Re: synaptic: Cannot change mark from "complete removal" to "removal"

[Philip Chung <philipchung1995@yahoo.com>]

On Mon, 25 Jul 2016 13:56:58 -0700 Kyanos <someanon126@gmail.com> wrote:
> Package: synaptic
> Version: 0.83+b1
> Severity: minor
>
> Dear Maintainer,
>
> If I mark a package for complete removal, I cannot change the mark
> directly to simple removal.
>
> Steps to reproduce:
>
> 1. Select a package.
>
> 2. Mark the package for complete removal.[1]
>
> 3. Mark the package for removal.[1]
>
> [1] This can be done by either right-clicking and selecting the
> appropriate option, going to the Package menu and selecting the option,
> or using a keyboard shortcut. Any of these will yield the same result.
>
> Actual result:
>
> The package is still marked for complete removal. There is still a
> yellow cross on the package mark, and the summary that is shown after
> clicking "Apply" lists the package under "To be completely removed
> (including configuration files)".
>
> Expected result:
>
> The package should be marked for removal. There should be a red cross on
> the package mark, and the summary that is shown after clicking "Apply"
> should list the package under "To be removed"

I'm revisiting this old bug report of mine and I confirm that this still 
occurs in the latest version of Synaptic.

Upon further investigation, I believe this can actually be traced to APT 
itself, or more specifically libapt-pkg. Synaptic uses the pkgDepCache 
class from libapt-pkg to handle these states, and the 
pkgDepCache::MarkDelete method is called to mark a package for removal 
or complete removal.

In the implementation of that method (source file apt-pkg/depcache.cc), 
there is a particular check that suggests a deliberate design decision 
to disallow moving from purging (complete removal) to simple removal but 
allow moving in the other direction:

> // Check that it is not already marked for delete
> if ((P.Mode == ModeDelete || P.InstallVer == 0) && 
>     (Pkg.Purge() == true || rPurge == false))
>    return true;

I've CC'ed the APT team on this e-mail to get their input. If there are 
good reasons for this decision then I am happy to have this bug closed 
as "won't fix." (I don't think there is a compelling reason for Synaptic 
to work around this design decision in APT.)

Kyanos