RE: Language characters

[Owen Taylor <otaylor redhat com>]

On Tue, 2003-07-15 at 11:30, martyn 2 russell bt com wrote:
> > The encoding has to be part of your network protocol. (Either 
> > by passing
> > it, or by defining what the encoding is as part of the protocol.)
> > You can't just have a random stream of bytes in an undefined encoding
> > and tell GTK+ to display it.
>
> This is TRUE :)
>
> The information I receive is coming in from a GIOChannel.  I would expect
> that is UTF8.  Therefore there should not be any translation needed.

GIOChannel has *no* ability to magically guess the incoming text
encoding either. The knowledge of what encoding is involved
has to come from the application/protocol leve.

> If, for example I search for a contact I have which has Russian details, the
> information is displayed perfectly:
> http://gnome-jabber.sourceforge.net/screenshot-russian.png
>
> Therefore, either umlauts are not supported by UTF8 OR the text that comes
> in on the socket is invalid UTF8 (and I admit I do not check it because I
> assume it already is).

UTF-8 certainly supports German.

Likely your Russian text was encoded as correct UTF-8 to start
of with, your German text wasn't.

Regards,
                                                Owen