal. "display: contents; contents: #foo;" I foresee two modes. #foo tag is slotted and another mode where #foo's children are slotted. So "contents: #foo" and "contents: #foo / children;" Not sure on the syntax.

[15:34:50.0926] <emilio>
That raises a lot of questions about all sorts of APIs. Lots of things assume an element has one principal box. E.g. how does something like getBoundingClientRect() behave?

[15:35:29.0245] <emilio>
Not trying to discourage necessarily, just pointing out that it's probably a bigger challenge than it seems 


2023-12-06
[19:28:46.0238] <sirisian>
You're completely right. Writing up some examples and running into design issues, so I'll continue thinking about it.

[19:31:03.0248] <sirisian>
emilio: Oh your name sounded familiar. I was just reading your comment the other day when I made this issue: https://github.com/WICG/webcomponents/issues/1038

[22:55:37.0709] <sirisian>
https://github.com/sirisian/slottedelements I think this would work for me. Any less ad-hoc solutions?

[02:28:29.0516] <Ms2ger>
Looking at [delta traverse](https://html.spec.whatwg.org/multipage/nav-history-apis.html#delta-traverse) from [#6315](https://github.com/whatwg/html/pull/6315), if delta is zero, should we run both step 3 "reload" and step 4 "Traverse the history by a delta"? Before #6315, we seem to have had an early return there. CC Domenic  Jake Archibald 

[02:30:05.0255] <Domenic>
> <@ms2ger:igalia.com> Looking at [delta traverse](https://html.spec.whatwg.org/multipage/nav-history-apis.html#delta-traverse) from [#6315](https://github.com/whatwg/html/pull/6315), if delta is zero, should we run both step 3 "reload" and step 4 "Traverse the history by a delta"? Before #6315, we seem to have had an early return there. CC Domenic  Jake Archibald 

Definitely a bug, thanks. PR or issue appreciated.

[02:30:27.0717] <Ms2ger>
Will do, thx

[02:35:43.0856] <Ms2ger>
https://github.com/whatwg/html/pull/9980

[13:59:27.0295] <akaster>
In the transfer steps for message ports (https://html.spec.whatwg.org/multipage/web-messaging.html#message-ports:transfer-steps), is there any actual implementation steps intended for step 2 to transfer a task source? My interpretation at the moment is that all this is saying is that on the other end, when you reconstitute the message port, to make sure to set up a task source for it.

Or is it possible that there are pending messages at the time you transfer the message port, and those messages should be stashed and delivered to JavaScript on the other side? 

[14:00:00.0407] <akaster>
Does anyone actually write code like that? postMessage to a message port, and then transfer the message port, expecting the message to be received on the other end?


2023-12-07
[23:54:51.0869] <annevk>
akaster: I suspect what happens in practice is that folks forward the remaining tasks

[23:57:21.0931] <annevk>
akaster: although I suppose here a queue is really just a list underneath so you might as well copy the list of tasks all at once, hmm

[23:58:26.0584] <akaster>
In the case of a message port, would all the tasks just be serialized data ready to be deserialized in some arbitrary realm?

[23:58:53.0316] <annevk>
And this could happen when you have some kind of capability-based system as Ian envisioned when he wrote this up. You pass your message port out, but whoever receiving it might pass it along to someone else, but at the same time it starts getting incoming messages.

[23:59:17.0846] <annevk>
akaster: yeah.

[00:00:10.0837] <akaster>
Hmm ok. I guess we can make that work. Still no barriers to allowing process-barriers between workers and windows yet 😸

[00:19:00.0973] <sirisian>
https://github.com/WICG/proposals/issues/131 Updated with a real proposal. Anyone want to fundamentally change DOM?

[00:36:45.0143] <Kaiido>
annevk: I'm looking at adding the historical tests for `commit()` into WPT as you suggested, however I find [html/canvas](https://github.com/web-platform-tests/wpt/tree/master/html/canvas) quite messy there with a mix of manual and auto-generated tests and an *element*/*offscreen* segregation. So I'm not sure where it would be best to add this. Do you think directly in `html/canvas` is just-fine™?
For reference, another such test might be needed for the `scrollPathIntoView` method, removed in https://github.com/whatwg/html/pull/8229 which would concern only *element*, though that one never had any implementation.

