On 06/14/2012 04:46 AM, Stephan Kulow wrote:
It's time we realize delaying milestones is not a solution. Instead,
let's use the delay of 12.2 as a reason to challenge our current
development model and look at new ways. Rather than continue to delay
milestones, let's re-think how we work.
Well, reality and you message beat me to the punch. I do have a
presentation on my box I was planing to give at oSC (assuming I can get
the necessary travel approved). But I digress....
openSUSE has grown. We have many interdependent packages in Factory. The
problems are usually not in the packages touched, so the package updates
work. What's often missing though is the work to fix the other packages
that rely on the updated package. We need to do a better job making sure
bugs caused by updates of "random packages" generate a working system.
Very fortunately we have an increasing number of contributors that
update versions or fix bugs in packages, but lately, the end result has
been getting worse, not better. And IMO it's because we can’t keep up in
the current model.
I think there are issues with the model, but I think that they cannot be
fixed if we do not address what I consider underlying issues. I believe
that an underlying issue to the overall state is the package maintenance
model. Look at any perl or python package in the devel projects and
you'll find tons of people listed as "maintainers" but almost non of
them is a "true" maintainer. Everyone of the listed "maintainers" is
inherited from some parent project and in the end no one really feels
That this is an issue is also evident in the increased mail traffic on
the list about lingering SRs.
I believe we need to clean up our package maintainer model. We should
have 2 or 3 people at the most that are responsible for a package. These
people will get messages about build failures for their packages, and
these cannot be disabled. Further, just because you are a maintainer of
a package that does not imply you become a maintainer of the parent
project. Any devel project should probably not have more then 10
maintainers. If we have packages that do not have "true" "committed"
maintainers they cannot go into Factory. This implies that in any given
devel project the maintainers of the devel project can decide to have
"unmaintained" packages that are nice to have in OBS, and that "drift
along". However, these would never show up in factory.
Basically I am proposing that we get stricter about maintainership of
packages. This may be sufficient to avoid the trouble we're in now in
the future. If it is not at least we will have a base to build a new
development model on for factory. With the current package maintainer
model and the way it is presented any new model we come up with will,
IMHO, be built on quick sand.
In summary, I think without fixing what we build on, i.e. the package
maintainer model a discussion about the Factory development model is
somewhat of a moot point.
Coincidentally, I just wrote about the same thing :-). I can only agree
that IMO less maintainers will lead to increased responsibility and
shorter response times when it comes to package submissions between
projects. Yet, I'm unsure if this will also help raising awareness of
issues within Factory. Nonetheless, it's worth discussing.