| 07:19 | <annevk> | keithamus: I still need to update the PRs, yes. The tests haven't been upstreamed, but that's an oversight as I understand it. (rniwa is on it: https://github.com/web-platform-tests/wpt/pull/50790.) |
| 09:35 | <hacknorris> | weird question but is it possible to make cursor sprites in css or i'm left only with cursor:none and JS and tracking cursor position? |
| 09:38 | <keithamus> | If you have an svg sprite-sheet you might be able to use something like cursor: url("data:image/svg+xml,<svg xmlns='http://www.w3.org/2000/svg' width='32' height='32'><use href="sprites.svg#icon"></use></svg>") |
| 09:38 | <hacknorris> | it's png sized 128px*128px and each tile is 32x32… |
| 09:39 | <hacknorris> | perfect for background-image hack but possibly not cursor… |
| 09:40 | <Noam Rosenthal> | Yea, I think you can do that with object-position |
| 09:40 | <Noam Rosenthal> | Ah cursor sprites, not sure |
| 09:41 | <keithamus> | if you sliced the images you could put them all in an svg as image embeds. You'd still end up with one resource, even if it's a bit hacky. But having said that I don't know if <use> in an svg cursor image will work. |
| 09:41 | <hacknorris> | yup. cursor image… |
| 09:43 | <hacknorris> | the one which changes to hand icon when over link |
| 09:45 | <hacknorris> | i don't want to slice… |
| 09:45 | <hacknorris> | i'll quicker do cursor:none + JS… |
| 11:42 | <sideshowbarker> | zcorpan: Noam Rosenthal If/when you have time to do an HTTP Archive query: I’m looking for data on role=radiogroup usage — and specifically, on how many have it on div or fieldset elements. |
| 11:42 | <sideshowbarker> | Context is https://github.com/w3c/aria/pull/2297#issuecomment-2660986380 |
| 11:43 | <Noam Rosenthal> | Happy to help. Can you supply a REGEX that would approximate this? |
| 11:44 | <Noam Rosenthal> | (REGEX is the best we can do in HTTP archive, otherwise you'd want something like a use counter) |
| 11:44 | <sideshowbarker> | yep, gimme a few minutes |
| 11:47 | <keithamus> | annevk: given foolip has approved I wondered if you had time to do one last pass on https://github.com/whatwg/html/pull/9841 ? While I have the time I'd love to be able to address any last review comments. |
| 11:51 | <sideshowbarker> | Noam Rosenthal: <(div|fieldset)[^>]+role=["']?radiogroup should work, I think |
| 11:56 | <sideshowbarker> | well actually, I guess it would be also be good to have the data for two as separate queries; so:
|
| 11:58 | <Noam Rosenthal> | The total is 0.01%, running the separate ones |
| 11:59 | <sideshowbarker> | Thanks! |
| 12:04 | <Noam Rosenthal> | 1 136 div+fieldset 2 3500 fieldset 3 58782 div 4 672552031 neither |
| 12:11 | <sideshowbarker> | ah, excellent — double thanks |
| 12:12 | <sideshowbarker> | 672552031 is the total number of documents in the entire archive? |
| 12:44 | <Noam Rosenthal> | total desktop responses in November 2024 |
| 13:24 | <zcorpan> | keithamus: annevk : was the naming of commandForElement vs commandFor discussed? Most existing IDL attributes in HTML that return an element (or null) don't have Element suffix |
| 13:27 | <keithamus> | zcorpan: it was modelled after popoverTargetElement. |
| 13:30 | <zcorpan> | keithamus: ok |
| 13:31 | <keithamus> | that's not necessarily an objection to your comment, but I don't think the question has been raised before. Naming was discussed quite extensively (it used to be invokeTargetElement) but no one raised questions around the Element suffix IIRC. |
| 14:06 | <annevk> | zcorpan: all IDL attributes that are Element(s)-reflecting should have such a suffix |
| 14:06 | <zcorpan> | annevk: ok, I've missed that |
| 14:20 | <zcorpan> | CommandEvent source doesn't have an Element suffix |
| 14:23 | <Luke Warlow> | I think the aria reflection stuff all has Element / Elements suffixes too |
| 14:24 | <Luke Warlow> | Neither does target etc, I think was the logic there (if there was any). It was named source based on the option bag added to showPopover/togglePopover. It's also readonly as opposed to the others which do two way reflection. |
| 14:57 | <zcorpan> | ah, foolip merged. I'll continue reviewing and comment here for now |
| 14:58 | <zcorpan> | keithamus: show modal dialog can throw an exception, e.g. if element is not connected. Is that ok? Example: https://software.hixie.ch/utilities/js/live-dom-viewer/saved/13512 |
| 15:02 | <keithamus> | Ah we should probably check if the element is connected just before showing as modal I guess |
| 15:06 | <annevk> | zcorpan: CommandEvent doesn't use reflect, so it's not applicable? |
| 15:09 | <zcorpan> | annevk: true. But then https://html.spec.whatwg.org/#dom-navigateevent-sourceelement was recently added |
| 15:15 | <zcorpan> | keithamus: the other possibility is that the document is not fully active, but I think the button will have the same node document, and the activation behavior steps check for that (step 2). So I think connected is the only missing case |
| 15:17 | <zcorpan> | I can file a bug |
| 15:25 | <zcorpan> | https://github.com/whatwg/html/issues/11043 |
| 15:26 | <Luke Warlow> | We could alternatively do the same as popover? Have a way to turn exceptions into early returns? Though that's a bit messy so just adding the check might be nicer |
| 15:37 | <annevk> | zcorpan: I think sourceElement there is reasonable because there's multiple types of sources, but we only expose it for elements. |
| 15:41 | <zcorpan> | I like the Auto state btw 👌 |
| 15:45 | <keithamus> | Yeah I appreciate all the reviews from everyone, I think this has shaped up nicely. |
| 15:46 | <zcorpan> | IIUC, you can now do <form><button command=... commandfor=...> (without type), and it will be a command button instead of a submit button |
| 15:47 | <Luke Warlow> | It'll currently be a no-op |
| 15:47 | <Luke Warlow> | But in future it can be a command button instead of submit |
| 15:48 | <Luke Warlow> | (Unless I can't remember the details which at this point wouldn't be surprising 😅) |
| 15:48 | <Luke Warlow> | The reasoning is that in older browsers it will submit the form still so this encourages you to add an explicit type |
| 15:52 | <zcorpan> |
|
| 15:52 | <Luke Warlow> | Good catch! |
| 15:55 | <zcorpan> | OK, done reviewing. Nice work 👏 |
| 15:55 | <Luke Warlow> | Thanks for everyones reviews btw! |
| 16:01 | <Luke Warlow> | https://github.com/whatwg/html/issues/11044 - fixed at https://github.com/whatwg/html/pull/11048 |
| 18:29 | <Luke Warlow> | https://github.com/whatwg/html/pull/11047 - is something like this considered normative or editorial? |
| 18:58 | <Luke Warlow> | https://github.com/whatwg/html/pull/11049 - fixed here |
| 21:01 | <zcorpan> | normative |
| 21:41 | <hacknorris> | and i still don't know if it's possible to make cursor from tileset… |