[00:47:34.0012] <Kaiido>
For the record, I believe I might do this sometimes: I often use a MessageChannel as a way to start various Promise based communication channels with a single multi-task worker. The first one will thus start the worker, ask to transfer its MessagePort and send a first message into the port synchronously. But if I'm not mistaken, the transferring of the port will actually be done asynchronously, once the worker is ready, and thus after I sent the first message to the port.

[01:08:59.0017] <annevk>
Kaiido: yes html/canvas/historical.any.js (not sure what environments you need, adjust extension accordingly) would be just fine

[01:09:05.0357] <annevk>
(tm)

[01:10:15.0409] <annevk>
Kaiido: and I think adding `scrollPathIntoView` seems reasonable, though perhaps with a comment that it's prolly safe to reuse that name if we wanted to

[15:04:07.0763] <TabAtkins>
annevk: btw, tiny markup fix in <https://github.com/whatwg/url/pull/808> ahead of me cutting a new Bikeshed release that'll start flagging it as an error. 

[15:04:25.0667] <TabAtkins>
(There might be more similar ones coming from me in a few minutes, I'm running thru all the WHATWG specs since they're easy.)


2023-12-08
[01:18:28.0892] <keithamus>
annevk & Domenic I believe https://github.com/whatwg/html/pull/9965 and https://github.com/whatwg/html/pull/9963 are the last 2 PRs before we can close https://github.com/whatwg/html/issues/9832.

[01:22:53.0758] <Domenic>
Sorry, I've been slammed with end of year stuff, and am now taking a short vacation. Hopefully next week!

[03:05:47.0196] <annevk>
keithamus: merged the first, left a nit on the second

[03:05:57.0266] <keithamus>
Thank you kindly

[03:06:10.0600] <keithamus>
Enjoy the vacation Domenic!


2023-12-10
[12:21:41.0537] <Cayden Nguyen>
Hi!

[13:55:35.0302] <Cayden Nguyen>
Nice to meet you!


2023-12-11
[06:37:24.0804] <chargeitall>
When I click on the rd-token> or n-token> links, they do not anchor to their respective definitions. Here is the link to the webpage where this issue is occurring: https://www.w3.org/TR/css-syntax-3/#error-handling.

[07:30:46.0031] <chargeitall>
 * I am trying to get my head used to read dry technical documentation and got stuck here 👉When I click on the rd-token> or n-token> links, they do not anchor to their respective definitions. Here is the link to the webpage where this issue is occurring: https://www.w3.org/TR/css-syntax-3/#error-handling.

[09:11:06.0193] <chargeitall>
> <@chargeitall:matrix.org> I am trying to get my head used to read dry technical documentation and got stuck here 👉When I click on the rd-token> or n-token> links, they do not anchor to their respective definitions. Here is the link to the webpage where this issue is occurring: https://www.w3.org/TR/css-syntax-3/#error-handling.

I figured it out. the ADHD reading extension using bionic reading format omitted the characters  before n-token

[09:19:05.0706] <TabAtkins>
Oh, how strange. I wonder why it's cutting things off like that.

[09:31:22.0637] <chargeitall>
it could be that the extension was originally for chrome. I deleted it and replaced it with similar extension dedicated to edge as this is the browser I am using and this resolved the issue

[09:31:55.0726] <chargeitall>
 * it could be that the extension was originally for chrome. I deleted it and replaced it with similar extension dedicated to edge as this is the browser, I am using of which the later resolved the issue.


2023-12-12
[01:25:44.0895] <chargeitall>
does mdn web doc discord channel work on element?

[05:37:18.0141] <Ms2ger>
annevk: would you take a PR to add `<div algorithm>` to some or all algorithms in DOM (no promises when I might have time, but the lack of them means that variable highlighting doesn't work)

[08:13:09.0681] <annevk>
Ms2ger: yes, see Fetch for my preferred style

[11:53:46.0628] <akaster>
Is it correct to assume that the kind of  "requirement on callers" listed here is discouraged for future specs/spec updates? https://html.spec.whatwg.org/multipage/structured-data.html#performing-structured-clones-from-other-specifications

> Call sites that are not invoked as a result of author code synchronously calling into a user agent method must take care to properly prepare to run script and prepare to run a callback before invoking StructuredSerialize, StructuredSerializeForStorage, or StructuredSerializeWithTransfer abstract operations, if they are being performed on arbitrary objects.

