Proposed udpates policy change

[Christof Damian]

On Tue, Mar 9, 2010 at 00:53, Jesse Keating <jkeating at redhat.com> wrote:
> On Mon, 2010-03-08 at 18:45 -0500, Steven M. Parrish wrote:
>> As a maintainer I have seen several of my packages sit in updates testing for
>> over 2 weeks with no comments and no karma.  In fact they sat so long I got
>> nag mail about not pushing them.  Requiring a karma of +3 to push is just not
>> going to work by itself.  If anything it should be 'An update cannot be pushed
>> to stable until either it reaches a Karma of +3 or it has been in updates
>> testing for 14 days with no negative karma."
>
> There is a chicken/egg problem here.  Karma isn't required, ergo it
> doesn't happen.  If karma is required, we might see an increase in karma
> registered, as well as an uptick in tools development to help provide
> karma (we're already seeing this in F13, one could conclude that part of
> that is because karma is required for critpath packages).

I think the 14 days with no negative karma is a good solution. At
least for the packages I maintain. That is the way I use it manually
at the moment, I push packages to stable once I get the nag mail. I
also started to do mayor updates only to rawhide, enhancements to F-12
and security/mayor bug-fix updates to F-11.

I maintain mainly php-qa packages. I never received any karma for
those. So far I haven't received a single bug report. I might be the
only person using those packages (I hope not - is there any way to
find out?).

I like the idea of easy karma scripts, but these will only help for
packages which are used every day by a number of persons which are
prepared to use update-testing.

This could mean that I will only add new versions to rawhide and
simply ignore any F-XX. This would save me a lot of time, but is
probably not what users are looking for.

Cheers,
Christof