2015-02-02 17:15 GMT+01:00 whayutin <whayutin(a)redhat.com>:
On Fri, 2015-01-30 at 14:52 +0100, Jakub Ruzicka wrote:
> For some time, I'm wondering if we should really call it rdopkg CI since
> it's not really tied to rdopkg but to RDO. You can use most of rdopkg on
> any distgit. I reckon we should simply call it RDO CI to avoid
> confusion. I for one don't underestimate the impact of naming stuff ;)
I would call this particular job "RDO update CI"
So I suppose Phase 1 would be a reduced to a basic packstack job
that
runs through w/ the update from the devels submission.
Phase 2 addressed what happens when a group of submissions are pooled
together. Does one submission break another submission that could not
have been detected in the individual submissions themselves. I see two
options here..
A. Let the issue get sorted out in the stage yum repo.
B. Create a temporary yum repo with collection of submissions with a job
hitting rdo production and the temporary repo.
Applying KISS I'd take A. Also let's rename stage to testing, to be
more like Fedora/EPEL updates process.
After RDO update passes Phase1 it will be pushed to the testing repo
and once stage/testing CI job passes, repo maintainer pushes all
updates live.
Cheers,
Alan