Attr is per-Element, and one ElementAttributeData can be used by multiple Elements. Hence, Attr should be managed by Element instead of ElementAttributeData.
Created attachment 172216 [details] Cool patch
Comment on attachment 172216 [details] Cool patch Attachment 172216 [details] did not pass qt-ews (qt): Output: http://queues.webkit.org/results/14726117
Comment on attachment 172216 [details] Cool patch Attachment 172216 [details] did not pass qt-wk2-ews (qt): Output: http://queues.webkit.org/results/14726116
Created attachment 172217 [details] Cooler patch
Comment on attachment 172216 [details] Cool patch View in context: https://bugs.webkit.org/attachment.cgi?id=172216&action=review > Source/WebCore/dom/Node.h:243 > + bool hasSyntheticAttrChildNodes() const { return getFlag(HasSyntheticAttrChildNodesFlag); } > + void setHasSyntheticAttrChildNodes(bool f) { setFlag(f, HasSyntheticAttrChildNodesFlag); } f?! I like the use of 'synthetic' here.
Comment on attachment 172217 [details] Cooler patch Attachment 172217 [details] did not pass qt-ews (qt): Output: http://queues.webkit.org/results/14721175
Comment on attachment 172217 [details] Cooler patch Attachment 172217 [details] did not pass qt-wk2-ews (qt): Output: http://queues.webkit.org/results/14721176
Created attachment 172219 [details] Coolest patch
Comment on attachment 172219 [details] Coolest patch Clearing flags on attachment: 172219 Committed r133394: <http://trac.webkit.org/changeset/133394>
All reviewed patches have been landed. Closing bug.
Committed r133492: <http://trac.webkit.org/changeset/133492>