[prev in list] [next in list] [prev in thread] [next in thread] 

List:       xfree-i18n
Subject:    Re: [I18n][OT] GNU gettext and LANGUAGE variable
From:       Bruno Haible <haible () ilog ! fr>
Date:       2001-04-10 12:32:11
[Download RAW message or body]

Tomohiro KUBOTA writes:

> ~$ LANG= LANGUAGE=zh_TW ls foo
> ls: foo: No such file or directory
> ~$ LANG= LC_CTYPE=de_DE LANGUAGE=zh_TW ls foo
> ls: foo: No such file or directory
> ~$ LANG= LC_MESSAGES=de_DE LANGUAGE=zh_TW ls foo
> ls: foo: ?????????
> 
> This is as if LC_MESSAGES specifies the encoding (and Gettext tried to
> convert Chinese message into ISO-8859-1), while I think the encoding
> must be specified by LC_CTYPE category...

No, the encoding is specified by the LC_CTYPE category. But gettext
refuses to do any translation if the LC_MESSAGES category is set to
"C". This is in order to follow the POSIX or SUSV2 standard; it
specifies the precise output of the programs "wc" or "ls" in the C
locale. If gettext() did pay attention to LANGUAGE even when
LC_MESSAGES=C, it would not be usable in a correct implementation of
"wc" or "ls".

Really the recommended setting is LANG, not LANGUAGE without
LC_MESSAGES.

> I am now working on "man" command for Debian GNU/Linux.

Let's discuss this on linux-utf8, where the groff developers can
participate.

Bruno
_______________________________________________
I18n mailing list
I18n@XFree86.Org
http://XFree86.Org/mailman/listinfo/i18n

[prev in list] [next in list] [prev in thread] [next in thread]