Bug#377581: tetex-base: Creating of missing formats failed with dpkg --configure started in /usr/bin
Oliver-Mark Cordes <firstname.lastname@example.org> wrote:
> On Mon, 10 Jul 2006, Frank Küster wrote:
>> Oliver-Mark Cordes <email@example.com> wrote:
>>> On Mon, 10 Jul 2006, Frank Küster wrote:
>>>> Strange that it says "(/usr/bin/amstex", on my system it's "(./amstex".
>>>> But anyway.
>>> Does tex translate any filename into realnames?
>> Pardon? What do you mean with this question?
> From my experience in many TeX documents I found something like
> \include amstex which includes the first file of amstex.* (or
> priviledged filename extensions). In the *tex output you always see
> which real file is included,right?
Yes, that's true. Substitute "realnames" by "name with qualified path"
and I would have said "yes" at once. I'm not sure under which
circumstances "qualified path" means "full absolute path", or when a
relative path like "./something..." can occur.
>> I still don't see a reason to change the current directory to
> Okay, from the beginning: When I did my upgrade my current directory
> was /usr/bin
Which is what I think should never happen. I understand what was going
on when you did that, why it went wrong and all. But I don't see why
your current directory should be /usr/bin at all, ever. And I'm asking
whether it's worth to fix a bug that's only triggered by a behavior of
the local admin that (so far) makes no sense to me.
> and the installation/configuration with dpkg
> failed. apt-get -f
> install started from the same directory worked.
That's interesting; it seems apt-get changes dir itself (although I
can't find it in the sources).
> Yes, I think the configuration should take care of the directory from
> which the configuration is started until this BUG can be closed. The
> idea from Florent to use a temporary directory in /tmp is easy to
> implement and should work. I don't have the time to test it but it
> really sounds like a good idea.
Yes, it does. People go ahead coding. Although I'd appreciate more if
you'd give teTeX's license problems some care...
Single Molecule Spectroscopy, Protein Folding @ Inst. f. Biochemie, Univ. Zürich
Debian Developer (teTeX)