[11:54:39.0220] <akaster>
And following that, is there any kind of tracking issue for places that "should be" updated to explicitly manipulate the surrounding agent's execution context stack?

[11:57:48.0102] <akaster>
As a side note, it seems that Domenic's navigation apis will do that sort of "deferred structured serialize" mentioned in the second example if a navigation history entry needs its history API state serialized/deserialized "in parallel" or on the traversal task source 🤔

[12:01:37.0991] <akaster>
Or maybe it was from a webidl callback context in the WaitForAll in navigation history event processing 🤔🤔. I'll have to revive my branch implementing that

[12:38:39.0646] <akaster>
I think we also ran into this kind of implicit execution context stack issue within the module loading implementation, though that was mostly Andreas this month

[12:41:34.0974] <TabAtkins>
Hm, why would you think it's discouraged? It's useful to know what implicit assumptions an algorithm makes, when relevant, which might affect how you can/should call it.

[12:45:40.0417] <akaster>
I would rather see the explicit steps on the actual callers as well, rather than make it implicit

[12:46:10.0115] <akaster>
When implementing algorithms that call that method, it's hard to keep track of all the context of which algorithms have special requirements 

[12:49:09.0883] <akaster>
In Ladybird at the moment, that manifests as a test failing with an assertion that the execution context stack is not empty, and we then have to go put a RAII helper into the part of the algorithm that needs one. Rather than seeing "prepare to run a script and a callback" as a spec step, we need to hit the code path from the event loop or as a result of the UI process poking the renderer and recognize the source of the assertion failure

[15:35:39.0082] <Domenic>
Those requirements on callers are not meant to be action at a distance. They're meant to be advice to future spec writers.

[15:36:42.0155] <Domenic>
Navigation API serialization/deserialization is always on the main thread.

[15:39:31.0602] <akaster>
Hmm okay, in that case I'll try to document all the "extra" prepare to run script/callback steps we've had to add and see if we're doing something wrong for them or if the steps should be added 

[15:41:01.0562] <akaster>
On the main thread and as a result of a JavaScript call aren't always the same thing though. Maybe we just always need to push a context when interacting from the UI process and it all goes away


2023-12-13
[17:33:41.0387] <Domenic>
To be more precise, I meant "as a result of either the author calling a Web IDL method or the browser invoking a Web IDL callback".


2023-12-14
[01:26:06.0866] <Ondřej Žára>
I am looking for a proper place to discuss various options that remain available in a post-3P-cookie world (related to sign-in management, bounce tracking, FedCM, ...). I have written a short summary of points that I would like to hear expert opinions on; where should I post that? Here? FedCM github issues? Some other github repo? I am not seeking clarification regarding one particular technology, but rather a more high-level overview of options and directions browser are heading to.

[02:28:54.0272] <annevk>
Ondřej Žára: some privacycg repo prolly

[02:43:12.0815] <Ondřej Žára>
> <@annevk:matrix.org> Ondřej Žára: some privacycg repo prolly

Thanks! Will have a look.

[03:16:03.0416] <annevk>
keithamus: I learned from Jeffrey Yasskin about https://docs.github.com/en/communities/moderating-comments-and-conversations/managing-reported-content-in-your-organizations-repository. Is there any way that you can configure that for an entire organization and manage for an entire organization?

[03:16:49.0508] <annevk>
keithamus: I was also thinking of making some small set of people moderators, but it seems moderators cannot be entrusted to delete issues so I'm not quite sure what the point would be

[03:18:00.0771] <keithamus>
I’ve been talking with the communities team about the WhatWG repos and getting better moderation tools. I can’t make any promises nor make any predictions on timelines but rest assured it’s being discussed! 

[03:18:28.0582] <annevk>
That's great, thanks for taking the initiative. Happy to leave things as-is

[03:18:33.0849] <annevk>
 * That's great, thanks for taking the initiative. Happy to leave things as-is for now.

[05:14:03.0633] <chargeitall>
how can I view table content on the left https://infra.spec.whatwg.org/?

