On Mon, 1 Mar 2010 12:11:20 -0600 (CST)
Mike McGrath <mmcgrath(a)redhat.com> wrote:
On Mon, 1 Mar 2010, Tom "spot" Callaway wrote:
> On 03/01/2010 12:48 PM, Peter Jones wrote:
> > I'd also like a policy in place to help us avoid situations like
> > the recent dnssec unpleasantness.
>
> Sure. I'm just not at all convinced that if those packages had sit
> in testing for $ARBITRARY_PERIOD_OF_TIME that they would have been
> tested and fixed.
>
> In EPEL, where there is a mandatory period of "testing" for
> updates, I almost never get Bodhi karma, and if I do, it is only
> because I have been aggressively harassing users to test and give
> karma.
>
This has been my experience too. While I like the ability to have a
karma / feedback system. In reality it doesn't seem to actually be
part of our package release workflow.
I will note again that 0 karma doesn't mean that it's not been tested.
I run here with updates-testing enabled, but often only report -karma
when things break. There are just so many things in updates-testing,
that I usually don't have time or energy to +1 all of them that work.
I have reported -karma on about 4-5 things in the f12 cycle and also
actively asked others to do so, preventing these things from being
pushed, so I don't agree that updates-testing is not useful.
kevin