Re: UTF8 locales

[Bruno Haible <haible@ilog.fr>]

Michael Sobolev writes:

> I am trying to build ru_RU.UTF-8 locale using information that comes with my
> glibc 2.1.2.  The command I use is:
> 
>     localedef -v -i ru_RU -f UTF8 ru_RU.UTF-8
> 
> I receive a lot of messages like:
> 
> ru_RU:3018: unknown character in field `yesexpr' of category `LC_MESSAGES'

glibc-2.1.2 does not yet "really" support multibyte locales. When creating
the locale, it replaces such strings containing "unknown characters" by
empty strings.

When I created a de_DE locale, I tried various things, and only the
following workaround finally did it: Create a copy ru_RU.fake of ru_RU,
replace all cyrillic strings with Ascii strings of the correct multibyte
length (remember to allocate two bytes for every cyrillic character), then
create the locale, and then replace the strings by the multibyte
representation of the desired cyrillic words, using Emacs. In the German
locale, there was only one error message that bothered me, a month name.
For the russian locale, it's certainly a tedious task.

Very recent glibc snapshots contain a new implementation of `localedef',
which should work with multibyte locales.

                          Bruno
-
Linux-UTF8:   i18n of Linux on all levels
Archive:      http://mail.nl.linux.org/lists/