JSChar is typedef'd as either "unsigned short" or "wchar_t" based on a number of platform checks. This is problematic for non-standard ports. To solve this, a way to determine which typedef to use without platform checks is required.
Created attachment 290384 [details] JSStringRef change
Comment on attachment 290384 [details] JSStringRef change View in context: https://bugs.webkit.org/attachment.cgi?id=290384&action=review > Source/JavaScriptCore/API/JSStringRef.h:40 > +#if __WCHAR_MAX__ > 0x10000 Don't we need to check if __WCHAR_MAX__ is #defined?
(In reply to comment #2) > Don't we need to check if __WCHAR_MAX__ is #defined? Looking more closely at gcc and msvc, a better check would be as follows: > #if !defined(_NATIVE_WCHAR_T_DEFINED) && (!defined(__WCHAR_MAX__) || (__WCHAR_MAX__ > 0xffffU)) _NATIVE_WCHAR_T_DEFINED being msvc specific __WCHAR_MAX__ being gcc specific However, I'm not familiar with mac implementation, it seems according to apple's libc (https://opensource.apple.com/source/Libc/Libc-825.24/include/_types.h) it is not guaranteed to exist: > #ifdef __WCHAR_MAX__ > #define __DARWIN_WCHAR_MAX __WCHAR_MAX__ > #else /* ! __WCHAR_MAX__ */ > #define __DARWIN_WCHAR_MAX 0x7fffffff > #endif /* __WCHAR_MAX__ */
Created attachment 290405 [details] updated patch
Comment on attachment 290405 [details] updated patch r=me
Created attachment 290419 [details] updated patch Covers a missing case for RVCT. Adds description for each set of compiler options.
Created attachment 290420 [details] updated patch
Comment on attachment 290420 [details] updated patch r=me
Comment on attachment 290420 [details] updated patch Clearing flags on attachment: 290420 Committed r206734: <http://trac.webkit.org/changeset/206734>
All reviewed patches have been landed. Closing bug.
ChangeLog entry does not match patch contents!
(In reply to comment #11) > ChangeLog entry does not match patch contents! How so?