Bug#781990: apt: Removing virtual packages doesn't work as you would expect
Control: found -1 2.5.0
On 2015-04-05 19:40:22 -0700, d3fault wrote:
> `apt-get install <virtual-package>` && `apt-get remove <virtual-package>`
> does not remove the packages that were implicitly installed, and should.
This has just happened to me, and I got very confused, in particular
because <virtual-package> was part of a long list of packages (to
rebuild a package). This makes some answers to
https://askubuntu.com/questions/247549/is-it-possible-to-undo-an-apt-get-install-command
fail (though they can also fail for other reasons).
A simple testcase from what I got:
# apt install dh-sequence-gir
Note, selecting 'gobject-introspection' instead of 'dh-sequence-gir'
The following additional packages will be installed:
python3-mako python3-markdown
The following NEW packages will be installed:
gobject-introspection python3-mako python3-markdown
(I've removed unrelated output.)
In the /var/log/apt/history.log file:
Commandline: apt install dh-sequence-gir
Install: python3-mako:amd64 (1.1.3+ds1-3, automatic), gobject-introspection:amd64 (1.72.0-1+b1), python3-markdown:amd64 (3.3.7-1, automatic)
One can see that gobject-introspection doesn't have the "automatic" tag.
# apt purge dh-sequence-gir
Note, selecting 'gobject-introspection' instead of 'dh-sequence-gir'
but nothing is done, and gobject-introspection is still regarded as
manually installed.
Note that in both cases, because of a lot of output, I did not see the
"Note, selecting 'gobject-introspection' instead of 'dh-sequence-gir'"
message (it was unexpected, so I focused on the "will be installed"
information).
On 2016-10-09 17:16:26 +0100, Eric Curtin wrote:
> PR open https://github.com/Debian/apt/pull/24 fixes this
Unfortunately it was never merged and became obsolete due to
ABI/API break.
--
Vincent Lefèvre <vincent@vinc17.net> - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)
Reply to: