mock-0.9.19 on CentOS5/RHEL5

[Florian La Roche]

Hello,

if you want to run the newest version of mock (0.9.19)
with RHEL5/CentOS5, you can use a backported version from:

http://www.jur-linux.org/rpms/el-updates/5/SRPMS/mock-0.9.19-1.el5.src.rpm
http://www.jur-linux.org/rpms/el-updates/5/i386/mock-0.9.19-1.el5.noarch.rpm

regards,

Florian La Roche

[Karanbir Singh]

On 11/12/2009 06:40 AM, Florian La Roche wrote:
> Hello,
>
> if you want to run the newest version of mock (0.9.19)
> with RHEL5/CentOS5, you can use a backported version from:

Have you done any tests to see how the output changes from what we have 
in extras/ now ? Are these test results public somewhere ?

[Mike McLean]

On 11/12/2009 01:40 AM, Florian La Roche wrote:
> if you want to run the newest version of mock (0.9.19)
> with RHEL5/CentOS5, you can use a backported version from:
>
> http://www.jur-linux.org/rpms/el-updates/5/SRPMS/mock-0.9.19-1.el5.src.rpm
> http://www.jur-linux.org/rpms/el-updates/5/i386/mock-0.9.19-1.el5.noarch.rpm

So basically just reverting the dbus patch (2046136e)?

Perhaps since RHEL5 is something of an important platform for mock we 
should consider making this codepath conditional, or otherwise fix this 
without forking.

[Clark Williams]

On Fri, 13 Nov 2009 13:09:14 -0500
Mike McLean  wrote:
> On 11/12/2009 01:40 AM, Florian La Roche wrote:
> > if you want to run the newest version of mock (0.9.19)
> > with RHEL5/CentOS5, you can use a backported version from:
> >
> > http://www.jur-linux.org/rpms/el-updates/5/SRPMS/mock-0.9.19-1.el5.src.rpm
> > http://www.jur-linux.org/rpms/el-updates/5/i386/mock-0.9.19-1.el5.noarch.rpm
>
> So basically just reverting the dbus patch (2046136e)?
>
> Perhaps since RHEL5 is something of an important platform for mock we 
> should consider making this codepath conditional, or otherwise fix this 
> without forking.

I've got a patch in my tree but I'm trying to figure out why my
validation tests are failing on epel4. Should ahve a 0.9.20 out early
next week though.

Clark

[Florian La Roche]

On Fri, Nov 13, 2009 at 01:09:14PM -0500, Mike McLean wrote:
> On 11/12/2009 01:40 AM, Florian La Roche wrote:
> > if you want to run the newest version of mock (0.9.19)
> > with RHEL5/CentOS5, you can use a backported version from:
> >
> > http://www.jur-linux.org/rpms/el-updates/5/SRPMS/mock-0.9.19-1.el5.src.rpm
> > http://www.jur-linux.org/rpms/el-updates/5/i386/mock-0.9.19-1.el5.noarch.rpm
>
> So basically just reverting the dbus patch (2046136e)?
>
> Perhaps since RHEL5 is something of an important platform for mock we 
> should consider making this codepath conditional, or otherwise fix this 
> without forking.

Hello Mike,

this small change is all that is needed, at least for a first start.
I've put this into bugzilla in August and Clark has done a clean patch
within two days, but it is not commited into the public repo and also
no new rpm has been pushed out. See bz 516355 for details.

Karanbir: You've already emailed the current CentOS position to have
one mock version across all of CentOS3/4/5 running and that the newest
bits don't qualify for that. This could maybe change next year once CentOS3
does not get updates anymore. I could see similar decisions like e.g.
keeping to a mock version starting at initial release of a new major
CentOS push-out. (Then CentOS-5 would always stick to the mock release
that was used for CentOS-5.0...)
Another idea that Red Hat would probably support is to just use the
same mock version that also Red Hat uses for their releases, just to have
another data point that CentOS is really 100% compatibel to RHEL.
(Keeping infrastructure the same should be extended to as many parts
as possible.)

regards,

Florian La Roche

[Florian La Roche]

On Fri, Nov 13, 2009 at 01:09:14PM -0500, Mike McLean wrote:
> Perhaps since RHEL5 is something of an important platform for mock we 
> should consider making this codepath conditional, or otherwise fix this 
> without forking.

Hello Mike,

thanks for getting the new release of mock out, also for the pykickstart
commit to keep working with older python releases. I've moved to newest
rpms and besides mock I've also tested koji a bit more now. Some recompile
tests with kernel,glibc,memtest86+ and a dozen other rpms did all work ok,
so the special koji config steps should be complete now...

(Also Michal Hlavinka has merged in some bits for dovecot to have newest
Fedora Rawhide rpms compile on RHEL5, so overall it's been a good week to
reduce extra patches for me.)

thanks,

Florian La Roche