Re: Chinese and Japanese (and Korean?) breakage in GNOME2

[Jody Goldberg <jody gnome org>]

On Mon, Jan 28, 2002 at 07:09:37AM +0900, Yukihiro Nakai wrote:
> > Owen Taylor <otaylor redhat com> writes: 
> > > Well, actually, if we want to support non-GNU gettext, or older GNU
> > > gettext, we can't rely on bind_textdomain_codeset(); instead we have
> > > to:
> > > 
> > >  a) ship UTF-8 .po files
> > >  b) call bind_textdomain_codeset() if available so that strings don't
> > >     get translated back from UTF-8 to the encoding of the locale.
> > > 
> > > This is what GTK+ does.
> > 
> > I think that's too complicated for people to get right (involves
> > configure checks, etc.) and translators seem to dislike the UTF-8 po
> > files thing.
> > 
> > We probably need to either automate this, or require new GNU gettext.
> > 
> 
> So what is the conclusion?
> 
> There is no reason to reject UTF-8 .po files if it works completely with GNOME2.

It sounds as if people are saying that there exist versions of
gettext that can not map non-utf8 po file -> utf8 on the fly.  If
this is true there we can either require better gettext routines or
ensure that the po files are in utf8.  The latter seems like a
better solution.