====================================================================== The results in this summary consider Test Updates! ======================================================================
Summary of broken packages (by owner):
Matt_Domsch AT dell.com dkms - 2.0.19.1-1.el4.noarch
chrisw AT redhat.com git-cvs - 1.5.3.6-1.el4.i386
devrim AT commandprompt.com postgresql-dbi-link - 2.0.0-3.el4.noarch postgresql-pgpoolAdmin - 1.0.0-7.el4.noarch python-psycopg2-zope - 2.0.7-1.el4.i386
foolish AT guezz.net perl-libwhisker2 - 2.4-3.el4.noarch
john AT ncphotography.com bugzilla - 2.22.3-0.el4.noarch
limb AT jcomserv.net roundcubemail - 0.1.1-3.el4.noarch
lmacken AT redhat.com python-sqlobject - 0.9.2-1.el4.noarch
lxtnow AT gmail.com specto - 0.2.0-4.el4.noarch
matthias AT rpmforge.net python-Coherence - 0.2.1-3.el4.noarch
====================================================================== Broken packages in fedora-epel-testing-4-i386:
bugzilla-2.22.3-0.el4.noarch requires perl(Template::Stash) dkms-2.0.19.1-1.el4.noarch requires kernel-devel git-cvs-1.5.3.6-1.el4.i386 requires cvsps perl-libwhisker2-2.4-3.el4.noarch requires perl(MD5) postgresql-dbi-link-2.0.0-3.el4.noarch requires perl-DBI >= 0:1.52 postgresql-pgpoolAdmin-1.0.0-7.el4.noarch requires php >= 0:4.4.2 postgresql-pgpoolAdmin-1.0.0-7.el4.noarch requires php-pgsql >= 0:4.4.2 python-Coherence-0.2.1-3.el4.noarch requires SOAPpy python-Coherence-0.2.1-3.el4.noarch requires python-twisted-core python-Coherence-0.2.1-3.el4.noarch requires python-nevow python-Coherence-0.2.1-3.el4.noarch requires python-twisted-web python-psycopg2-zope-2.0.7-1.el4.i386 requires zope python-sqlobject-0.9.2-1.el4.noarch requires python-sqlite2 roundcubemail-0.1.1-3.el4.noarch requires php-pear-Mail-Mime roundcubemail-0.1.1-3.el4.noarch requires php-pear-Net-SMTP roundcubemail-0.1.1-3.el4.noarch requires php-pear-DB roundcubemail-0.1.1-3.el4.noarch requires php-pear-Net-Socket roundcubemail-0.1.1-3.el4.noarch requires php-pear-Auth-SASL specto-0.2.0-4.el4.noarch requires notify-python
On Sun, Jun 15, 2008 at 12:38:43PM -0000, Fedora Extras repoclosure wrote:
====================================================================== The results in this summary consider Test Updates! ======================================================================
Summary of broken packages (by owner):
Matt_Domsch AT dell.com dkms - 2.0.19.1-1.el4.noarch
dkms-2.0.19.1-1.el4.noarch requires kernel-devel
Last time I looked, kernel-devel is provided by Red Hat in their packages, as well as CentOS. Is it somehow excluded from the package set for the purposes of this repoclosure?
Thanks, Matt
On Sun, Jun 15, 2008 at 8:52 AM, Matt Domsch Matt_Domsch@dell.com wrote:
On Sun, Jun 15, 2008 at 12:38:43PM -0000, Fedora Extras repoclosure wrote:
====================================================================== The results in this summary consider Test Updates! ======================================================================
Summary of broken packages (by owner):
Matt_Domsch AT dell.com dkms - 2.0.19.1-1.el4.noarchdkms-2.0.19.1-1.el4.noarch requires kernel-develLast time I looked, kernel-devel is provided by Red Hat in their packages, as well as CentOS. Is it somehow excluded from the package set for the purposes of this repoclosure?
Given that some of my packages were complaining about missing glibc I'd say that the script is broken and the results cannot be relied upon.
Jeff
On Sun, 15 Jun 2008 08:55:09 -0500, Jeffrey Ollie wrote:
On Sun, Jun 15, 2008 at 8:52 AM, Matt Domsch wrote:
On Sun, Jun 15, 2008 at 12:38:43PM -0000, Fedora Extras repoclosure wrote:
====================================================================== The results in this summary consider Test Updates! ======================================================================
Summary of broken packages (by owner):
Matt_Domsch AT dell.com dkms - 2.0.19.1-1.el4.noarchdkms-2.0.19.1-1.el4.noarch requires kernel-develLast time I looked, kernel-devel is provided by Red Hat in their packages, as well as CentOS. Is it somehow excluded from the package set for the purposes of this repoclosure?
Depends on how it is configured. If run for i386, the report is true, because kernel-devel is available for i586/i686 only (in the CentOS 4.6 repo at least). The arch here must be i686 when running repoclosure.
Given that some of my packages were complaining about missing glibc I'd say that the script is broken and the results cannot be relied upon.
