Kevin Fenzi wrote:
> - I don't distrust our maintainers. I very much value the work they do
> and without them we would have no Fedora. However, I also want to
> help them do the right thing for our users (who I also would like to
> see happy). I'm open to ideas on how to reduce 'red tape' for them,
> while increasing the standard of packages for our users. I know you
> have many such ideas, but I don't agree that we should not have
> testing or help our maintainers find problems before our users get
> the package.
Helping our maintainers is one thing, FORCING them to work the way FESCo
wants is another. And sadly, FESCo's update policy does the latter, NOT the
former. "Helping" means ADVISING people, not REQUIRING them to follow some
bureaucratic process. Update guidelines should be purely informative, not
hardcoded in Bodhi as you, the other 8 FESCo members, decided (over my
strong dissent).
And again, it is simply NOT TRUE that I'm arguing that "we should not have
testing". I'm arguing that SOME updates should, at the maintainer's
discretion, bypass testing because the urgency largely outweighs the risk
(be it because the risk is extremely small, because the urgency is extremely
high or both). The maintainer is in the best position to make such a call. I
DO complain in the rare occasions where some update which breaks things is
pushed directly to stable. But it means the maintainer screwed up and we
need to teach the maintainer how to avoid such a mistake the next time, not
to outright ban all direct stable pushes, many of which are legitimate. (In
the cases I complained about, it was always quite obvious to me, and I
believe to any sufficiently experienced maintainer, that the decision to
push to stable was inappropriate, even without the hindsight.) But I ALSO
complain when an urgent fix is NOT pushed to stable in a timely manner,
sitting around in testing for no good reason.
> - I read this list every day, and am very mindful of feedback from
> developers. Any communication media is good, IMHO. My mailbox is also
> always open. I think many become discouraged with the mailing list
> these days because a few people reply to EVERY SINGLE POST with no
> new thoughts or information. Make a reasoned argument, wait and reply
> (at a high level) to feedback. Posting a reply to every post
> repeating yourself just makes less people able or interested in
> following the discussion.
Then why did you not take such feedback into account when voting? Several
people, including some very experienced high-profile packagers, objected to
the new update policies. They have been ignored, by you and by the other 7
members.
> - I would like to hope that we can look beyond ourselves. We
> shouldn't be looking at "My packages" or "My Desktop". We should all be
> working for a Fedora that we can be proud of our users using. We should
> be consistent about how much testing we do and when we update things so
> ALL our users will know whats going on.
I'm sorry, but I don't agree with you at all about this kind of bureaucracy
being the way to make our users happy. I strongly believe our maintainers
are the ones who know best how to make their users happy, in particular,
what to push to users of the stable updates and when.
As for consistency: our users have explicitly asked for non-conservative, or
even "adventurous", updates, see Adam Williamson's poll, so I believe the
way to make our users happy while being consistent is to consistently push
new versions as updates unless there's a reason not to (and I already
detailed possible reasons not to push an update on several occasions, so I
won't do it again). But of course this should also be an indicative policy
and ultimately the maintainer's decision, as they know best whether there's
a reason not to push the update. We should just make it clear that the
general policy is for new versions to be pushed unless there's a reason not
to.
> Can we Improve things? I absolutely think so, but change takes time,
> well reasoned argument, and people willing to do the work to make it
> happen.
True change mainly takes a change in attitude in FESCo. Otherwise all the
"change" we'll get will be towards more and more bureaucracy. :-(
The fact that a change requires implementation work is a strong hint that
the change may be technobureaucratic. Most non-bureaucratic approaches
require little to no work to implement.
Kevin Kofler
--
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxxx
https://admin.fedoraproject.org/mailman/listinfo/devel