[05:20:37.0695] <chargeitall>
 * how can I view table content on the left "sidebar" https://infra.spec.whatwg.org/?

[05:25:54.0495] <annevk>
Domenic: defending some 2015-decision making from you here, you might wanna chime in or at least be aware: https://github.com/tc39/proposal-arraybuffer-base64/pull/33

[05:35:52.0294] <chargeitall>
this extension solved it https://microsoftedge.microsoft.com/addons/detail/onetoc/jkgapfniamkoblbmbhdjlnfklihlpjmc

[05:42:30.0937] <chargeitall>
https://microsoftedge.microsoft.com/addons/detail/mindpane/ihpkadephhbmhgmlmbconphgneloaeic

[11:11:12.0442] <TabAtkins>
someone with mod in whatwg wanna close this? https://github.com/whatwg/meta/issues/273

[14:51:12.0052] <Melati Melati>
Selamat bergabung

[14:52:37.0576] <Melati Melati>
@melatimela.matrix


2023-12-15
[19:12:46.0959] <Domenic>
annevk: any chance we can snipe you into implementing and updating tests for https://github.com/whatwg/dom/pull/1079 ?

[23:27:46.0696] <chargeitall>
why the whatwg webpages do not have a sidebar index like the w3c or workgroups do?

[00:18:51.0147] <annevk>
Domenic: I think we might have tried to roll something like that out and got bitten by a router so I'd prefer a more coordinated rollout. If there's someone on the Chromium side who is willing it might be something we can try next year though.

[00:19:19.0375] <annevk>
See whatwg/whatwg.org for an issue on this. 

[00:19:35.0297] <Domenic>
I think that particular one was designed to only be backward-compatible? But who knows; it's fair to want something more coordinated.

[00:24:12.0916] <annevk>
It's prolly also very hot code making it more tricky to navigate, but it's worth looking into as the status quo sucks.

[00:42:00.0906] <chargeitall>
the link is invalid https://whatwg/whatwg.org

[00:59:42.0343] <chargeitall>
another pretty handy extension. I did not understand your proposal.

[01:12:05.0877] <chargeitall>
I could not find an issue on git.

[01:48:58.0932] <annevk>
I thought https://github.com/whatwg/whatwg.org/issues had an issue for this, but it seems not

[01:54:47.0245] <chargeitall>
I am intrigued as to why it was not brought up.

[02:41:44.0846] <annevk>
It has been brought up, but maybe just on Chat.

[05:51:14.0135] <Ms2ger>
Where does it say that you can't set `EventTarget.prototype.__proto__`? I don't think https://webidl.spec.whatwg.org/#platform-object-setprototypeof applies

[05:51:16.0753] <Ms2ger>
cc annevk 

[05:54:53.0642] <chargeitall>
can I create an issue or you would rather do it yourself.

[06:18:52.0225] <Sam Sneddon [:gsnedders]>
AFAICT, per WebIDL, it has the normal `[[SetPrototypeOf]]`?

[06:20:37.0966] <Ms2ger>
That would be fine by me if that's what browsers did :)

[06:20:57.0602] <Sam Sneddon [:gsnedders]>
I am making no judgement about what browsers do 🙃

[06:21:23.0028] <Andreu Botella>
I once, when working on something for Deno, saw a WPT test that checked that

[06:21:28.0086] <Andreu Botella>
let me find it

[06:21:37.0260] <Ms2ger>
Oh, I believe it's tested

[06:21:49.0225] <Ms2ger>
I'm just not sure about the spec part

[06:22:28.0747] <Ms2ger>
https://github.com/web-platform-tests/wpt/blob/master/webidl/ecmascript-binding/global-immutable-prototype.any.js

[06:22:52.0975] <Ms2ger>
(I would know, I wrote it)

[06:23:18.0566] <Sam Sneddon [:gsnedders]>
yeah I was about to say, "didn't you write those tests"

[06:23:44.0693] <Andreu Botella>
I'm pretty sure I found some spec text for it

[06:23:50.0565] <Andreu Botella>
but I don't remember where

[06:24:10.0471] <Ms2ger>
Ah, https://webidl.spec.whatwg.org/#interface-prototype-object step 8

[06:24:54.0832] <Ms2ger>
Thanking my 2016 self for having the foresight to link to the spec changes 😅

[06:25:52.0537] <Ms2ger>
And thanks Andreu Botella  for making me look for them :)

[06:26:01.0067] <Sam Sneddon [:gsnedders]>
i.e., there exists some interface with `[Global]` that inherits from `EventTarget`?

