On Wed, 2008-03-05 at 13:20 -0500, Tom "spot" Callaway wrote:
On Wed, 2008-03-05 at 18:18 +0000, Daniel P. Berrange wrote:
> On Wed, Mar 05, 2008 at 01:11:32PM -0500, Tom spot Callaway wrote:
> > On Wed, 2008-03-05 at 13:03 -0500, Colin Walters wrote:
> > > Hi,
> > >
> > > I've spent a little bit of time on the shuttle to/from work hacking
on
> > > Makefile.common. The general idea is to remove the Release: and %
> > > changelog fields from spec files, and have them generated automatically.
> >
> > Changelog I might buy into, but Release? There are too many complicated
> > cases (Pre/Post immediately spring to mind) to automatically do this.
> > Also, we'd be forcing a hardcoded dist tag here.
>
> We should optimize for the common case - not the edge-case. I'd welcome
> this automated release / changelog management. If a handful of packages
> need to opt-out that's still a huge win for the common case.
Its not an optimization, its excluding all possible edge cases. The
kernel is a notable "edge case" here.
I'm not suggesting we immediately deploy the patch and force everyone to
use it - the goal was to prototype and get feedback, and this is exactly
the kind of issue that I wanted to know about.
The versionscript.py I mentioned in another response is one possible
solution. Another would be to add a new field to the spec file like:
VersionTransform: pre
(or post).
That would tell the build system to tweak its Release: generating
accordingly.