Unicode and Win32

João Garcia jgarcia@uk2.net
Fri May 30 04:15:00 GMT 2003


Hi,

fernando@lozano.eti.br wrote:
> 
> You're right, unicode is unicode. But beware UTF-8 and UTF-16 :-)

Wrong link... :-)
Sorry. It was irrelevant anyway. UTF-16BE and UTF-16LE seem to depend 
only on the architecture (big endian and little endian), so everything 
should be "in sync" between the OS and the compiler.

But I am not so sure about other things (ie. surrugates). I also don't 
know if those things are relevant for the problem. We shall see...

> > What about System.getProperty("os.name") and "os.version" ? Aren't they
> reliable? If not, they should be fixed and used on your code.

Thanks. I guess I could use them but, since the code is in C++, I 
thougth it would be nice to use the w32 API directly.

And I did find a solution (it wasn't that difficult...): 
GetVersionEx(LPOSVERSIONINFO lpVersionInfo).

lpVersionInfo is a structure. One of its fields holds the platform code 
(value 1 for W9x and value 2 for NT branch). I can use ( platform >= 2 ) 
as a condition to use wchar functions. I hope the code will be able to 
work with platform 3 (whatever that may be). ;-)

A better alternative would be to have Linux dominant on the desktop. :-)

João



More information about the Java mailing list