| 17:12 | <Luke Warlow> | Regarding the pseudo-element discussion I think it would be very helpful to have concrete use cases. Because I worry this is designing something that doesn't actually solve use cases. It gives web authors maybe some of the stuff they want to solve use cases. For example, ::backdrop keeps being mentioned. What do authors want to do with it? We have closedby=any for the light dismiss case now. Can we perhaps look at more things like that? |
| 17:14 | <Noam Rosenthal> | I used backdrop because it's easy to demo with, but the actual use case this started from is about scroll markers |
| 17:15 | <Noam Rosenthal> | This issue started from view-transitions: https://github.com/w3c/csswg-drafts/issues/11559 |
| 17:16 | <Noam Rosenthal> | and https://github.com/w3c/csswg-drafts/issues/12163 is about scroll-markers |
| 17:16 | <Noam Rosenthal> | I think if we find something that's general enough for these two use cases we'll be off for a good start |
| 17:19 | <Noam Rosenthal> | Luke Warlow: with a11y - proxying with an element can have some interesting benefits if designed right... like the HTML aria attributes can override the default pseudo-element attributes, giving more control |
| 17:20 | <Noam Rosenthal> | (benefits over having this as a JS-only thing) |
| 17:21 | <Luke Warlow> | There was mention of popovertarget, can you speak more to what the thinking there is? Given they only work on buttons right now? |
| 17:22 | <Luke Warlow> | For some of these, just hiding the the pseudo and using HTML as a replacement provides all those benefits though. Which is why I'm wondering some of the use cases other than backdrop. |
| 17:23 | <Noam Rosenthal> | Yea, for some of these, sure... but for the style/layout-generated pseudos there is no good alternative |
| 17:24 | <Noam Rosenthal> | Like an equivalent alternative would be to create an actual shadow-dom for these and give you access to the ShadowRoot |
| 17:24 | <Noam Rosenthal> | (which is perhaps an alternative we can consider) |
| 17:25 | <Luke Warlow> | Sorry if I'm missing something obvious but these are both quite solution oriented? Is there a summary of use cases rather than 'how do make pseudo elements real'? |
| 17:26 | <Noam Rosenthal> | Look above, the use cases come from people trying to do things with view transitions and scroll markers and not being able to. Definitely problem oriented, but the solution went into the more general space after years of adding piecemeal pseudo-element supports in specific APIs like getComputedStyle and web animations |
| 17:29 | <Noam Rosenthal> | It's easy to say "let's just fix this point problem" over and over, and sometimes it's the right thing to do, but I think that's sort of how we went about pseudo-elements from the start and it got us to a place where it's not a concept that's well defined enough at the platform level |
| 18:09 | <TabAtkins> | Exactly that, yes. We keep growing more pseudo-elements, and it's clear the piecemeal solutions are not viable for our additions, let alone the clear path for the future. |
| 18:24 | <Noam Rosenthal> | having said that, a list/table of use cases for this would be helpful either way. |
| 18:28 | <Noam Rosenthal> | (there might be one that I don't have a link to, I asked Daniil who was supposed to present the issue today) |
| 23:46 | <jmdyck> | In recent merge to HTML spec re deferred commit, https://html.spec.whatwg.org/#inner-navigate-event-firing-algorithm, step 35.1 defines variable 'precommitController', but doesn't use it. |
| 23:46 | <jmdyck> | Might be a mistake. |
| 23:54 | <Domenic> | Noam Rosenthal: ^ |