On Tue, 29 Sep 2009, Troy Dawson wrote: <snip> > Hi Eva, > I bet it stems from us having a epoch of 2 in our version of R. > That stems back from the first time we compiled R, I didn't see that we had > an epoch of 2. I didn't notice that we had the epoch until I was reading the > spec file from Fedora or EPEL, which has the epoch in, with a note saying > they had to do it because Scientific Linux had it. Once you have an epoch, > it's almost impossible to get it out. > > But how to fix your problem? > You need to exclude R that is coming from Scientific Linux. That's not too > hard. You need to add the following at the end of /etc/yum.repos.d/sl.repo > and /etc/yum.repos.d/sl-security.repo > > exclude=R R-devel libRmath libRmath-devel > > That should keep it from trying to install Scientific Linux's R. I have a horrible feeling that the epoch=2 evil might well be my fault. The changelog in the sl53 (R-2.8.1) specfile contains: * Fri Feb 08 2008 Jon Peatfield <[log in to unmask]> - 2:2.6.2-1 - Update to 2.6.2 tweaks to build on SL5x (and add DAMTP evil hacks in a slightly cleaner way). Include fix from SL5x 2.5.0-2 srpm (zip must be a BuildRequires or it won't get listed in etc/Renviron when built in brewbuilder). Epoch needs to be 2 for SL5x. ... before that all the epoch numbers mentioned seem to be 1... If I did cause this mess then I'm sorry for all the trouble is has caused. In any case I wonder it it would be worth us asking (nicely) if the CRAN people would set epoch=2 on their rhel5 rpm builds since otherwise anyone using EPEL on RHEL/SL/CENTOS will be unable to directly use the CRAN version... -- /--------------------------------------------------------------------\ | "Computers are different from telephones. Computers do not ring." | | -- A. Tanenbaum, "Computer Networks", p. 32 | ---------------------------------------------------------------------| | Jon Peatfield, _Computer_ Officer, DAMTP, University of Cambridge | | Mail: [log in to unmask] Web: http://www.damtp.cam.ac.uk/ | \--------------------------------------------------------------------/