This is the mail archive of the
libc-help@sourceware.org
mailing list for the glibc project.
Re: the description of LANGUAGE
- From: Petr Baudis <pasky at suse dot cz>
- To: Zhang Le <r0bertz at gentoo dot org>
- Cc: libc-help at sourceware dot org, drepper at redhat dot com, roland at frob dot com
- Date: Thu, 4 Dec 2008 01:20:02 +0100
- Subject: Re: the description of LANGUAGE
- References: <20081130133656.GA22220@adriano.hkcable.com.hk> <20081130134717.GB22220@adriano.hkcable.com.hk>
Hi,
On Sun, Nov 30, 2008 at 09:47:17PM +0800, Zhang Le wrote:
> Sorry, please take this one. Just did a little modification.
>
> LANGUAGE:
> When determining the language (e.g. en or en_US) of a locale category (e.g.
> LC_MESSAGES), if the locale category is itself already set to a valid locale
I think you should make clear this is only for the translation
purposes, and it _usually_ only affects LC_MESSAGES - exceptions are
rare, e.g.:
http://google.com/codesearch?hl=en&lr=&q=dcgettext.*LC_[^M]
> name (note, must be a locale name, a languge name won't work. e.g. 'en_US' is
> ok, but 'en' is not) in a system, no matter it is set explicitly or implicitly
> using LC_ALL or LANG, LANGUAGE's value will override this locale category's
> original setting. LANGUAGE's value is a colon separated list of language codes.
This is unnecessarily entangled - the locale category must simply be
set to a valid locale different than C (POSIX). Also, LANGUAGE's value
is a list of locale names, not language codes.
> The list will be scanned from left to right, the first valid language code will
> take effect.
This is not quite correct, only the first valid language code with
translation available will take effect.
--
Petr "Pasky" Baudis
People who take cold baths never have rheumatism, but they have
cold baths.