Let's get moving

[David Botsch]

See Johnny Strom's email.... you and he are having different interpretations of
what means what...

my interpretation this point of PUBLISH is when an SRPM is built and is on
someone's public server someplace. People download it, QA it, and say PUBLISH
or DONTPUBLISH.

When it gets two gpg signed PUBLISHes, it gets stuck in updates-testing

Two weeks after that, w/o objection, it should go to updates.

So, if that is not what is supposed to be happening, then, we've got a problem.
One can have too much QA, and if I'm reading your interpretation of what is
going on, then I would say we have too much QA, perhaps.

On Tue, Sep 07, 2004 at 02:23:57PM -0500, Eric Rostetter wrote:
> Quoting David Botsch <dwb7 at ccmr.cornell.edu>:
> 
> > If I read the QA doc correctly, what it actually refers to is getting stuff
> > put
> > into updates-testing, not moved out of updates-testing to updates.
> 
> The bottom section about publishing is about moving it from updates-testing
> to updates.  It should be made more clear.
> 
> > After talking w. some folk in the channel, my understanding is that after 2
> > weeks, if no one objects, packages should be moved from updates-testing to
> > updates and released as final.
> 
> No, IRRC.  If the package has passed QA for one OS version, but not for 
> other OS versions, then after 2 weeks all versions can be passed.  This
> is there so that if one SRPM is built for multiple OS versions, they won't
> all hang up because only one OS version wasn't tested.  It really only 
> applies when the same SRPM, or at least the same patch code, was used for
> all of the versions.  If a different patch was used for the different versions,
> then this probably shouldn't be applied to that case.
> 
> At least that is my memory of the situation.
> 
> > Unless I misunderstood, it also seems that only one person currently has the
> > ability to actually do this. And, 2 weeks is also way, way, way tooo long. A
> > couple days is more like it.
> 
> Yes, currently it is one person.  But that doesn't usually matter as the
> RPMs never even get the QA publish votes for them to be moved.  So one
> person or 10 people, it doesn't matter if there is nothing to be moved.
> 
> And 2 weeks is a timeout if no one votes on it.  To speed that up, vote on
> it.  Participation is the key.
> 
> -- 
> Eric Rostetter
> 
> 
> --
> fedora-legacy-list mailing list
> fedora-legacy-list at redhat.com
> http://www.redhat.com/mailman/listinfo/fedora-legacy-list

-- 
********************************
David William Botsch
Consultant/Advisor II
CCMR Computing Facility
dwb7 at ccmr.cornell.edu
********************************