Bug#209395: teTeX: language.dat mislinked
Atsuhito Kohda <kohda@pm.tokushima-u.ac.jp> schrieb:
> From: frank@kuesterei.ch (Frank Küster)
> Subject: Re: Bug#209395: teTeX: language.dat mislinked
> Date: Fri, 31 Oct 2003 11:08:11 +0100
>
>> I don't know Hilmar's ideas ;-), but yes, it is. One can combine debconf
>> sections and manually edited sections in one conffile. /etc/locale.gen
>> and /etc/environment are examples everybody should have.
>
> I didn't investigate these, but before you are doing
> too much, I'd like to note,
>
>> Since language.dat doesn't depend on the order of the entries, on could
>> easily do it like this. The only thing one cannot do manually AFAIS is
>> unload patterns that are loaded by debconf - but that doesn't make much
>> sense, anyway.
>
> I myself accept your reasoning that it doesn't make much
> sense but there will be members who don't accept your
> reasoning and say it is clear that tetex violates the policy
> because it doesn't preserve local modification.
Debconf would check wether the "debconf section" signature has been
deleted. If yes, then it would do nothing. If it is still there, it will
only change stuff between the Begin/End pair. So if one wants to add a
pattern that is not shipped with Debian, he can either generate a
completely new language.dat or simply add it outside the debconf
section. If he wants to exclude something he had previously selected
with debconf, again he has two options: dpkg-reconfigure (which would be
the easy, sensible way) or generating an a-la-carte language.dat from
scratch, without the debconf section. This would never be touched.
We should, however, still keep the debconf question wether to manage
language.dat with debconf. And if debconf find's that, although the
question had been answered with yes, there's no debconf section left,
then we migth think about asking again.
> As I said before I didn't care where to put language.dat
> but I still guess to put it in /var keeps tetex in a much
> safer state. (I didn't mean I objected to put it in /etc
> but it would be very dangerous, IMHO.)
Dangerous only because of possible inconsistencies after local
modification, or more?
Bye, Frank
--
Frank Küster, Biozentrum der Univ. Basel
Abt. Biophysikalische Chemie
Reply to: