UTF oddity

[Eric Pouech]

Jan Sporbeck wrote:
> Hallo,
> 
> while porting a friends windows code, i encountered the following:
> 
> WideCharToMultiByte returns incorrect string length when using CP_UTF8 
> for conversion
> 
> the following code snippet correctly sets nLen1 to 2 but sets nLen2 only
> to 1
> 
> int nLen1 = WideCharToMultiByte(CP_UTF8, 0, L"Ü", 1, NULL, 0, NULL, NULL);
> // query buffer size
> int nLen2 = WideCharToMultiByte(CP_UTF8, 0, L"Ü", 1, szUtf, nLen1, NULL, 
> NULL);
> // do conversion
> 
> 
> I tested this with Wine <=20030408
> 
> am I missing something?
the L"" constructs are not yet supported while porting code (which I 
assume you're doing)
if you're using a recent gcc, try using the -fshort-wchar option (in 
gcc), it might help
(for a bit of explanation, unicode strings are stored as unsigned short 
in Windows, and by default, unsigned long by gcc, hence some issues)

A+


-- 
Eric Pouech