"Action Items" From FUDCon?

[Jeff Spaleta]

On Jan 13, 2008 9:57 PM, Axel Thimm <Axel.Thimm at atrpms.net> wrote:
> I would consider "forced retagging" a feature not a bug.   Otherwise for
> each failed build during development you get a release tag inflation
> with no useful %changelog entries (sure one can pile up all comments
> about how many bugs one introduced during an upgrade of a package, but
> that's not what the %changelog is about ;).

Perhaps you read this wrong.  I'm not looking to take away the ability
of contributors to retag like they are currently doing. But I want a
way for koji to add an additional tag back into cvs that maintainers
don't need to touch as part of the workflow for creating tags that
koji uses for input.

>
> And having scratch builds to avoid this and a proper one at the end
> sounds like too much CPU cycles burning.

Again perhaps i wasn't clear.  I don't care about scratch builds at
all. I don't need people to do scratch builds, I don't need koji to
write a tag back into cvs for scratch builds. I just want koji to
write a tag back to CVS for things that completely a build with a
release candidate target of some sort that we can rely on to exist.
So if I need to regenerate the srpm associated with that koji build
which we released to a public tree.

>
> But I fully agree on the need to have a reproducable method for
> packages. Just don't remove the retagging *feature* for packagers :)

I'd very much like to avoid removing retagging.  If koji was writing
new, specially formatted tags back to CVS at the end of each completed
build would that get in way of using retagging like you have been
using?

-jef