Harald Alvestrand <[EMAIL PROTECTED]> wrote:
>>[qrczak ~]$ find /usr/share/man/pl -type f | wc -l
>> 149
>>
>>They are ISO-8859-2. And there is more of them, I don't know how much
>>(my system is far from any vanilla distribution).
>
>the point may be that the subdir under /usr/share/man should be named by
>something that indicates character set (locale?) rather than just language?
>
>when we know where to find out what it is, converting is a bit easier....
Yes. Some `man` implementations knows about MANPATH=
environment, where MANPATH= is substitution of $LANG's fields,
like NLSPATH=":%N.cat:/usr/share/locale/%L/%N.cat" :
%N The value of the name parameter passed to catopen().
%L The value of the LC_MESSAGES category.
%l The language element from the LC_MESSAGES category.
%t The territory element from the LC_MESSAGES category.
%c The codeset element from the LC_MESSAGES category.
%% A single % character.
So, it is possible to call `groff -T%c` .
However, this is so called 'outbound' encoding-controlling method.
Another way is to define manpage's charset "inline" in text, like
in MIME messages :
Content-Type: text/plain; charset="koi8-r"
Subject: Re: =?KOI8-R?Q?=C1?=
or in glibc2 .po messages. Unfortunately, I don't know portable
method.
--
-=AV=-
-
Linux-UTF8: i18n of Linux on all levels
Archive: http://mail.nl.linux.org/lists/