[06:26:21.0517] <Andreu Botella>
`Window` does

[06:26:27.0872] <Ms2ger>
All of them do

[06:26:28.0362] <Andreu Botella>
and `WorkerGlobalScope`

[06:50:40.0642] <Sam Sneddon [:gsnedders]>
oh, right. duh. because they're event targets.

[13:00:36.0513] <keithamus>
For spam, please don’t close any issues but instead forward them to me. I’ve enabled a flag on the repo which gives GH staff access to fast flag spammy accounts but only if the issue/pr remains open. Of course if you’d rather continue closing and labelling as spam please continue to do so, but if you are comfortable linking or otherwise leaving spammy issues to me then I’ll be able to tag them properly in our systems. 

[13:26:38.0324] <TabAtkins>
The issue is https://github.com/whatwg/meta/issues/222


2023-12-18
[04:46:59.0998] <chargeitall>
TabAtkins: Thank you for sending me the link. I appreciate your help. since I am not very familiar with web development and I don’t think I can offer much help. I will just wait for the issue to be resolved.

[06:31:28.0052] <annevk>
I could have sworn we had tests for attribute selectors with HTML case-insensitive attribute values, but I can't find any

[06:43:40.0657] <Ms2ger>
Values? I vaguely remember something about names

[07:15:37.0611] <annevk>
Ms2ger: you might be able to review as well: https://github.com/web-platform-tests/wpt/pull/43709 :-)

[07:28:29.0168] <chargeitall>
I noticed that the "Note", "Example", "1.", and "2." cannot be highlighted, which means the AI will not consider them when interacting with the highlighted paragraph. Is there a way to modify this behavior from your side, or do I need to do something from my side?

[07:32:50.0402] <TabAtkins>
No problem. Yourv question reminded me about it, so I'll try to get it done over the quiet break here.

[07:54:52.0253] <Ms2ger>
You asked... :)

[09:28:39.0771] <TabAtkins>
They're CSS generated content, which browsers still aren't great at handling in some ways.

[12:06:10.0548] <zcorpan>
html now has > 10,000 issues/PRs

[12:06:34.0752] <bkardell>
Wow

[15:59:27.0257] <TabAtkins>
Y'all beat CSS, we're still lagging at 9700ish right now


2023-12-19
[16:47:55.0709] <Alan Stearns>
And it looks like we started earlier

[23:29:54.0594] <annevk>
I think CSS allows commits to main? We don't.

[03:36:12.0885] <chargeitall>
However, if the content inserted via ::before or ::after pseudo-elements is indeed semantically important, it should be included in the HTML instead to ensure it is accessible to all users. 

[03:37:12.0367] <chargeitall>
 * if the content inserted via ::before or ::after pseudo-elements is indeed semantically important, it should be included in the HTML instead to ensure it is accessible to all users. 

[09:33:42.0949] <TabAtkins>
Ah true, if y'all are all-PR-all-the-time then we're probably still ahead *in spirit*

[12:43:59.0669] <snek>
do html or css specify what should happen when different characters of a ligature are styled differently https://gc.gy/aea050d0-8475-4e0e-8324-027a225ddf1d.png

[12:45:56.0629] <snek>
i didn't see anything in the html spec or css fonts spec but i'm not really sure what i'm looking for 😔

[12:48:03.0919] <snek>
 * do html or css specify what should happen when different characters of a ligature are styled differently https://gc.gy/aea050d0-8475-4e0e-8324-027a225ddf1d.png (in this case there is a span around the `i` with a background color applied)

[12:58:33.0479] <Alan Stearns>
snek: As far as I know this (and where you might put the cursor as you arrowkey through a ligature) is not defined. And from what I recall Firefox may be the only engine that tries to do something in this case

[13:53:04.0157] <TabAtkins>
No, this is unspecified.