It isn't broken. It was misconfigured on Fri, 13 Jun 2008 16:51:25 +0000 (UTC) according to the diff that was sent to sysadmin-members. For ppc and x86_64 the RHEL repository URLs are bad. As why this doesn't create a fatal error in yum remains to be investigated.
On Sun, 15 Jun 2008 08:52:22 -0500 Matt Domsch Matt_Domsch@dell.com wrote:
On Sun, Jun 15, 2008 at 12:38:43PM -0000, Fedora Extras repoclosure wrote:
====================================================================== The results in this summary consider Test Updates! ======================================================================
Summary of broken packages (by owner):
Matt_Domsch AT dell.com dkms - 2.0.19.1-1.el4.noarchdkms-2.0.19.1-1.el4.noarch requires kernel-develLast time I looked, kernel-devel is provided by Red Hat in their packages, as well as CentOS. Is it somehow excluded from the package set for the purposes of this repoclosure?
The broken deps checker has gone mad again today. I'd just ignore the reports.
Paul.
Matt Domsch wrote:
On Sun, Jun 15, 2008 at 12:38:43PM -0000, Fedora Extras repoclosure wrote:
====================================================================== The results in this summary consider Test Updates! ======================================================================
Summary of broken packages (by owner):
Matt_Domsch AT dell.com dkms - 2.0.19.1-1.el4.noarchdkms-2.0.19.1-1.el4.noarch requires kernel-develLast time I looked, kernel-devel is provided by Red Hat in their packages, as well as CentOS. Is it somehow excluded from the package set for the purposes of this repoclosure?
repoclosure (last time I checked) compared requirements within repositories rather then spanning several repositories. I had written a similar utility to check dependencies across repositories[1] but it didn't pass.
Anyway, might that (not crossing repos to fulfill deps) be the actual problem? I haven't looked at it recently.
-Jeroen
https://www.redhat.com/archives/fedora-devel-list/2007-August/msg01431.html
On Sun, 15 Jun 2008 16:29:29 +0200, Jeroen van Meeuwen wrote:
Summary of broken packages (by owner):
Matt_Domsch AT dell.com dkms - 2.0.19.1-1.el4.noarchdkms-2.0.19.1-1.el4.noarch requires kernel-develLast time I looked, kernel-devel is provided by Red Hat in their packages, as well as CentOS. Is it somehow excluded from the package set for the purposes of this repoclosure?
repoclosure (last time I checked) compared requirements within repositories rather then spanning several repositories.
What does that mean?
I had written a similar utility to check dependencies across repositories[1] but it didn't pass.
?
Anyway, might that (not crossing repos to fulfill deps) be the actual problem? I haven't looked at it recently.
Repoclosure (albeit here it is the modified one for Fedora Extras) examines a _set_ of repositories. Packages in Repo B can fulfill requirements of packages in Repo A and vice versa.
On Sun, 15 Jun 2008, Michael Schwendt wrote:
On Sun, 15 Jun 2008 16:29:29 +0200, Jeroen van Meeuwen wrote:
Summary of broken packages (by owner):
Matt_Domsch AT dell.com dkms - 2.0.19.1-1.el4.noarchdkms-2.0.19.1-1.el4.noarch requires kernel-develLast time I looked, kernel-devel is provided by Red Hat in their packages, as well as CentOS. Is it somehow excluded from the package set for the purposes of this repoclosure?
repoclosure (last time I checked) compared requirements within repositories rather then spanning several repositories.
What does that mean?
I had written a similar utility to check dependencies across repositories[1] but it didn't pass.
?
Anyway, might that (not crossing repos to fulfill deps) be the actual problem? I haven't looked at it recently.
Repoclosure (albeit here it is the modified one for Fedora Extras) examines a _set_ of repositories. Packages in Repo B can fulfill requirements of packages in Repo A and vice versa.
Sorry for the noise guys but can _anyone_ find a single run of this script that worked since January? Just curious because it seems like I've been spending a whole lot of time on it since then and its honestly starting to piss me off. This thing should just work but for whatever reason (not blaming anyone here) it never seems to work right.
-Mike
On Sun, 15 Jun 2008 11:40:53 -0500 (CDT), Mike McGrath wrote:
Sorry for the noise guys but can _anyone_ find a single run of this script that worked since January? Just curious because it seems like I've been spending a whole lot of time on it since then and its honestly starting to piss me off. This thing should just work but for whatever reason (not blaming anyone here) it never seems to work right.
Well, you cannot use $basearch in the yum.conf because that expands to the host arch, not the desired target repo arch.
As why the single reports has been split into multiple reports again, I've contacted the author of that change just recently.
The other [never-ending] problems are with the various RHEL repos, which either are moved from time to time or are not in the place that's specified in the yum config file.
That also causes problems with the buildsys where packages are not found.
JFTR, when running the repoclosure against public CentOS 4.6 repos, the broken deps report is different.
The run posted to this list on 2008-06-11 was close, probably just using i386 instead of i686.
epel-devel@lists.fedoraproject.org
-
Fedora Koji Build System -
Jeffrey Ollie -
Jeroen van Meeuwen -
Matt Domsch -
Michael Schwendt -
Mike McGrath -
Paul Howarth