Looking at the "failure," it may actually be a "fix" because from a quick glance only one child is being added; not two. I should have some time today to verify this and update the expectations (or debug the failure). Deliberately not skipping the test for now.
<rdar://problem/26515630>
Created attachment 279969 [details] Patch
Chris: Should be a super-quick review. Rather than just rebaseline the test, I started including the child being added/removed in the data sent to the listener so we can verify we're getting what we assume we're getting. Thanks!
Comment on attachment 279969 [details] Patch Clearing flags on attachment: 279969 Committed r201458: <http://trac.webkit.org/changeset/201458>
All reviewed patches have been landed. Closing bug.