Fixed up some #include dependencies so the WriteBarrier class can actually call Heap::writeBarrier
Created attachment 92917 [details] Patch
Comment on attachment 92917 [details] Patch Attachment 92917 [details] did not pass efl-ews (efl): Output: http://queues.webkit.org/results/8655802
Created attachment 92922 [details] Patch
Comment on attachment 92917 [details] Patch Attachment 92917 [details] did not pass mac-ews (mac): Output: http://queues.webkit.org/results/8655819
Comment on attachment 92922 [details] Patch Attachment 92922 [details] did not pass mac-ews (mac): Output: http://queues.webkit.org/results/8645819
Comment on attachment 92922 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=92922&action=review r=me, but I think you need to find out why this failed on the Mac EWS > Source/JavaScriptCore/heap/MarkStack.h:42 > + template <typename T> class WriteBarrierBase; I think we should consider a style guideline of omitting the space between the word template and the <> arguments, by analogy with how there is no space after a function name before its () arguments. > Source/JavaScriptCore/heap/MarkStack.h:65 > + inline void appendValues(WriteBarrierBase<Unknown>*, size_t count, MarkSetProperties = NoNullValues); I believe the inline keyword here has no effect and so should be omitted.
> r=me, but I think you need to find out why this failed on the Mac EWS This builds locally, so I think the EWS may have gotten confused with the other path attached here. I'll watch the bots to make sure. > > Source/JavaScriptCore/heap/MarkStack.h:42 > > + template <typename T> class WriteBarrierBase; > > I think we should consider a style guideline of omitting the space between the word template and the <> arguments, by analogy with how there is no space after a function name before its () arguments. You keep telling me this, and I keep forgetting. Maybe today's the day I remember? > > Source/JavaScriptCore/heap/MarkStack.h:65 > > + inline void appendValues(WriteBarrierBase<Unknown>*, size_t count, MarkSetProperties = NoNullValues); > > I believe the inline keyword here has no effect and so should be omitted. Actually, it opts you in to this warning in the case where you forget to #include the defining header: "scratch.cc:9: warning: inline function ‘void f()’ used but never defined". I think that's a little better than just getting a linker error, since linker errors on Mac OS X can be cryptic.
Committed r86209: <http://trac.webkit.org/changeset/86209>