[13:53:31.0253] <TabAtkins>
Exactly what you're even *capable* of doing is dependent entirely on your font rendering engine, so we just don't say anything.

[14:57:46.0707] <snek>
makes sense, thx

[15:05:16.0213] <TabAtkins>
WHATWG darkmode PR up at https://github.com/whatwg/whatwg.org/pull/429 !

[15:28:02.0890] <TabAtkins>
The specs say that ::before/etc should be highlightable/etc. Firefox does this, Chrome doesn't yet.


2023-12-20
[10:48:47.0784] <TabAtkins>
Where can I see a Review Draft? I'll need to check up on some CSS bits.

[11:06:31.0067] <annevk>
TabAtkins: /review-drafts/

[11:07:07.0942] <TabAtkins>
ah, got it. thanks

[14:21:42.0209] <Chris de Almeida>
> <@tabatkins:matrix.org> WHATWG darkmode PR up at https://github.com/whatwg/whatwg.org/pull/429 !

splendid!  and reminds me I'm overdue for my PR for `tc39.es`..  maybe I'll do it before year end

[14:24:53.0309] <Chris de Almeida>
> <@tabatkins:matrix.org> WHATWG darkmode PR up at https://github.com/whatwg/whatwg.org/pull/429 !

 * splendid!  and reminds me I'm overdue for my PR for `tc39.es`..  maybe I'll do it before year end.  edit: it's done, I just need to create the PR.  although there might be one or two remaining items for the spec page


2023-12-21
[05:12:43.0995] <chargeitall>
My understanding was that the DOM does not include the content created by pseudo elements. How does Firefox manage to make it selectable?

[05:15:37.0442] <chargeitall>
I was unable to select the word "Note" in Firefox.☝️

[06:21:34.0584] <chargeitall>
https://drafts.csswg.org/css-ui-4/#propdef-user-select

[08:15:48.0547] <chargeitall>
none of these worked 

.note::before {
    background: green;
    -webkit-user-select: text;
    -moz-user-select: text;
    -ms-user-select: text;
    user-select: text;
}


2023-12-22
[21:49:13.0890] <sideshowbarker>
jarhar: https://bugs.webkit.org/show_bug.cgi?id=228843#c5 makes me wonder how for auto-expanding `details` y’all dealt with AT/screen-reader handling. Did you yourself implement something for it? Or else somebody else did in in a separate patch? Or else for Chromium it was already working and you didn’t have to add anything new?

[21:49:28.0605] <sideshowbarker>
 * jarhar: https://bugs.webkit.org/show\_bug.cgi?id=228843#c5 makes me wonder how for auto-expanding `details` y’all dealt with AT/screen-reader handling. Did you yourself implement something for it? Or else somebody else did it in a separate patch? Or else for Chromium it was already working and you didn’t have to add anything new?

[02:47:37.0258] <chargeitall>
https://infra.spec.whatwg.org/

[05:10:38.0737] <chargeitall>
 * <!DOCTYPE html>
<html lang="en">
<head>
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
    <title>Document</title>
    <style>
        .note::before {
           background: green;
           content: "Note";
           -webkit-user-select: text;
           -moz-user-select: text;
           -ms-user-select: text;
            user-select: text;
             }
         </style>
   </head>
<body>
    <p class="note" role="note">
           ::before
           "Text" </p> 
</body>
</html>


[05:10:52.0368] <chargeitall>
 * `\<!DOCTYPE html>\<html lang="en">
\<head>
    \<meta charset="UTF-8">
    \<meta name="viewport" content="width=device-width, initial-scale=1.0">
    \<title>Document\</title>
    \<style>
        .note::before {
           background: green;
           content: "Note";
           -webkit-user-select: text;
           -moz-user-select: text;
           -ms-user-select: text;
            user-select: text;
             }
         \</style>
   \</head>
\<body>
    \<p class="note" role="note">
           ::before
           "Text" \</p> 
\</body>
\</html>`

[05:12:40.0442] <chargeitall>
 * I am unable to select the css generated content "Note" by the pseudo class ::before

```
<!DOCTYPE html>
<html lang="en">
<head>
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
    <title>Document\</title>
    <style>
        .note::before {
           background: green;
           content: "Note";
           -webkit-user-select: text;
           -moz-user-select: text;
           -ms-user-select: text;
            user-select: text;
             }
         </style>
   </head>
<body>
    <p class="note" role="note">
           ::before
           "Text" </p> 
</body>
</html>
```


