962628 - Sibling combinators should work on roots of anonymous content subtrees

Status: RESOLVED FIXED

(Core :: CSS Parsing and Computation, defect)

Description

[Boris Zbarsky [:bzbarsky]]

Right now they don't because we implement them on top of IndexOf().  But we don't want to use IndexOf anyway, and I think that if your anon content looks like this:

  <span class="a"></span>
  <span class="b"></span>

then ".a + .b" should match the second span.

Comment 1

[Boris Zbarsky [:bzbarsky]]

Attachment: Make '+' and '~' combinators work at the top level of an anonymous content forest.

Cameron, are you OK reviewing this?  If not, please punt to dbaron.

Comment 2

[Cameron McCormack (:heycam)]

Comment on attachment 8363719 [details] [diff] [review]
Make '+' and '~' combinators work at the top level of an anonymous content forest.

Review of attachment 8363719 [details] [diff] [review]:
-----------------------------------------------------------------

No, I'm happy with this.  Maybe replace the for loop entirely with a GetPreviousSiblingElement call?  And a test that exercises the NODE_HAS_SLOW_SELECTOR_LATER_SIBLINGS flag having been set on the parent, e.g. starting with #a having some text content, using script to remove the text content, and having this rule:

  .a:empty + .b { ... }

Comment 3

[Boris Zbarsky [:bzbarsky]]

> Maybe replace the for loop entirely with a GetPreviousSiblingElement call?

Good call.  It's _maybe_ a tiny bit slower due to not being inlined, but seems totally worth it.

Comment 4

[Boris Zbarsky [:bzbarsky]]

Attachment: Updated to review comments

Comment 5

[Boris Zbarsky [:bzbarsky]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/f88b13e2ae0a

Comment 6

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/f88b13e2ae0a