Re: patch

[Petter Reinholdtsen <pere@hungry.com>]

[Tomohiro KUBOTA]
> Such a normalization is not needed if the platform is limited to
> GNU/Linux.  However, I tested FreeBSD and it had different naming
> way for locales (for example, ISO8859-1 in FreeBSD vs ISO-8859-1
> in GNU libc).  Moreover, there are many proprietary UNIX systems.

Sure.  If you are sure it will never mistake one charset for another,
it is OK with me.  There is a IETF registration for charset names.  I
suspect it would be better to accept the names registered there, and
report errors on the others, to make it possible to detect when an
incorrect charset is used.  Or perhaps ask the C library if the
charset is known before trying to guess which charset it is.

Anyway, as long as it is hidden in some library, I do not have to
worry about it. :)

> I cvs-committed it.  I enabled --with-textwrap of ./configure in
> debian/rules because it is not used unless libtextwrap-dev is
> installed (and I expect rare people have installed it now).  For
> people who want to test my patch, just install libtextwrap-dev (and
> libtextwrap1 which is depended) and rebuild cdebconf.

And I uploaded the new version, with this support enabled.  The
Japanese d-i install looked quite good. :)

But I need to know which charset should be selected by default among
the two japanese entries in /usr/share/i18n/SUPPORTED.  UTF-8 or
EUC-JP?  This is the locale that will be included in /etc/locale.gen,
and possibly in /etc/enviroment.

Would you like to add Japanese support to Skolelinux too?  Check
<URL:http://i18n.skolelinux.no/new-language.html> for instructions.