We currently determine whether a CSSInitialValue is "implicit" or not based on its CSSValue::ClassType at construction. We should move this to a CSSValue bit instead to make the code a little less horrible.
Created attachment 115375 [details] Patch
Comment on attachment 115375 [details] Patch Yeah, seems cleaner. r=me.
Comment on attachment 115375 [details] Patch Clearing flags on attachment: 115375 Committed r100449: <http://trac.webkit.org/changeset/100449>
All reviewed patches have been landed. Closing bug.
Isn't this worse? m_implicit as a separate bit no longer communicates that only InitialClass can be implicit, and now costs a full bit rather than being one value of many in m_classType.
(In reply to comment #5) > Isn't this worse? m_implicit as a separate bit no longer communicates that only InitialClass can be implicit, and now costs a full bit rather than being one value of many in m_classType. I actually plan to remove both CSSInitialValue and CSSInheritedValue and move their representation into CSSValue.