[JSC] Move calling convension flags to WTF
Created attachment 281100 [details] Patch
Comment on attachment 281100 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=281100&action=review r=me with comments. > Source/JavaScriptCore/ChangeLog:3 > + [JSC] Move calling convension flags to WTF typo: convension => convention > Source/JavaScriptCore/ChangeLog:8 > + Due to some calling convension flags and JIT_OPERATION falgs, MathCommon.h inclues MacroAssemblerCodeRef and JITOperations.h. ditto. > Source/JavaScriptCore/ChangeLog:10 > + As the same to other calling convension flags like JSC_HOST_CALL, we move these flags to WTF. I think this should be: As with other calling convention flags like JSC_HOST_CALL, these flags should be in WTF
Comment on attachment 281100 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=281100&action=review Thanks!! >> Source/JavaScriptCore/ChangeLog:3 >> + [JSC] Move calling convension flags to WTF > > typo: convension => convention Oooops, fixed. >> Source/JavaScriptCore/ChangeLog:8 >> + Due to some calling convension flags and JIT_OPERATION falgs, MathCommon.h inclues MacroAssemblerCodeRef and JITOperations.h. > > ditto. fixed. >> Source/JavaScriptCore/ChangeLog:10 >> + As the same to other calling convension flags like JSC_HOST_CALL, we move these flags to WTF. > > I think this should be: > > As with other calling convention flags like JSC_HOST_CALL, these flags should be in WTF Looks very nice! Changed.
Created attachment 281101 [details] Patch for landing Check Windows port EWS
Comment on attachment 281100 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=281100&action=review >>> Source/JavaScriptCore/ChangeLog:8 >>> + Due to some calling convension flags and JIT_OPERATION falgs, MathCommon.h inclues MacroAssemblerCodeRef and JITOperations.h. >> >> ditto. > > fixed. Another typo: /falgs/flags/.
Created attachment 281174 [details] Patch for landing Check Windows port EWS
Created attachment 281177 [details] Windows EWS checking Check Windows port EWS, part 2...
Created attachment 281179 [details] Windows EWS checking Check Windows port EWS, part 3...
Created attachment 281360 [details] Windows EWS checking Check Windows port EWS, part 4...
Created attachment 281361 [details] Windows EWS checking Check Windows port EWS, part 5...
Ah OK, it seems that I've just found the existing bug in Windows port. In Windows port, previously, we typedef the JIT operations as follows, typedef EncodeJSValue JIT_OPERATION (*J_JITOperation_E)(ExecState*); And we assumed that JIT_OPERATION is CDECL (And CDECL is __cdecl in Windows port). But it was not correct, since typedef EncodeJSValue __cdecl (*J_JITOperation_E(ExecState*); is syntax incorrect. In previous Windows port, we accidentally define JIT_OPERATION here as "". (Maybe, CALLING_CONVENTION_IS_STDCALL is not defined accidentally in JITOperation.h) So, the above typedef is recognized as typedef EncodeJSValue (*J_JITOperation_E)(ExecState*); I checked it by manually typedefining `typedef EncodeJSValue __cdecl (*J_JITOperation_E)(ExecState*);` and seeing the build failure on Windows EWS. To correct this, we need to define it as follows in Windows MSVC environment. typedef EncodeJSValue (__cdecl *J_JITOperation_E)(ExecState*); So, we need to typedef the function pointer as follows, typedef EncodeJSValue (JIT_OPERATION *J_JITOperation_E)(ExecState*);
Created attachment 281362 [details] Windows EWS checking Check Windows port EWS, part 6... Maybe, the last check.
Comment on attachment 281100 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=281100&action=review >>>> Source/JavaScriptCore/ChangeLog:8 >>>> + Due to some calling convension flags and JIT_OPERATION falgs, MathCommon.h inclues MacroAssemblerCodeRef and JITOperations.h. >>> >>> ditto. >> >> fixed. > > Another typo: /falgs/flags/. Thanks! Fixed.
Committed r202092: <http://trac.webkit.org/changeset/202092>
Committed r202094: <http://trac.webkit.org/changeset/202094>