Removals keep removing dependencies if removal of a package fails
Bug #1844634 reported by
Julian Andres Klode
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
apt (Debian) |
Fix Released
|
Unknown
|
|||
apt (Ubuntu) |
Fix Released
|
High
|
Julian Andres Klode | ||
Disco |
Won't Fix
|
Undecided
|
Unassigned |
Bug Description
[Impact]
Assuming packages A and B, with A depending on B. A has a failing prerm script.
Expected behavior:
- A fails to be removed, A and B stay unchanged
Actual behavior:
- A fails to be removed
- B is still removed
This might crash their system (e.g. if A is systemd and B is libsystemd0).
[Test case]
See Impact. An automated version of the test case (test-apt-
[Regression potential]
We now abort earlier in removal failures, that might be harder to recover from or not, nobody really knows.
description: | updated |
Changed in apt (Ubuntu): | |
importance: | Undecided → High |
status: | New → Fix Committed |
assignee: | nobody → Julian Andres Klode (juliank) |
Changed in apt (Debian): | |
status: | Unknown → New |
Changed in apt (Debian): | |
status: | New → Fix Released |
Changed in apt (Ubuntu Disco): | |
status: | Fix Committed → Won't Fix |
To post a comment you must log in.
This bug was fixed in the package apt - 1.9.4
---------------
apt (1.9.4) experimental; urgency=medium
* CMake: Pass -Werror=return-type to gcc
* CMake: Produce a fatal error if triehash could not be found
* apt.systemd.daily: Do not numerically check if intervals equal 0
(LP: #1840995)
* srvrec: Use re-entrant resolver functions
* Pass --abort-after=1 to dpkg when using --force-depends (Closes: #935910)
(LP: #1844634)
* Fix use of GTest to adjust for GTest 1.9
-- Julian Andres Klode <email address hidden> Thu, 19 Sep 2019 11:13:47 +0200