Fwd: Re: releasing updates-testing packages without VERIFY votes

[Eric Rostetter]

Arg, sent with wrong From: address, so here it is again, since the moderator
probably won't get to it for a while...

----- Forwarded message -----
 Subject: Re: releasing updates-testing packages without VERIFY votes
      To: fedora-legacy-list at redhat.com

Quoting Pekka Savola <pekkas at netcore.fi>:

> I suggest changing the policy so that packages in updates-testing
> which haven't got any VERIFY votes could:

First, let me say that it would take less time for the people invloved in these
"lets publish without QA" discussions to just QA the packages than they are
spending arguing if we should publish them without any QA.  But, back to
the current point of discussion...

>   - after 2 weeks, marked with a timeout
>   - after the timeout of 4 weeks [i.e., 6 weeks total] be
>     officially published

This goes against everything this group was founded on, and all Best
Practices.  However, it does seem to be popular with the few folks
involved in these conversations.  So, I'll approve of this, but only
if ammended to include the following:

After the 2 week period when it is marked with a timeout, a message MUST
be posted to the fedoral-legacy-list at redhat.com informing people that it
is being marked for timeout, and making a plee for people to QA test it.

In addition, if it is released without any QA, the bug ticket for the
package MUST note that it was released without any QA.
 
> (And rp-pppoe and squid currently in updates-testing could be released
> immediately upon the acceptance of this policy.)

I've just submitted QA for those for RHL 7.3 and RHL 9.  Take those votes
for what you will.

I guess there is still a question: If I QA a package on RL 7.3 and RHL 9
is that one vote (since one person did the QA) or two votes (since I did
two OS versions)?

> > I vote to just release them after a long timeout period. If there are
> > any issues, we can quickly fix them afterwards. We most often use
> > patches that came from upstream or from another distro anyway, so most
> > of them have already gone through QA.

I agree, with the above stipulation about announcing it to the mailing list
and marking it as such in Bugzilla.

-- 
Eric Rostetter
The Department of Physics
The University of Texas at Austin
 
Why get even? Get odd!


----- End forwarded message -----


-- 
Eric Rostetter