[css-scoping-1] ::placeholder selector inside ::slotted() not working

[jorenbroekema]

Referencing this issue #1747

Working as intended:
::slotted(::hover){ color: red; }

Not working:
::slotted(::placeholder){ color: red; }

From the issue I am referencing, it seems like it is supposed to not work in both scenarios, as option 3 was chosen to be allowed and option 2 was not. So basically, ::slotted()::hover should work but ::slotted(::hover) should not. At least, this was my interpretation of that issue.

Could someone give some clarification on what should work and what should not, and perhaps why in my case the ::placeholder selector within ::slotted does not work?

[emilio]

:hover is a pseudo-class, ::placeholder a pseudo-element. ::slotted(::placeholder) should not work, though arguably ::slotted(*)::placeholder should, per that issue.

[tabatkins]

Yeah, the argument to ::slotted() is a selector that is matched against the elements that get slotted. By definition, none of these elements are pseudo-elements, so ::slotted(::placeholder) can never match anything.

But yeah, ::slotted()::placeholder should work, and if it doesn't, that's a browser bug.

(::slotted(::hover) definitely shouldn't be working, as there's no pseudo-element named ::hover. Are you sure you actually have that precise selector in your stylesheet, or do you have ::slotted(:hover) instead?)

[emilio]

Actually I think this is a valid issue. The resolution in #1747 only referenced tree-abiding pseudos (::before / ::after).

I actually have an implementation that works for all pseudos, but I don't plant to land it without a resolution here.

[lilles]

According to css-pseudo-4, the tree-abiding pseudos are: ::before, ::after, ::marker, ::placeholder. So it currently follows from the specs that ::placeholder should be allowed after ::slotted().

There's no limitation in Blink that makes it hard to support any kind of pseudo elements after ::slotted(), but the previous CSSWG resolution was to limit this.

Reported Blink issue: https://crbug.com/902518

[emilio]

Ah, alright, true that. I'll fixup firefox as well.

[emilio]

(I agree this is invalid then, just a browser bug)

[emilio]

https://bugzilla.mozilla.org/show_bug.cgi?id=1505213 fwiw :)

[lilles]

Wasn't quite as simple in Blink: https://chromium-review.googlesource.com/c/chromium/src/+/1323049

[jholt1]

Adding here in case others are searching for this:

Webkit has yet to fix the bug https://bugs.webkit.org/show_bug.cgi?id=223814