Otherwise we may get confused in layout, thinking that the node is part of the flattened tree when it is not.

Comment on attachment 8929684 [details] Bug 1418560: Properly return null from GetFlattenedTreeParent if the node is XBL fallback content and the binding is active. https://reviewboard.mozilla.org/r/200974/#review206296 Would be nice to have a test for this, if possible. Re GetXBLBinding, agreed; it might be worth trying to devirtualize that. ::: dom/base/nsIContentInlines.h:85 (Diff revision 2) > - // Check if the node is an explicit child of an element with a shadow root, > - // re-bound to an insertion point. > + // Check if the node is an explicit child of an element with a shadow root or > + // an element with an XBL binding, re-bound to an insertion point. In the comment can you mention that the check for parent->GetXBLBinding() is specifically for the case of children that are not distributed to insertion points? ::: layout/base/nsCSSFrameConstructor.cpp:7494 (Diff revision 2) > + for (nsIContent* node = iter.GetNextChild(); node; node = iter.GetNextChild()) > + if (node == aChild) > + return true; Nit: braces around statements.

Pushed by ecoal95@gmail.com: https://hg.mozilla.org/integration/autoland/rev/ff9caa303cc1 Properly return null from GetFlattenedTreeParent if the node is XBL fallback content and the binding is active. r=heycam https://hg.mozilla.org/integration/autoland/rev/eb531fabadea Disable test_content_element.html. r=smaug

Pushed by ecoal95@gmail.com: https://hg.mozilla.org/integration/autoland/rev/6fba2818b1ee Don't build IsFlattenedTreeChild in release builds. r=bustage

https://hg.mozilla.org/mozilla-central/rev/ff9caa303cc1 https://hg.mozilla.org/mozilla-central/rev/eb531fabadea https://hg.mozilla.org/mozilla-central/rev/6fba2818b1ee