>>35
wchar.h here (for wchar_t).
As for Win32, in MSVC, MS had the brilliant idea of making a type(macro TCHAR) which would represent normal characters(char) or multibyte ones in one case, and "UNICODE" (UTF-16, WCHAR, or wchar_t) in the other case (depends if UNICODE macro is defined). Similar _t* macros are defined for various API calls and libc functions. The idea is that you can build both ANSI and UNICODE applications form the same code. ANSI applications can be useful if compatibility with even old Win32 platforms is desired (9x had crappy unicode support), and in fact, most applications these days tend to be ANSI apps and only a few properly support UNICODE.
It might be worth mentioning that even if this is a MSVC extension, there is nothing preventing you from implementing a setup like this in portable C, as long as your C library has unicode versions of standard libc functions.