Re: Conflicts in Debian : a bit of data
On Mon, 2007-03-12 at 13:44 +0100, Lucas Nussbaum wrote:
> On 12/03/07 at 12:38 -0000, Regis Boudin wrote:
> > Oops... Just noticed I had a bug and was testing versions the wrong way
> > around, so all the cases where the version of a package is the same as in
> > the declared conflict gave the wrong result. I will upload pages with
> > correct data this evening.
>
> Could you also add a column with a "conflict number", so it's easy to
> just see how many useless conflicts there are in each case ?
>
> Or just display the total number of conflicts on top ?
Here we go. Still at http://www.imalip.info/debian/ . I also added Slink
to the checked releases.
Overall, it looks like we have :
- 106 pre-slink conflicts
- 124 pre-potato
- 506 pre-woody
- 920 pre-sarge
- 1152 pre-etch
Also added the list of conflicts against packages not found in any
release. Some are packages removed before slink, some are non-i386
packages (aboot), many are completed transitions (python,...), some are
virtual-package-like (libdb-dev).
Some notes : a lintian test can hardly be done, as it would require
checking Packages files from old distributions. The
package-doesn't-exists cases would need more different cross-checks to
know when the package was in, and what sort of delay is reasonable to
consider cleanups for transitions like the python one.
Any suggestion/idea is welcome
Regis
PS : I'm subscribed to the list, no need to CC me anymore :)
Reply to: