- From: Rune Lillesveen via GitHub <sysbot+gh@w3.org>
- Date: Fri, 04 Mar 2022 07:47:21 +0000
- To: public-css-archive@w3.org
It was not accidental from my side, and the implementation is in line with the spec. The rationale for the current spec, is that #child will always match my-component for the container query in the example below. If we use flat tree descendants instead of shadow-including descendants, my-component's shadow tree may intercept the query, either by an implicit size container, if names are not used, or by a name clash (perhaps not likely). That behavior is a bit surprising too as it exposes internals of my-component. ```html <style> @container foo size(width >= 400px) { #child { color: green; } } </style> <my-component style="width:400px; container: foo / inline-size"> <div id="child">Not green?</div> </my-component> ``` I don't have very strong opinions here, and I think there are pros and cons to both behaviors. I think the part about starting from the originating element for selectors with pseudo element still holds, so not exposing shadow internals for ::part() is not affected by the choice. @tabatkins Was this surprising to you too? I think what we are looking at here is to possibly change the spec from using "shadow-including descendants" to "flat tree descendants". Changing the implementation in Blink is trivial. -- GitHub Notification of comment by lilles Please view or discuss this issue at https://github.com/w3c/csswg-drafts/issues/5984#issuecomment-1058917431 using your GitHub account -- Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config
Received on Friday, 4 March 2022 07:47:23 UTC