04:18
<jmdyck>
In the HTML spec, in https://html.spec.whatwg.org/#the-source-element (4.8.2 The source element), in the first Note, the phrase "react to changes in the environment" is a link to https://html.spec.whatwg.org/#img-environment-changes. However, if you follow the link and click on "react to changes in the environment", the popup doesn't list the occurrence in 4.8.2.
04:19
<jmdyck>
This is presumably because the 4.8.2 occurrence is an <a href> rather than a <span data-x>. Is that intentional?
04:25
<jmdyck>
zcorpan: You added the 4.8.2 <a href> in 9e9a6730 about 2.5 years ago.
04:40
<jmdyck>
Similarly, in 4.8.5 "The iframe element", at https://html.spec.whatwg.org/#:~:text=Furthermore%2c%20notice , "optional start and end tags" uses an <a href> to link to https://html.spec.whatwg.org/#syntax-tag-omission, whose popup doesn't link back to 4.8.5.
04:41
<jmdyck>
That one goes back to 2013, before data-x, so I think maybe it evaded conversion to <span data-x> when that came along.
04:42
<jmdyck>
I think those are the only two such cases.
04:46
<jmdyck>
Wondering if I should submit a fix for them.
09:39
<annevk>
Luke Warlow: we had agreement that form pseudo-elements would not expose underlying API details right so ::picker(select):popover-open would not work?
09:39
<annevk>
I'm seeing WPT tests for the opposite. Working on a correction now.
09:40
<Luke Warlow>
Hmm it might be worth checking GitHub for usage? But yeah :open::picker(select) does the same job so should be fine.
09:42
<Luke Warlow>
The picker stuff might have pre-dated that discussion. Also I think currently the fact it's popover is exposed in other ways? The fact it closes other popovers for example. See Keith's recent issue on that. Personally I think it's probably harmless but there's an alternative so I think it's also fine to drop it.
09:44
<annevk>
I'm pretty sure I said this from the when we proposed ::picker so I'm rather annoyed.
09:47
<annevk>
I haven't really made up my mind what would be best for Keith's issue. I do think we want consistency between native and base appearance as ultimately that should just be (mostly) appearance.
09:48
<Luke Warlow>
Isn't this (the picker being a popover) something well defined though? The implementation has to use a popover else there will be differences? It's not something that affects appearance:auto too? I can't remember the discussions when picker originally came up unfortunately. I suspect usage is low and it's changeable. Is this something specified one way or another? I guess it might just be a consequence of picker being a popover? I can't speak for whether chromium will want to change it but I'm fine with it from a css forms perspective.
09:52
<Luke Warlow>
Hmm it might be worth checking GitHub for usage? But yeah :open::picker(select) does the same job so should be fine.
Thinking about it most of the time the styles don't need to apply to the picker conditionally because it's only shown when open so usage is probably even lower. Perhaps it's used for some animations or something. Do the tests shine a light on the intention?
09:52
<annevk>
It's a popover, but what pseudo-classes a pseudo-element exposes is a separate matter.
09:53
<annevk>
For animations maybe? I'm going to change html/semantics/forms/the-select-element/customizable-select/select-picker-animations.html and html/semantics/forms/the-select-element/customizable-select/switch-picker-appearance.html. Will put up a PR shortly.
09:57
<zcorpan>
jmdyck: I don't recall why I used <a href>, it's clearly inconsistent with the other note that has a normal xref
10:11
<annevk>
PR mentioned above: https://github.com/web-platform-tests/wpt/pull/57739
15:11
<Dominic Farolino>
If you create a cross-origin iframe whose name=foo via the iframe attr, then window.foo should be undefined, right? That's my reading of https://html.spec.whatwg.org/#document-tree-child-navigable-target-name-property-set. I'm asking because it seems no browser follows this; all let you reference the child Window via window.foo
15:35
<Andreu Botella>
Hey, I'm not sure why the pr-preview isn't working for https://github.com/whatwg/webidl/pull/1568
15:35
<Andreu Botella>
I thought it was because the bikeshed build was failing, but I fixed that and it's still not working
15:47
<annevk>
Dominic Farolino: I wonder if that regressed with the navigable refactor. name=foo should work. The cross-origin frame setting window.name="foo" is what should remain hidden.
15:50
<Dominic Farolino>
annevk: Just to be clear, by "name=foo should work" do you mean <iframe name=foo src=x-origin> should let the parent reference the child via window.foo?
15:50
<Dominic Farolino>
I agree we should change the spec to match browsers on this one.
16:00
<annevk>
Dominic Farolino: yup, you can always address cross-origin iframes, you should just not be able to obtain information from them (though we are bad at that too of course)
16:16
<zcorpan>
annevk: is https://github.com/whatwg/html/pull/12071 ok now?
16:37
<Dominic Farolino>
annevk: One thing I'm not clear on is: if the x-origin iframe sets window.name to something new, should window.parent.newName === window from within the iframe? This q is mostly about spec machinery... should window.parent.newName from within the frame succeed bc the targeted frame is same-origin to the one doing the accessing? Or should it fail because the targeted frame is cross-origin to the window.parent object from which the named access is done?
16:38
<Dominic Farolino>
I'm 99.9% sure it should fail, because we use the relevant settings object in https://html.spec.whatwg.org/#named-access-on-the-window-object:relevant-settings-object for the same-origin check, which disregards the fact that we're "inside the iframe" at the time of access. Sound right?
16:40
<annevk>
Dominic Farolino: yeah I don't think we could make that work because it also impacts supported property names and such. A good thing to put on the "let's make sure this is tested" list though.
16:41
<annevk>
zcorpan: shall I do a final review Monday?
16:41
<zcorpan>
annevk: sounds good, thanks