[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Problems with Provides/Replaces/Conflicts


tetex-bin previously only declared

Replaces:... dvipdfm
Provides: ... dvipdfm

but no Conflicts. This has the effect that dvipdfm is not removed when
tetex-bin is installed, and trying to remove afterwards fails because of
some dpkg-divert stuff (see #269235). Note that dvipdfm does no longer
exist in sarge or sid.

So I thought the natural thing would be to add a Conflicts, as described
in Policy, Section 7.5.2. But when testing this, the dist-upgrade from
woody does no longer work, which I don't understand:

bin/bash-2.05a# apt-get dist-upgrade
Reading Package Lists... Done
Building Dependency Tree... Done
Calculating Upgrade... Starting
Starting 2
Investigating libconsole
Package libconsole has broken dep on console-tools-libs
  Considering console-tools-libs 4 as a solution to libconsole 9
  Added console-tools-libs to the remove list
  Fixing libconsole via remove of console-tools-libs
Investigating tetex-bin
  Or group remove for tetex-bin
Package tetex-bin has broken dep on dvipdfm
  Considering dvipdfm 0 as a solution to tetex-bin 0
  Holding Back tetex-bin rather than change dvipdfm
Investigating tetex-base
Package tetex-base has broken dep on tetex-bin
  Considering tetex-bin 0 as a solution to tetex-base 3
  Added tetex-bin to the remove list
  Fixing tetex-base via remove of tetex-bin

dvipdfm itself has the following Dependency lines:

Depends: tetex-base, libc6 (>= 2.1.97), libkpathsea3 (>= 1.0.7+20000807-6), libpaperg (>= 1.0.4), libpng2, zlib1g (>= 1:1.1.3)
Suggests: gs

sarge's tetex-base declares a Replaces for dvipdfm (which is correct),
and a Conflict with woody's tetex-bin (which is also necessary), and the
new tetex-bin depends on the new tetex-base.

Does anybody have an idea why apt decides "Holding Back tetex-bin rather
than change dvipdfm"? tetex-bin also Replaces/Conflicts/Provides:
texdoctk, and there's no problem with it. There's a similar problem with
cweb, which still exists in sarge. How could I debug this?

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

Reply to: