strange (64bit) requirements in mock i386

Michael Schwendt bugs.michael at gmx.net
Mon Dec 31 21:12:18 UTC 2007


On Mon, 31 Dec 2007 14:39:45 -0600, Dennis Gilmore wrote:

> > > setarch 'i386'
> > > DEBUG util.py:260:  0:xorg-x11-devel-6.8.2-1.EL.33.i386
> > > DEBUG util.py:260:  0:tcsh-6.13-9.el4.1.i386
> > > DEBUG util.py:260:  0:xbae-devel-4.60.4-8.el4.1.x86_64 <---------- this
> > > one is wrong
> > > DEBUG util.py:260:  0:lapack-3.0-25.1.i386
> > > DEBUG util.py:260:  0:blas-3.0-25.1.i386
> > > DEBUG util.py:260:  0:gawk-3.1.3-10.1.i386
> > > DEBUG util.py:260:  0:desktop-file-utils-0.9-3.el4.i386
> > > DEBUG util.py:260:  0:gcc-g77-3.4.6-9.i386
> > >
> > > No idea why the 64bit version of xbae-devel gets pulled in, especially
> > > since this happens in my local mock, too. Looks like setarch i386 is not
> > > enough.
> >
> > This looks like some sort of yum problem. Does EL4 have combined
> > i386/x86_64 repos? Why are there 64-bit rpms in the i386 EL4 repository?
> >
> > You may try "fixing" this by putting an "exclude=*.x86_64" in the
> > yum.conf section of the mock config.
> > --
> this is a long standing problem where if there is a /etc/rpm/platform file 
> things get confused  the plague needsign repo contains ppc i386 and x86_64   

worse than that, plague needsign is an "all in one" repo, it contains
_all_ arches and even the source rpms

> removing the /etc/rpm/platform file fixes the problem.  which i did.  things 
> have long reverted to normal.




More information about the epel-devel-list mailing list