[Rpm-maint] querry on "rpm warning"

devzero2000 pinto.elia at gmail.com
Wed Apr 23 16:11:53 UTC 2008

On Wed, Apr 23, 2008 at 2:04 PM, Panu Matilainen <pmatilai at redhat.com>

> On Wed, 23 Apr 2008, devzero2000 wrote:
>  I don't think it is right to use context marked dependency (e.g.
> > scriptlet deps ordering ) for ordering package installation, if the
> > packages
> > don't use these in first place
> > (
> > http://www.rpm.org/max-rpm-snapshot/s1-rpm-depend-manual-dependencies.html
> > ).
> >
> > If  absolutely necessary and the automatic deps resolutions of rpm is
> > not
> > sufficient  just use plain  Requires: baz
> >
> There's nothing "wrong" in using context marked dependencies. In fact
> > they're often the only way to ensure correct ordering in complex package
> > sets / transactions (notably due to dependency loops that creep into
> > packaging all too easily).
> >
>        - Panu -
> Yes, I know. Perhaps this one is an example

My position is that solving the rpm problems on dependency loop or bad rpm
package would be better instead of using workaround solutions. I hope that
you agreed on this.

I call workround because isn't elegant to have in the, for example, "baz"
SPEC file Requires(pre): bar and "baz" not having an %pre section that call
something of "bar". Al least, the meaning isn't obvious, almost for most

I could be wrong. If so it would be useful  to update the document, for
example, "Fedora Packaging Guideline" and defining the "context marked
dependecy" how a best practices, in some situation, to assure the rpm
install ordering, also if aren't defined any scriptlet section.

OK. In effect we disgress from the original question. *

*Sorry if the subject wasn't on the point. In every case look interesting*.*

Best Regards

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.rpm.org/pipermail/rpm-maint/attachments/20080423/a73b9988/attachment-0001.htm 

More information about the Rpm-maint mailing list