Bug#267299: Installation of tetex-base (2.0.2b-3) fails.
"Maciej A. Patelczyk" <maf@rudy.who.vectranet.pl> wrote:
> On Mon, 30 Aug 2004, Frank Küster wrote:
>
>> But this cannot have been true at the moment the bug occured, because
>> then it would not even have tried to generate those formats.
>>
>
> Ugh, can't tell you. I fixed it manually and it works now.
> But the problem occured for the first time while i was uprading
> the system 2-3 weeks ago. It could be that for the first time it
> was in tetex-bin and in next upgrade it occured while setting up
> tetex-base. So packages were installed but not configured.
The time is what I expected; and it will indeed occur every time one of
the packages tries to configure. But of course not when tetex-bin is
purged, as it was when you made the bug report.
> Hm tetex-base's postinst is using file that is installed by tetex-bin ?
> Is it ok ?
It's a little more complicated - the file comes from tetex-base, and is
used by /usr/bin/fmtutil (from tetex-bin), which is called by
tetex-bin's postinst, and also by tetex-base's postinst if tetex-bin is
installed.
The bug in the provided file has been fixed in tetex-base_2.0.2a-3, but
because language.dat is a conffile, it might still persist on users'
systems. Therefore we have introduced code in tetex-bin's postinst
script to fix the configuration file.
Unfortunately I didn't notice that if tetex-bin is installed, and
tetex-base is upgraded simulataneously, tetex-base's postinst is
executed first, before the file is corrected, and will fail. Therefore
the correcting code will never be called...
Regards, Frank
--
Frank Küster, Biozentrum der Univ. Basel
Abt. Biophysikalische Chemie
Reply to: