scratch-build on untagged code [Was: Re: Proposal: Better force-tag]

[Nils Philippsen]

On Tue, 2008-09-16 at 11:53 +0200, Dominik 'Rathann' Mierzejewski wrote:
> On Tuesday, 16 September 2008 at 10:02, Matej Cepl wrote:
> > On 2008-09-15, 20:49 GMT, Doug Ledford wrote:
> > > I *could* do an srpm build to test ppc/ppc64 kernel builds, but 
> > > it takes 30 minutes for a kernel srpm to upload to the build 
> > > system from here (admittedly, I do this on my rhel kernel 
> > > builds, I haven't been messing with Fedora kernels, but the 
> > > point is valid regardless), so I commit, tag, build, fix, 
> > > force-tag, build, fix, etc.
> > 
> > I don't want to immerse myself into this discussion (which 
> > I haven't heard and read before), but this looks like bad 
> > workaround around stupid problem. The problem here is IMHO that 
> > koji is apparently not able to make a scratch build from untagged 
> > (but commited) code.
> 
> Wrong. You can use the tag "HEAD".

Which doesn't help you if somebody else is concurrently working on the
same package and checks in things while you're waiting for your scratch
build. We could use "scratch-tags", but I find this a bit silly. IMO,
this is rather a point where using an SCM with a repo-wide revision-id
(instead of a per file one as used by CVS) would help. Who again was
looking for use-cases for different SCMs ;-)?

Nils
-- 
Nils Philippsen      "Those who would give up Essential Liberty to purchase 
Red Hat               a little Temporary Safety, deserve neither Liberty
nils at redhat.com       nor Safety."  --  Benjamin Franklin, 1759
PGP fingerprint:      C4A8 9474 5C4C ADE3 2B8F  656D 47D8 9B65 6951 3011