RE: GCC vs. typedef unsigned short XMLCh

[Nikko]

Hi, We had the same problem at my company recently. wchar_t size varies according to the platform and event the compiler on a platform. What we did is that : - we suppresed all wchar_t and wsprintf, etc... stuff - we developed an in-house C++ class handling 2 bytes long characters (we typedef it with #ifdef VISUALC, #ifdef GCCLINUX, etc...). We can convert an XMLCh* to that object or from that object to a 2 bytes char array and assign it to an XMLCh*. It does not handle local page problem yet, we rely on XMLCh* for that.   HTH   Nikko          -----Message d'origine----- De : Endre Magyari [mailto:[EMAIL PROTECTED]] Envoy� : mardi 10 septembre 2002 17:12 � : [EMAIL PROTECTED] Objet : GCC vs. typedef unsigned short XMLCh Dear list,         I'm porting an application using Xerces from WIN32 to Linux/gcc.       I have this piece of code:(a is a DOMString)     -------8><--------------     const wchar_t *buf = a.rawBuffer();     wchar_t * endptr;     unsigned long retval = wcstoul(buf+2, &endptr, 16);     -------8><---------------   Under win32, it seems appropiate because both wchar_t and XMLCh are typedef-ed to unsigned short. However, this is not the case under Linux/G++. wchar_t and wcstoul are defined there as well, but wchar_t becomes typedef-ed to unsigned long. Now you probaly can guess my question... How to convert a from a 16bits long wide char array to a 32 or 64 bits long wide char array, assuming that I would like to use the wcstoul function.   Thank you, Endre Magyari