Mailinglist Archive: zypp-devel (78 mails)

[Jan Kupec <jkupec@xxxxxxx>]

Duncan Mac-Vicar Prett wrote:
> On Friday 09 November 2007 09:35:33 Jan Kupec wrote:
> > Why do we need more information than the info we currently have to do
> > this? Why do we need package history? I can't think of any problem or
> > case. Maybe you missed the *when removing a package* in the mail
> > subject. Or it's just me missing something :O) Please explain.
> >
> > J.
>
> Does your algorithm work for leaf packages that I install and that are
> not required by another packages? lets say libwt, which I use for
> development. As nobody depends on it, it will be removed,

Yes, it could be removed by the algorithm, but that is exactly what i
want from it. Nota bene, the 'zypper remove foo' would work as before!
Would would have e.g. 'zypper remove --remove-unused foo' or something
like that to trigger the algorithm. And yes, that _could_ result in
removing libwt in the end, if no installed package uses it.

But you see my point. Upon removing foo (called with this special
option), i want zypper to check for those unused packages resulting from
removing foo. It will display the install/remove summary and if i don't
like it (wants to remove libwt, which i want to keep), i choose not to
continue and reach my goal by other means. At leasts it will give me a
list of such packages, that's great too.

> but it should only be removed if it
> was dragged by the solver , not if it was a explicit user transaction.

I see! Yes, that would be a 'zypper remove --remove-unused foo' and what
i am basically proposing is 'zypper remove --remove-unused
--ignore-package-history foo'. *Both have sense* IMO!

So it seems we should understand each other right now, so what do you think?

> Anyway, it sucks that rpm can save this metadata by itself.

Can you explain?

Jano

-- 
To unsubscribe, e-mail: zypp-devel+unsubscribe@xxxxxxxxxxxx
For additional commands, e-mail: zypp-devel+help@xxxxxxxxxxxx