Trying to install rawhide: madness and where do we hope to go ?

Jeff Spaleta jspaleta at gmail.com
Wed Sep 7 13:34:30 UTC 2005


On 9/7/05, Daniel Veillard <veillard at redhat.com> wrote:
>  I would like to point out that there is a lot of human and physical
> resources affected in getting rawhide out and tested, and in comparison that
> looks small to me.

Size of the effort is relative to perspective.  If you asking dave to
set this up, I doubt he's going to come to the same decision as to
worth.  Though i seriously doubt this is a volunteer situation inside
the redhat fenceline. If this sort of hardware is going to be setup
and be relied on to hold back builds I think you'll have to have a
series of meetings about provisioning this. Would it be cool to do
this sort of testing before rawhide lands..sure.. it'd be super
cool... if it can be automated reliably.  You'd have to ask a script
to pull the binaries from the buildsystem before they are all pushed
to rawhide if there is a problem.

Though at this point, I'm not sure its fair to single out x86 as
special anymore. If a kernel boots on ppc and x86_64 systems but
doesnt boot on x86, I'm not sure its appropriate to hold back that
kernel. But thats a policy decision, that comes after there is
hardware to do the regression testing with appears.

>   In that case it's not a transcient failure, I think last time I tried to
> use anaconda of rawhide back in July I had the exact same error. I think
> everybody agree it is known to be broken, but hasn't been fixed 

anaconda is going through what I would call major surgery.  Its not so
much about trying to fix it daily... its more about making sure it
works by test1 release.  If memory serves the anaconda development
that is going on was one of the reasons why the fc5 schedule is what
it is...to make room in part to get anaconda retooled. I don't
remember there being a promise to anyone to make day-to-day rawhide
installability a high priority. There is a target that the anaconda
devs are shooting for, fc5test1.  Are you suggesting they maintain a
dead-end branch of anaconda in rawhide while working on the new
version somewhere else?

-jef




More information about the fedora-devel-list mailing list