[css-pseudo] Default UA styles for markers don't affect ::before::marker and ::after::marker

[Loirooriol]

From https://drafts.csswg.org/css-pseudo-4/#marker-pseudo,

UAs must add the following rule to their default style sheet:

::marker {
  unicode-bidi: isolate;
  font-variant-numeric: tabular-nums;
}

But ::marker only selects markers originated by elements, not by other pseudo-elements.

So I guess the selector should be ::marker, ::before::marker, ::after::marker.

[faceless2]

Yes, I agree. That's almost exactly what we have in our user-agent sheet (we also reset text-indent, letter-spacing and word-spacing to initial)

[AmeliaBR]

I wonder if it would be easier to expand selector matching so that ::marker as a standalone selector does include ::before::marker etc? This is going to trip up authors as much as it did the spec editors! (Who even had a note immediately following them in the spec warning against the error.)

I think this is our first standard stacked pseudo-element selector, so it should still be possible to redefine how it works.

[Loirooriol]

@AmeliaBR But ::marker is equivalent to *::marker, where * can only select elements, not pseudo-elements. In #3826 @MatsPalmgren, @tabatkins and me agreed on this, though I don't remember if it was discussed by the whole CSSWG.

I wonder if a new kind of universal selector that includes tree-abiding pseudo-elements would be worth it. Let's say e.g. **, then the selector could be **::marker. It could also be useful for things like *, ::before, ::after { box-sizing: border-box }.

[fantasai]

@Loirooriol OK, I updated the UA rule selector.

@faceless2 I think adding more properties to the rule would be a separate issue. Would you mind filing that separately, along with some explanation of why you think these are a good idea? We'll want to get @MatsPalmgren and @dauwhe to check over that list, too. (text-indent seems pretty clear to me, I'm less sure about the text spacing properties.)

@AmeliaBR No, as @Loirooriol points out, the * selector (which is what's implied if you leave out a tag selector) only matches elements, not pseudo-elements.

@Loirooriol If you want to propose ** file a separate issue? :)

[Loirooriol]

Fixed in b7f55d4