[05:42:55.0254] <chargeitall>
is there a better workaround? https://jsfiddle.net/danield770/0v5ayz8j/1/

[06:27:45.0317] <chargeitall>
https://stackoverflow.com/questions/24385171/is-it-possible-to-select-css-generated-content


2023-12-23
[04:32:40.0685] <keithamus>
https://github.com/whatwg/html/pull/8467 looks like the spec has got approvals from annevk and Domenic. I'm curious what is the next step here? I've raised https://github.com/WebKit/WebKit/pull/22223 for WebKit to ship it by default, but I don't know if that's going to be blocked by the spec PR still being open? (I guess this could be counted as a friendly ping 😉).


2023-12-24
[02:16:57.0861] <annevk>
keithamus: merged. I read through the blink-dev thread again and it seems like it should be fine even if they're not prepared to make changes just yet. Hopefully once it hits Safari stable.

[02:17:57.0042] <keithamus>
Thanks! I’ve been chatting with jarhar about it and I’m hopeful 🙌


2023-12-25
[05:20:49.0471] <annevk>
keithamus: I just noticed you suggested to leave spammy issues open and notify you. Will attempt to do that from now on. My goto is banning the person and deleting the issue, might be a bit hard to unlearn, but doable. Once you flag them will they appear as deleted to others?

[05:22:58.0455] <keithamus>
If the spam team assesses that the user is spamming and flags them, then their comments and issues will be hidden from regular users’ views, and their account may be removed. 

[05:26:25.0377] <keithamus>
You’re welcome to continue deleting issues and banning people; it means that the users activity will be in the regular spam queue so it might take longer for the spam team to evaluate the users behaviour. As an employee though, any issue I mark as invalid floats to the top of the queue. We’re working to bring some extra functionality to repos and orgs to get similar functionality but I don’t know what that’ll shape up to be nor the timelines involved. 

[05:57:25.0055] <annevk>
Thanks, sounds good. And yeah, happy to report for the greater good and I used to do that, but they eventually created something that was too onerous and involved too many steps.

[05:57:32.0360] <annevk>
 * Thanks, sounds good. And yeah, happy to report for the greater good and I used to do that, but they eventually created something that was too onerous.


2023-12-27
[00:30:48.0652] <snek>
with the recent change to the websocket constructor, what's the expected result of `new WebSocket(null)`? is it equiv to the string `"null"` and you should get something out like `https://example.com/null`?

[00:33:09.0368] <snek>
 * with the recent change to the websocket constructor, what's the expected result of `new WebSocket(null)`? is it equiv to the string `"null"` and you should get something out like `wss://example.com/null`?

[00:37:28.0502] <snek>
i guess that's what webkit implemented https://github.com/WebKit/WebKit/pull/13751/files#diff-d989d5bdf667a51d97365c1a801fc33de5101b60be507264ff0eb3617c6f2d59R11


2023-12-28
[23:26:03.0384] <annevk>
snek: yup, follows from IDL


2023-12-30
[05:33:55.0398] <Maxim Vaarwel>
Hello everyone. Could somebody explain one thing that is used in html spec named **custom element reactions stack**. What are that cases which form stack from many **element queue**? For example this decorator https://html.spec.whatwg.org/#cereactions uses actions at first "push" and then "pop" with stack, between actions it says: "Run the originally-specified steps for this construct...". But how that step can increase size of **custom element reactions stack**?


2023-12-31
[00:46:32.0819] <annevk>
Maxim Vaarwel: if you read that CEReactions section carefully (and past the notes) you'll find it references algorithms that cause it to be relevant

[09:11:31.0608] <Maxim Vaarwel>
annevk: maybe I don't understand something that is implicit. I investigated all relevant links with CEReactions and really hard to get it what could make size of **custom element reactions stack** more than 1. Would be nice if you precise specify at least one case when it possible. Thanks.

[11:05:59.0168] <annevk>
Maxim Vaarwel: if you run the innerHTML setter and it queues up many reactions for custom elements?