UTF problem?

[Bernhard Herzog]

Werner Koch <wk@gnupg.org> writes:

> BTW, are you sure Windows uses UTF-8?  I have my doubst:  Most
> functions come in a fooA and fooW version, where the latter takes
> what MS calls Unicode but actually is the UCS-16 (2 bytes per
> characters) encoding which is not suitable for some languages
> becuase Unicode is a 32 bit character set. 

The windows port of GTK uses GTK 1.3 which uses UTF8 on both Unix and
Windows. What Windows API functions use doesn't matter here.

There's a similar Problem on Unix because we currently use GTK 1.2 for
that which doesn't do Unicode. GPG uses UTF8 in batchmode but neither
gpapa nor the GUI part of gpa do any sort of conversion.

One solution would be to use GTK 1.3 on Unix as well, but 1.3 is an
unstable development version.

> IIRC, the gettext version for Windows takes UTF8 and translates it
> to the Windows characters set.  We should go this way, which might
> mean that we have to convert our .po files from whatever encoding
> they use (de.po Latin-1 or -15) to UTF-8 which is just a simple
> shell one liner.

For the windows version of GPA it would probably be best to have the mo
files in UTF8 and a gettext that doesn't do conversions at all. GTK
takes care of any further conversion that may be necessary to display
the text.


  Bernhard

-- 
Intevation GmbH                                 http://intevation.de/
Sketch                                 http://sketch.sourceforge.net/
MapIt!                                               http://mapit.de/