newsplex.com map does not load after r204441. It loads fine in Firefox and Chrome.
<rdar://problem/28753438>
Created attachment 294687 [details] Patch
Comment on attachment 294687 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=294687&action=review > Source/WebCore/dom/TagCollection.h:46 > + AtomicString m_qualifiedName; I think it is confusing to call these strings "qualified names" and also call the data structure QualifiedName the same thing. We should figure out what our terminology is. > Source/WebCore/dom/TagCollection.h:86 > + return m_qualifiedName == element.tagQName().toString(); We should find a way to do this comparison without allocating a string each time. > Source/WebCore/dom/TagCollection.h:100 > + return m_loweredQualifiedName == element.tagQName().toString(); > + return m_qualifiedName == element.tagQName().toString(); Ditto.
Comment on attachment 294687 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=294687&action=review >> Source/WebCore/dom/TagCollection.h:46 >> + AtomicString m_qualifiedName; > > I think it is confusing to call these strings "qualified names" and also call the data structure QualifiedName the same thing. We should figure out what our terminology is. I following the naming in the specification. It is a qualified name stored as a String rather than a QualifiedName type. >> Source/WebCore/dom/TagCollection.h:86 >> + return m_qualifiedName == element.tagQName().toString(); > > We should find a way to do this comparison without allocating a string each time. In the common case, there is no prefix and therefore it will return the localName. It will not allocate a new String, only cause some refcounting churn.
Comment on attachment 294687 [details] Patch Clearing flags on attachment: 294687 Committed r208674: <http://trac.webkit.org/changeset/208674>
All reviewed patches have been landed. Closing bug.