[Aptitude-devel] Bug#531030: Re: Aptitude: "why" gives invalid reasons (e.g. using packages that are not installed).

[Daniel Hartwig]

On 27 February 2012 18:51, Niels Thykier <niels at thykier.net> wrote:
>
> """
>  Explains the reason that a particular package should or cannot be
>  installed on the system.
> """
>
> Personally I read that as:
>
> """
> "why" explains to you why package X is installed on your system.
> "why-not" explains to you why package X cannot be installed on your system.
> """
>
> So I can see the reason for "why-not" giving me a package not installed
> on my system.  But I have difficult in see what use I could have of
> "why" if it ignores my installed packages.
>

Perhaps a package was previously auto-installed, now it's dependencies
have been removed and you would like some suggestion as to why it was
installed in the first place.

> For me, the "why" command would much more useful if it showed me the
> chain from a manually installed package to the package I asked for.
> Admittedly I do not feel strongly about this as I stopped using
> "aptitude why", so feel free to ignore this.
>

The chain will start at a manually installed package if possible, only
if it can't do that are not installed packages considered.

>> "aptitude why" is not a tool to locate orphaned packages.
>>
>
> I was not using "aptitude why" to locate orphaned nor unused
> packages[1].  I was using to double check suggestions from
> deborphan/$tool.  Most of these tools rely on guesses/heurestics to
> produce its results and I found it can be useful to double-check what
> packages "keeps" it.
>
> These days I just feed the results to aptitude purge and see "what falls
> off".
>

The package states are displayed, so there is enough information there
to see whether a package is orphaned or not.  Though admitedly because
this is not as easy as if "aptitude why" simply returned no reason
for the orphaned packages (or put a big notice at the top/bottom of
it's output.)

>
> [1] Btw, I hope we both agree on "orphaned" here means
> "unused"/"discovered by deborphan" and not "orphaned" as in "package
> without (active) maintainer".
>

Yes, that is how I understood you.