1425757 - Don't support a list of selectors in ::slotted yet

Status: RESOLVED FIXED

(Core :: CSS Parsing and Computation, enhancement, P1)

Description

[Emilio Cobos Álvarez (:emilio)]

Other browsers don't, and that allows them to use the rule hash for fast selector-matching.

Comment 1

[Emilio Cobos Álvarez (:emilio)]

Attachment: Bug 1425757: Don't support a list of selectors in ::slotted yet.

Review commit: https://reviewboard.mozilla.org/r/208040/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/208040/

Comment 2

[Xidorn Quan [:xidorn] UTC+10]

If others don't... we probably should make the spec match it, and maybe add notes for future extension...

Comment 3

[Xidorn Quan [:xidorn] UTC+10]

Comment on attachment 8937359 [details]
Bug 1425757: Don't support a list of selectors in ::slotted yet.

https://reviewboard.mozilla.org/r/208040/#review214378

::: servo/components/selectors/parser.rs
(Diff revision 1)
>          assert!(parse("::slotted(div + bar)").is_err());
>          assert!(parse("::slotted(div) + foo").is_err());
>          assert!(parse("div ::slotted(div)").is_ok());
>          assert!(parse("div + slot::slotted(div)").is_ok());
>          assert!(parse("div + slot::slotted(div.foo)").is_ok());
> -        assert!(parse("div + slot::slotted(.foo, bar, .baz)").is_ok());

Maybe changes to `assert!(parse(...).is_err());` and add a TODO here that if at some point we start supporting it again, we can simply flip the assertion here.

::: servo/components/style/selector_map.rs:418
(Diff revision 1)
>              Bucket::LocalName {
>                  name: &selector.name,
>                  lower_name: &selector.lower_name,
>              }
>          }
> +        Component::Slotted(ref selector) => find_bucket(selector.iter()),

I'm a bit uncertain about this. So a rule with `::slotted()` may be put into a bucket which is either for the originating slot element, or the assigned element. Would that cause any trouble in selector matching?

This may be worth a bit more comment, I guess.

Comment 4

[Xidorn Quan [:xidorn] UTC+10]

Could you file an issue to relevant spec to propose a spec change given the implementation status of this among browsers as well as the fact that doing it makes it easier to optimize?

Comment 5

[Emilio Cobos Álvarez (:emilio)]

https://github.com/w3c/csswg-drafts/issues/2158

(Also, this landed in https://github.com/servo/servo/pull/19610)