[Aptitude-devel] Bug#246672: aptitude: quit directly instead of pressing a key to continue

[Manuel A. Fernandez Montecelo]

Control: tags -1 + moreinfo


Hi,

2004-04-30 11:10 martin f krafft:
>Package: aptitude
>Version: 0.2.14-3
>Severity: wishlist
>
>When aptitude finished installing packages, it would be nice to have
>it say "Please press 'q' to quit or any key to continue". ctrl-c
>seems to work, but it's not very nice to just SIGINT the running
>process like that, especially because the terminal sometimes gets
>messed up then.


2011-12-14 08:08 Daniel Hartwig:
>> When aptitude finished installing packages, it would be nice to have
>> it say "Please press 'q' to quit or any key to continue". ctrl-c
>> seems to work, but it's not very nice to just SIGINT the running
>> process like that, especially because the terminal sometimes gets
>> messed up then.
>
>Whilst it is possible to SIGINT at that point, there are subtle
>reasons why you might not want to do that.  See #429388 for some
>discussion of the issue.
>
>With aptitude's current state tracking it does not look like it is
>possible to avoid the cache reload (i.e. exit quickly) after running
>dpkg and keep the cache consistent.

There are some other reasons, like aptitude being able to detect
packages to be "upgraded"/"downgraded" once the action is performed.

If a package is marked for upgrade and was indeed upgraded by dpkg but
aptitude doesn't refresh the information in its database (we had bug
#721426 related with that), and then "apt update" happens somewhere, the
package will be marked for upgrade the next time that aptitude is
fired-up.

So basically aptitude needs to re-read the status of the system and
write back its own database, which is what takes longer when coming back
to curses mode (which is what the merged bug #296726 complains about).


I don't know if your complaint is more the speed side of reloading the
cache, or that you just don't see the need and don't want curses to be
restored just for quitting.  In the latter case, maybe it would not be
very difficult to implement what you ask, but the reloading of the cache
would continue to be there.


Cheers.
-- 
Manuel A. Fernandez Montecelo <manuel.montezelo at gmail.com>