Only thing this is used for anymore is SVGFontFaceElement. That can be handled without confusing extra stylesheets.
Created attachment 133889 [details] patch
Comment on attachment 133889 [details] patch Attachment 133889 [details] did not pass qt-wk2-ews (qt): Output: http://queues.webkit.org/results/12141126
Comment on attachment 133889 [details] patch Attachment 133889 [details] did not pass qt-ews (qt): Output: http://queues.webkit.org/results/12076033
Comment on attachment 133889 [details] patch TD;DR but it looks pretty good. Fix the build and slap me for an r+ tomorrow.
Created attachment 134010 [details] Try to fix qt build
Comment on attachment 134010 [details] Try to fix qt build Attachment 134010 [details] did not pass qt-wk2-ews (qt): Output: http://queues.webkit.org/results/12147392
Comment on attachment 134010 [details] Try to fix qt build Attachment 134010 [details] did not pass qt-ews (qt): Output: http://queues.webkit.org/results/12142409
Comment on attachment 134010 [details] Try to fix qt build View in context: https://bugs.webkit.org/attachment.cgi?id=134010&action=review > Source/WebCore/dom/Document.h:638 > + HashSet<SVGFontFaceElement*>& svgFontFaceElements(); As discussed on IRC, this needs ENABLE(SVG) guards, and should be moved right into SVGDocumentExtensions.
Created attachment 134017 [details] moved the map to SVGDocumentExtensions
Comment on attachment 134017 [details] moved the map to SVGDocumentExtensions Attachment 134017 [details] did not pass qt-wk2-ews (qt): Output: http://queues.webkit.org/results/12147423
Comment on attachment 134017 [details] moved the map to SVGDocumentExtensions Attachment 134017 [details] did not pass qt-ews (qt): Output: http://queues.webkit.org/results/12142434
Created attachment 134021 [details] test SVG_FONTS instead of SVG
Comment on attachment 134021 [details] test SVG_FONTS instead of SVG Me gusta.
Comment on attachment 134021 [details] test SVG_FONTS instead of SVG View in context: https://bugs.webkit.org/attachment.cgi?id=134021&action=review r=me with comments: > Source/WebCore/svg/SVGDocumentExtensions.cpp:421 > +void SVGDocumentExtensions::registerSVGFontFaceElement(SVGFontFaceElement* element) How about checking in the SVGDocExt dtor that this set is empty? I'm not sure we do it for the others, but we should do so to catch missing unregister() calls. I hope removedFromDocument is guaranteed to be called :-)
http://trac.webkit.org/changeset/112258
Reopen, because it broke debug builds on Qt: ../../../../Source/WebCore/svg/SVGDocumentExtensions.cpp: In destructor ‘WebCore::SVGDocumentExtensions::~SVGDocumentExtensions()’: ../../../../Source/WebCore/svg/SVGDocumentExtensions.cpp:53: error: ‘m_svgFontFaceElements’ was not declared in this scope Additionally it made almost all tests assert: STDERR: ASSERTION FAILED: m_svgFontFaceElements.isEmpty() STDERR: ../../Source/WebCore/svg/SVGDocumentExtensions.cpp(53) : WebCore::SVGDocumentExtensions::~SVGDocumentExtensions()
Removed the assert in ~SVGDocumentExtensions in http://trac.webkit.org/changeset/112269. It seems insertedInto/removedFromDocument don't always pair.