CSSRuleData and CSSRuleDataSet don't really need to be in CSSStyleSelector.h (they are only used in the .cpp file). Moving them to the .cpp file will reduce the number of accidental recompiles. In addition, it's possible that CSSRuleDataSet can just be replaced with a std::list container. (Filed as an offshoot of bug 27748).
I'm not working on this now nor do I expect to be in the near future, so I am disclaiming ownership in case someone else wants to pick this up.
Created attachment 77286 [details] Patch Move CSSRuleData and CSSRuleDataList from the .h to the .cpp file. Not sure if changing the custom linked list implementation to one of the generic WTF containers is still preferred.
Comment on attachment 77286 [details] Patch Seems sane. There shouldn't be any perf impact if this is the only class using these.
The commit-queue encountered the following flaky tests while processing attachment 77286 [details]: fast/workers/storage/use-same-database-in-page-and-workers.html bug 50995 (author: dumi@chromium.org) The commit-queue is continuing to process your patch.
Comment on attachment 77286 [details] Patch Clearing flags on attachment: 77286 Committed r74642: <http://trac.webkit.org/changeset/74642>
All reviewed patches have been landed. Closing bug.
http://trac.webkit.org/changeset/74642 might have broken GTK Linux 32-bit Release
(In reply to comment #7) > http://trac.webkit.org/changeset/74642 might have broken GTK Linux 32-bit Release Seems unlikely to be caused by this.
They're all media/ test failures from a media change at the same time as this one. The Gtk bot is kinda slow, so many changes get built at once.