rawhide package lag

Michael Schwendt mschwendt at gmail.com
Fri Apr 15 17:58:16 UTC 2011


On Fri, 15 Apr 2011 11:24:51 -0600, OP wrote:

> On 04/15/2011 09:02 AM, Michael Schwendt wrote:
> > On Fri, 15 Apr 2011 08:27:39 -0600, OP wrote:
> >
> >> On 04/15/2011 03:59 AM, Peter Robinson wrote:
> >>> Its because people are just building packages on F-15 branch. They
> >>> don't get into rawhide until they hit F-15 stable.
> >>
> >> This has probably been discussed before, but why not import all newer f(n-1)
> >> builds automatically into rawhide?  Might provide some additional testing.
> >
> > Testing of what? A target that isn't anything like well-defined.
> > People ought to test F-15.  Rawhide is pretty much unimportant while F-15
> > is being worked on.
> 
> If that's the case, what's the point of NoFrozenRawhide?  F15 builds will land 
> in rawhide as soon as they are push to F15 stable, so why not push them to 
> rawhide earlier?  Seems like rawhide should always have the latest build.

The point of NoFrozenRawhide is to _keep it open_ for any progress there
may be instead of holding up a growing pile of builds somewhere until F-15
release. That does not imply that package maintainers MUST give it the
same love as F-15. But it's possible to work on Rawhide and actually
have the new builds appear in the repo, too. Including ABI breaks and
necessary rebuilds.

Of course, it would be nice if git f15 not contained stuff newer than git
master, but git master containing stuff newer or equal to git f15 does not
imply that it builds for Rawhide. Progress in Rawhide may need someone to
actively work on the Rawhide packages, and that is less likely short before
F-15 release.


More information about the test mailing list