Portable way to accept unicode keyboard input

[Sam Varshavchik]

On Linux, and probably most modern platforms, when I look at what's in 
wchar_t, I see unicode values. Or, more precisely, UCS-4 in the native byte 
order. However, elsewhere I've been led to believe that this is really 
implementation defined; an implementation might put something else for a 
wchar_t.

So I'm reading my wchar_ts from get_wch(), and my application uses unicode 
to throw text strings around. If wchar_t is not guaranteed to be UCS-4, then 
I'm trying to figure out what's the most portable way to convert keyboard 
input to unicode.

The first thing that came to mind was use wcrtomb() to convert wchar_ts to a 
multibyte sequence, then stack iconv() on top of that, and end up with UCS-4 
values that way. But if I need to do that, shouldn't I just be able to read 
the same multibyte sequence with getch(), instead?

Basically -- what's the most portable approach to reading unicode values 
from the keyboard?

pgp6GCmrd9Ofl.pgp
Description: PGP signature