Bug#233886: Bug #233886: Re: tetex-base upgrade from woody pretend I have modified /etc/texmf/mktex.cnf

[Frank Küster <frank@debian.org>]

Bill Allombert <allomber@math.u-bordeaux.fr> wrote:

> Hello, I have relaunched my tests and it looks like the problem
> was fixed for the files below:
>
> /etc/texmf/texmf.d/95NonPath.cnf
> /etc/texmf/texmf.d/05TeXMF.cnf
> /etc/texmf/texmf.d/45TeXinputs.cnf
> /etc/texmf/texmf.d/75DviPS.cnf
> /etc/texmf/texmf.d/85Misc.cnf

Yes, these are in tetex-bin. The others are in tetex-base, which has not
yet been uploaded. 

> On the other hand, I get asked a question by ucf about /etc/texmf/texmf.cnf
> while I did not change it.

Thank you, I also noticed this a couple of minutes ago. I'll explain you
why, for the ones that know yet there's a question below.

texmf.cnf is not a content of the tetex-bin package proper, it is
generated from the files in texmf.d. However, we got complaints that
this generated file was silently overwritten after a user had made
changes in it. Well, we had warned him, but anyway. It seemed easy to
use ucf for this file, too, and every version that is on the system from
that time on will be registered with ucf. But I simply forgot about the
old versions.

I will have to get the md5sums of old versions of the file. This is no
problem for current woody, which is of major concern for upgrades. But
it is harder for other versions: Either pre-woody, or everything after
it. To be really sure to not prompt without need, we would have to
figure out the md5sum texmf.cnf after every change we made to files in
texmf.d. Furthermore, we would have to consider additions by other
packages, and every combination.

Therefore, please:

Everybody reading this should 

a) run "md5sum /etc/texmf/texmf.cnf" on every system he/she thinks might
   have an unusual configuration, status with regard to
   testing/unstable/backports, and compare the result with 

103eb2de6df60ea2153095a16927a60f 1.0.7+20011202-7.1 (current woody)
09b4a79e09833a5207fa858ea3d7c834 2.0.2-13           (next version)

as well as with the contents of /usr/share/tetex-bin/texmf.cnf.md5sum
which should come to existence in the next upload. If you discover
something new, please send it to me.

Regards, Frank
-- 
Frank Küster, Biozentrum der Univ. Basel
Abt. Biophysikalische Chemie