On Mon, Jun 23, 2014 at 12:44:31 -0400, Matthias Clasen <mclasen@xxxxxxxxxx> wrote:
I would suggest that the fix for this is to not push broken kernels so frequently that 'the oldest one is the only that works' becomes an issue, and to introduce automatic testing that ensures you can at least boot rawhide to the login screen, _before_ a build gets pushed out to the masses.
While this usually isn't a problem for the updates repo (though there have been times where kernels were not usable by me for several months even in the released branches), it's not so nice for rawhide (possibly plus the rawhide nodebug repo) where there can be new kernels pushed to rawhide each day (and more than one if you're testing things from koji).
For example right now I need to use a 3.15 kernel on one of my machines because of an issue affecting md raid and the running kernel is the oldest and only working kernel installed on that machine. And on another machine, I think I have a Nouveau regression (I'm not sure yet and haven't had time to diagnose that problem) and again I need to use a 3.15 kernel which is both the running kernel and the oldest installed kernel.
-- devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
