https://bugs.webkit.org/show_bug.cgi?id=86019 Comment #5 From Julien Chaffraix 2012-05-10 13:44:33 PST (-) [reply] (From update of attachment 141023 [details]) View in context: https://bugs.webkit.org/attachment.cgi?id=141023&action=review > Source/WebCore/rendering/RenderObject.cpp:-331 > - // cannot affect the positioned object list, and the floating object list is irrelevant (since the list gets cleared on Mhh, re-reading the ChangeLog, it implies that the comment is wrong and should thus be fine. Dave was fine with the approach and I agree it's better to move this to the lower-level plumbing to avoid missing any callers. Could you file a bug about the artificial split we have between removeChild / removeChildNode? (I think those bugs arise because what we expect both functions to do and what should be the differences is not clear)