On Sun, Sep 28, 2008 at 05:30:10PM +0200, Lucas Nussbaum wrote: > On 28/09/08 at 15:06 +0200, Frans Pop wrote: > > On Sunday 28 September 2008, Adeodato Simó wrote: > > > Isn't this a bug of our upgrading tools, that should prefer to install > > > a new, non-obsolete package A that Provides: B, rather than keeping a > > > no-longer-present-in-the-lists B installed? > > > > Possibly. But as long as those tools don't support that it will have to be > > solved in the packages themselves. > > I'm wondering how many cases like that are still in lenny. That is: > - binary package in etch > - that is not in lenny > - that is Replaced/Provided by another package You can use UDD to find it out. I failed to create something fast, but here is my code (it's too slow to output something):
select package from packages_summary where (
release='etch'
and package not in (select package from packages_summary where
release='lenny')
and exists (
select package from packages where (
release='lenny'
and ARRAY[packages_summary.package] <@
string_to_array(provides, ', ')
)
)
);
Please note that I have almost no experience with SQL stuff. This
is just the basic idea and needs real improvements. (Would be a good
QA test, in my opinion).
>
> On the other hand, lots of libraries are probably in this case.
>
> Frans, did you just run into this bug by luck, or did you specifically
> looked for such cases?
>
> > Feel free to clone the bug to apt/aptitude/whatever if you feel that
> > should be implemented, but I'd think that would be squeeze material, not
> > lenny.
>
> Agreed.
> --
> | Lucas Nussbaum
> | [EMAIL PROTECTED] http://www.lucas-nussbaum.net/ |
> | jabber: [EMAIL PROTECTED] GPG: 1024D/023B3F4F |
>
>
>
signature.asc
Description: Digital signature
