On Sun, Aug 26, 2012 at 10:38 PM, "S.Çağlar Onur" <caglar(a)10ur.org>
wrote:
Hey Steve,
On Aug 24, 2012, at 8:42 PM, Steve Salevan wrote:
> Hey guys,
> I apologize for being a bit out of the loop over the past few weeks, but
I wanted to drop in with a bang with my first func-related RFC. I've spent
a fair bit of time recently thinking about where next to take Func, as
development has stalled, but it has a wide array of users, almost all of
whom have experience in hacking it into submission.
>
> If there's anything that I've learned about systems management software
in my experience, hacking like this is in its very nature, and the more of
it that happens in the upstream community, the more capable the tool ends
up becoming. As the project's new maintainer, my first goal is to get
these sorts of commits going again, to start a flow of project improvements
that fundamentally round out in a better tool for all of us. Towards this
end, it occurs to me that it's pretty damn hard to contribute to Func, as
e-mail patch submission is a bit awkward and not very conducive to the
sorts of technical discussions we should be having when new code is
presented.
>
> Thus, for my first move, I'm thinking about moving Func to github, as
that pull request feature of theirs makes upstream interaction a fair bit
simpler and allows us to conduct code reviews, which will help us keep the
code tight going out. My second thought is that it's about time we brought
you guys into the fold as committers if you've spent serious time improving
Func, so if you've submitted several substantial patches/pull requests, you
will gain commit access.
That sounds like a great idea.
> Finally, to get project development going, I've found that
setting a
roadmap with goals helps drive development, so here's a rough draft of what
I'm thinking about feature-wise for Func 0.30, alongside their respective
priorities (0==highest):
>
> 0 - Improve client-side daemon stability
Agreed, we run monit just for that reason :)
Nice! Monit has some godlike powers, we use it out here for much the same.
In fact, that initscript change was our attempt to make it a bit more
stable under monit, we found that the process table grep would often catch
the grep itself.
> 0 - Prevent forks from listening on :51234
> 1 - Improve forking system, prevent zombie fork processes
That would be great, though 2c616830 introduced a workaround for that
problem. At least for me, func no longer keeps zombie processes around.
That, sir, is a solid commit, well played. Our code is a few clicks
younger than that, so I'm going to advance tomorrow and see where it takes
us, thank you for the awesome. :)
> 1 - Make certmaster daemon more stable, and fix bug which causes
it to
issue malformed certs
> 2 - Improve delegation parallelism
>
> If you have any ideas for the roadmap, and/or would be interested in
taking ownership in something, let me know, as open discussions like this
will drive Func going forward, and I can promise beer and committer status
:)
Maybe IPv6 support and dynamic module loading can be added to the list.
Also I think it would be helpful to everyone if you release a tag a new
version as "git diff --stat v0.28" shows a enormous amount of changes.
LOL that number is staggering, and your suggestion is a good one. I've
created a v0.29 tag to version the current state of the repository, and
from here we can begin working on v0.30 if that's cool.
Thanks for your hard work on Func so far, I'm looking forward to our future
collaboration!
> Thanks much for the read, let me know what you think about all
this, and
have a great weekend!
> --
> Steve Salevan
> steve(a)tumblr.com
>
> _______________________________________________
> func mailing list
> func(a)lists.fedorahosted.org
>
https://lists.fedorahosted.org/mailman/listinfo/func
Cheers,
--
S.Çağlar Onur <caglar(a)10ur.org>
_______________________________________________
func mailing list
func(a)lists.fedorahosted.org
https://lists.fedorahosted.org/mailman/listinfo/func
--
Steve Salevan
steve(a)tumblr.com