Re: Should we move texmf.cnf back to /etc?
Atsuhito Kohda <kohda@pm.tokushima-u.ac.jp> wrote:
>> Indeed the default answer is no. However, usually a three-way-merge will
>> work - and there's always the option to view the diff.
>
> I doubt that such diff is really helpful for a user.
> There might be users who install, for example, ptex
> package but don't understand its seetings at all.
You got a point. It would be great if we could let ucf run compare the
files in /etc/texmf/texmf.d/ to the corresponding parts in texmf.cnf,
and only ask questions if the user has edited that now has changed
upstream. Perhaps I'll submit it as an upstream bug, but it's also just
fairy tales.
But still: We cannot prevent the user from editing `kpsewhich
texmf.cnf`. If he's silly enough to do it now, his changes will be
overwritten, if texmf.cnf is ucf-managed, he'll be warned by some stuff
he doesn't understand. What's worse? I don't know, but probably the
overwriting, because if ucf "barfs" at him, he can ask somebody else.
What we should do if we put it under ucf is to make the note in
texmf.cnf more deterring, and more clear. I'd suggest
%%% This file is automatically generated by update-texmf
%
% PLEASE DO NOT EDIT THIS FILE DIRECTLY. It is meant to be generated from
% files in /etc/texmf/texmf.d/.
%
% While changes made by users will not be overwritten, they will cause
% you trouble. You will be shown the differences between the edited and
% the newly created file. We will try to merge our and your changes, but
% that might not always work, and you will probably have to edit again.
%
% Therefore, if you want a smooth upgrade, please edit the files
% in ${CNFDIR}, or create an additional one (with the extension '.cnf'),
% and invoke update-texmf.
%
%%%
>> If they edit /etc/texmf/texmf.cnf, but the one in /var/lib/texmf is
>> used, nothing happens at all (and we're getting a bug report, or some
>> TeX package is loosing a user). If they edit `kpsewhich texmf.cnf`, they
>> are not even asked, but their customization is silently overwritten.
>
> Well, the latest tetex has already renamed /etc/texmf/texmf.cnf
> as /etc/texmf/texmf.cnf.dpkg-old if the one in /var/lib/texmf
> was used.
Yes. But I bet if somebody ask in a mailing list or in usenet, there
will be a good-willing guy who explains kpsewhich to him, and then he'll
end up editing /var/lib/texmf. Given that he's dum^Winexperienced enough
to overlook the "do not edit" note.
>> > Right, but I suspect dpkg also has enough information
>> > of conffiles (md5sum?) so this feature could be a bug
>> > of dpkg.
>>
>> No, if you upgrade from woody, dpkg knows nothing about 05TeXMF.cnf,
>> only about 05TeXMF.
>
> Yes of course. But dpkg know about 05TeXMF so if dpkg
> supports some mapping(?) feature and could map the
> info/data of 05TeXMF to 05TeXMF.cnf then dpkg might
> behave more smart manner, IMHO. (yes, this is yet a
> kind of fairy tale at present.)
Ah, you meant it that way. Well, if I recall some sidenotes on dpkg
development I read on -devel recently, it doesn't seem that submitting
wishlist bugs to dpkg is very promising currently, even if they come
with a patch (what probably won't happen).
> Anyway I might be reluctant to use the ucf for texmf.cnf
> but not object to it.
>
> I guess Hilmar's comment if any might help Frank ;-)
Hey Hilmar, help here!
Regards, Frank
--
Frank Küster, Biozentrum der Univ. Basel
Abt. Biophysikalische Chemie
Reply to: