RESOLVED FIXED 56346
Made JSWrapperObject and subclasses moving-GC-safe
https://bugs.webkit.org/show_bug.cgi?id=56346
Summary Made JSWrapperObject and subclasses moving-GC-safe
Geoffrey Garen
Reported 2011-03-14 17:21:10 PDT
Made JSWrapperObject and subclasses moving-GC-safe
Attachments
Patch (11.46 KB, patch)
2011-03-14 17:28 PDT, Geoffrey Garen
oliver: review+
Geoffrey Garen
Comment 1 2011-03-14 17:28:13 PDT
Darin Adler
Comment 2 2011-03-14 17:35:06 PDT
Comment on attachment 85745 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=85745&action=review > Source/JavaScriptCore/ChangeLog:21 > + no longer does so for us. Also added an ASSERT to catch a latent bug, > + where DatePrototype stomped on its base class's anonymous slot. Hard-coded What was the symptom of this “stomping”? > Source/JavaScriptCore/ChangeLog:30 > + data member which was not always visited during GC. Now, we only use the > + data member, and we always visit it. I don’t see any change to the code to always visit the data member. Is the change just the change to StructureFlags? If so it might be better to say so outright. > Source/JavaScriptCore/ChangeLog:34 > + * runtime/NumberObject.cpp: > + (JSC::NumberObject::NumberObject): > + * runtime/NumberObject.h: Removed meaningless declaration. NumberObject.cpp here doesn’t have a comment. > Source/JavaScriptCore/runtime/BooleanObject.cpp:31 > +BooleanObject::BooleanObject(JSGlobalData&, NonNullPassRefPtr<Structure> structure) > + : JSWrapperObject(structure) Do we want to change the constructor arguments eventually to not even pass the global data?
Geoffrey Garen
Comment 3 2011-03-14 17:40:01 PDT
Geoffrey Garen
Comment 4 2011-03-14 17:41:42 PDT
(In reply to comment #2) > What was the symptom of this “stomping”? Updated ChangeLog to explain the symptom. > I don’t see any change to the code to always visit the data member. Is the change just the change to StructureFlags? If so it might be better to say so outright. Yes. Updated ChangeLog to mention the flag. > NumberObject.cpp here doesn’t have a comment. Fixed. > Do we want to change the constructor arguments eventually to not even pass the global data? Yes.
Note You need to log in before you can comment on or make changes to this bug.