> I'm not highly skilled in either win32 or NSIS, so excuse me if I say > something stupid but; how can an additional charset imply so much work? > NSIS supports a lot of charsets for different languages already. It feeds > text with a given charset and tells the OS which charset is it.
> Perhaps the problem is not about Unicode but about wide chars? (in which > case, I wonder why they use utf-16 instead of utf-8). Well, win32 has two sets of APIs, one for wide chars which is UTF-16 and the ANSI charset (multi-byte, potentially). There's also two sets of API for the standard C library, the ANSI and the wide-char, which you are also probably aware of. Why? Because the two kinds of strings are arrays of two different types: unsigned short vs char. So you can see the complexity there already. Now in the Unix world wchar_t is a 32-bit type but not so in the Windows world. That's why UTF-16 is the preferred encoding of Unicode in Unicode NSIS. So simply put, it's not a matter of different charsets (Unicode is not an ANSI codepage), it's a matter of different types altogether. - Jim -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
