988655 - nsIContent::IsInAnonymousSubtree returns true for shadow DOM content

Status: RESOLVED FIXED

(Core :: DOM: Core & HTML, defect)

Description

[Bobby Holley (:bholley)]

The current implementation of this method is:

return IsInNativeAnonymousSubtree() || GetBindingParent() != nullptr;

Since shadow DOM nodes return a non-null GetBindingParent(), this returns true for those cases. Talking with wchen, it doesn't sound like it was intentional, and probably isn't something that callers are expecting. Though I am concerned that we're relying on various anonymous content features (like event retargetting) without realizing it.

I'm going to try fixing it - hopefully nothing breaks.

Comment 1

[Bobby Holley (:bholley)]

Attachment: Stop returning true for shadow DOM nodes in nsIContent::IsInAnonymousSubtree. v1

Comment 2

[Bobby Holley (:bholley)]

https://tbpl.mozilla.org/?tree=Try&rev=8a92415d913d

Comment 3

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/74a617a50287

Comment 4

[u279076]

(In reply to Bobby Holley (:bholley) from comment #0)
> I'm going to try fixing it - hopefully nothing breaks.

What could this break, potentially? Is there anything QA should be on the lookout for?

Comment 5

[Bobby Holley (:bholley)]

(In reply to Anthony Hughes, QA Mentor (:ashughes) from comment #4)
> (In reply to Bobby Holley (:bholley) from comment #0)
> > I'm going to try fixing it - hopefully nothing breaks.
> 
> What could this break, potentially? Is there anything QA should be on the
> lookout for?

Possibly stuff related to Web Components. But the breakage I was expecting would have been visible on tinderbox, I think.

Comment 6

[William Chen [:wchen]]

Another thing to keep in mind is that web components are pref'ed off and not exposed to the web by default. The only people using it are developers experimenting with the feature.

Comment 7

[u279076]

Thanks Bobby and William. I'm tagging this [qa-] for now. If you need testing around this please don't hesitate to needinfo me.