Re: Gettext and LC_MESSAGES
> From: Changwoo Ryu <cwryu@idis.co.kr>
> Subject: Re: Gettext and LC_MESSAGES
> Date: 24 May 2000 12:00:47 +0900
>
> > Tomohiro KUBOTA <kubota@surfchem0.riken.go.jp> writes:
> >
> > [..]
> > > There should be two discussions: (1) Why setlocale(LC_MESSAGES, "de")
> > > fails? Should we improve locale database to have "de" locale?
> >
> > Because the glibc maintainer decided so. There is no locale db named
> > as "de" in glibc. But the message catalogs are stored in "de"....
> >
> > > (2) Why German message is displayed while setlocale() fails? Is this
> > > a bug of gettext or glibc?
> >
> > Gettext does not depends on the success of setlocale(). I guess this
> > is because some systems lack setlocale().
gettext depends on the success of setlocale(), in most case.
In case of that the system does not have setlocale() function,
it try to get locale information using own library.
> I understand. However, I don't understand why....
Mr. Changwoo already wrote reasons :)
> From: "Petr Kolar" <Petr.Kolar@vslib.cz>
> Subject: Re: Gettext and LC_MESSAGES
> Date: Wed, 24 May 2000 13:10:38 +0200
>
> > If I understand it well, you must set that you're using German language
> > in Germany (de_DE), while most catalogs are just in German language. But
> > if you would have some program containing messages eg. about the capital,
> > money, etc. (which depends on the state), you should have catalogs de_DE
> > (with Berlin and marks) de_AT (with Wien and schilings), de_CH (with Bern
> > and franks).
>
> Yes, so, setlocale(LC_MONETARY, "de") should fail while
> setlocale(LC_MONETARY, "de_DE") should work. However,
> I am discussing about LC_MESSAGES, not LC_MONETARY.
> (I don't know the name of the capital should be treated
> with locale system.)
? LC_MONETARY ?
> > If you set de_DE and the catalog is not found in de_DE, it is looked for
> > in de. So it is better to use de_DE or german or deutsch (the last two
> > are alises to de_DE.ISO-8859-1).
>
> If gettext works well for LANG=de, many people would think
> LANG=de is correct. I posted this thread because I feel many
> European-language-speakers prefer XFontStruct (conventional
> text output faculty of X11) to XFontSet (internationalized text
> output of X11R5) because they think XFontSet doesn't work with
> 8bit characters. I suspect that this is because they set LANG
> in wrong way, like "de".
gettext and XFont* are different issue. Separate.
However, I agree that we should use not XFontStruct but XFontset.
Regards,
-- GOTO Masanori
Reply to: