rpmlint [was]: Intent to Package: BloGTK

Iago Rubio iago.rubio at hispalinux.es
Tue Mar 22 06:33:17 UTC 2005

On Mon, 2005-03-21 at 22:14 +0200, Ville Skyttä wrote:
> On Mon, 2005-03-21 at 20:28 +0100, Iago Rubio wrote:
> > Is rpmlint a valid information source for rpm sanity ?
> No, it is an imperfect tool to catch _some_ packaging insanity.
> It does often report reasonable stuff, and quite often produces false
> positives.

OK, thanks for the hint.

> > Before reading Brian's post, I tried rpmlint at it seems it breaks some
> > of the fedora's guidelines.
> If this happens, please report a bug against rpmlint, or the guidelines.

I will do it as soon as I know where the bug is :)

> > As example it screams when a package tries to obsolete itself.
> Why would a package obsolete itself?  The only case I can come up with
> right now is that a package tries to avoid an Epoch bump by using a
> versioned Obsoletes on itself.  But that sounds, well, unusual, and
> unmaintainable and non-deterministic after a few package revisions.

Frankly I don't see a reason. I suppose there should be a reason as
someone wasted his time writing it down in the guidelines, but I think
RPM is clever enough to deal with such kind of package - say -F or --
replacepkgs - without self-obsoletion.

I was just checking the guidelines against the available rpm tools.

I'd like to see consistency among both.

> > This error clashes with the guidelines at:
> > http://fedora.redhat.com/participate/developers-guide/s1-rpm-guidelines.html
> > 
> > 5.- The package may obsolete itself.
> > 
> > Any hint ?
> Could someone who authored those guidelines comment?

I hope someone find some time to guide us :)

> If self-obsoletion is considered ok in packages, I'll add a filter to
> the rpmlint config so we'll no longer see that complaint.

Yes, it's needed if self-obsoletion is ok. 

And ... to be more specific in the guidelines won't hurt:

5.- The package may obsolete itself when (fill here).

Iago Rubio

More information about the fedora-extras-list mailing list