> I agree totally with your descriptions of the types. However
> you seem to have missed the two points I was making: you
> can put any value you want into a wchar_t - after all, C
> and C++ permit such things.
That's true.
(Similarly, you can insert the "Unicode" character U+FFFF in a text file, or
insert smashed potatoes in the video recorder, as my son did recently :-)
> My other point was that the de facto usage of wchar_t is
> to store a Unicode character, and that one should be
> wary of any code that uses it for something else: it's a
> flag that perhaps other assumptions and unusual behaviors
> have been coded that conflict with good practices in
> internationalization.
Type wchar_t has been designed to ease using "wide" character sets having
more than 256 characters, such as Unicode and CJK national character sets
(e.g. JIS, GB, Big-5).
The first example I've ever seen of wchar.h and wchar_t was a JIS
implementation for the DJGPP complier (http://www.delorie.com/djgpp/).
> By the way, what's the size of a wchar_t on ICL kit?
I am afraid I never heard that before. Next coffee break I'll go down to
marketing dept and ask someone...
Ciao. Marco
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:00 EDT