[users] Dependency trouble with wine 0.9.31-1.el3.rf

Dag Wieers dag at wieers.com
Sun Mar 4 13:46:29 CET 2007


On Sat, 3 Mar 2007, Bart Schaefer wrote:

> On 3/3/07, Dag Wieers <dag at wieers.com> wrote:
> > On Sun, 25 Feb 2007, Bart Schaefer wrote:
> >
> > > It appears that desktop-file-utils-0.3-10 on RHEL3 is too old.
> 
> I think you misunderstood what I meant by that.  The
> desktop-file-utils that is packaged with RHEL3 is version 0.3-10.  The
> one you built with is newer than that.  But the newer one that you
> built with does not appear to be available as an RHEL3 package from
> RPMforge, so the dependency fails because there's nowhere in my yum
> config where *I* can get an upgrade for desktop-file-utils.

The update-desktop-database dependency was wrong. We do run the script if 
it is available, but it should not have been there as a dependency.

You don't require a newer desktop-file-utils and I am certainly not going 
to replace what Red Hat provides :)


> > Sorry about that. I've fixed it now and updated to 0.9.32.
> 
> Well, thanks for the attempt, but this is still happening:
> 
> Resolving dependencies
> ......Unable to satisfy dependencies
> Package wine-core needs /usr/bin/update-desktop-database, this is not
> available.
> Package wine-core needs /usr/bin/update-desktop-database, this is not
> available.

I wasn't able to push the new version, so nothing has been fixed when you 
tried. The transfer only just finished (generating metadata and uploading 
6 wine versions takes a long time).


> > Fact is that we should catch such incidents before the repository is
> > pushed.
> 
> Ideally you'd build each package on a virgin RHEL3 (or whatever)
> installation with only the minimal set of dependencies upgraded no
> further than is available via either the standard repository for the
> distribution or via RPMforge.  However, I realize that's asking for a
> tremendous amount of effort and/or a lot of virtual hardware.

I do build on a clean RHEL3, the wrong dependency had nothing to do with 
the build environment but was a human error. It could have been catched if 
we'd checked for repository consistency before pushing, however there are 
a lot of false negatives when doing that so it's not an easy task.

Kind regards,
--   dag wieers,  dag at wieers.com,  http://dag.wieers.com/   --
[all I want is a warm bed and a kind word and unlimited power]


More information about the users mailing list