Worthless updates

[Adam Williamson]

On Wed, 2010-03-03 at 03:03 -0500, Jon Masters wrote:

> This isn't $Enterprise_Linux, it doesn't come with a guarantee and does
> expect to be a moving target, but that doesn't mean there can't be a
> predictable update cycle and a reasonable expectation that updates are
> necessary and won't break systems.

I'm trying to avoid posting to these threads any more since there's far
more heat than light, but as a general statement -

I think we've diverged a long way from the origin and started discussing
things that are far more radical. I'm generally in favour of a small
tweak to the current updates policy which would require updates to
*either* spend a certain amount of time (not too long - I think 4 days
or so would be fine) in -testing, *or* acquire a set level of positive
karma and no negative karma, before being released. Through some channel
such a proposal should wind up in front of FESco at some point and get
voted on.

I don't see this as being at all the same thing as instituting a set
update cycle, or a stricter definition of what people should ship as
updates, or anything like that. I think those are much bigger questions
and should not be decided at mailing list or FESco level; they are board
questions, if anything. Ultimately they're part of the ongoing 'what is
Fedora' question.

I'm certainly not in favour of releng or QA or anyone else using the
small proposed tweak to the existing system as a pretext to block
updates we don't consider 'acceptable'. Without a much wider and more
comprehensive discussion, I don't think anyone can take it upon
themselves to define the character of the project and start refusing
updates for not being security issues or documented bugfixes or
something like that. As far as the original proposal went, it's my
understanding that maintainers would still be able to submit whatever
they liked as updates, and they would be promoted to the regular updates
repo as long as they got the positive karma, or stayed in -testing for a
few days without acquiring negative karma, even if they're version bumps
or whatever. Probably at present the maintainer should be able to force
through an update that got multiple negative karma (to avoid the
possibility of someone 'DoSing' updates), though this should be tracked
somewhere for general review so we can see if it actually winds up in
people still pushing updates that *really* shouldn't be pushed.

I'm a bit uncomfortable at the seeming extension of the debate to far
more radical things about update cycles and refusing non-bugfix updates
and stuff like that. I just wanted to make it clear that personally I
don't think anything that radical should be in play without some serious
consideration, and I hope we can keep different proposals clearly
separated.
-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org
http://www.happyassassin.net