CVS for Cfengine (Was: Re: Bugzille for cfengine)

[Eric Sorenson]

On Tue, 14 Dec 2004, John Sechrest wrote:

> It would be good to have a stronger idea of who wants to
> engage in coding. And what the process for getting
> things into the archive should be.
>
> My expectation is that a small number of people will take on the
> developer role and that others will decide to take on a supporting
> role for submitting patches. But I could be way off. My current
> plan is to move slowly to set up the system. My intent was to
> set up CVS, because it was supported by Sourceforge.

John -- there are many points worthy of discussion in this and
subsequent emails, but I wanted to focus on this comment. I would
argue that, rather than move slowly to get everything planned and
set up before going "live" on sourceforge, we ought to move quickly
to get one critical piece working, and let the rest of it fall into
place "as slowly as possible; as fast as necessary".

Specifically, I would very much like to see the cfengine project
CVS tree on sourceforge populated and *written to by Mark* as quickly
as is reasonable, perhaps before the end of the year. Everything
else -- mailing lists, bug tracking, developer access, *everything*
is secondary to this. If that is the only change that happens to
the cfengine development process, and nobody else ever ends up 
getting write access, that would be fine with me. But I plead to
both you and Mark, please work to make this happen ASAP.

Over-dramatic? Maybe. But I broke 1200 production machines for several days
last week (due to the 'cfagent -Q' splaytime behavior), and I really think it
would not have happened if we had a systematic way of comparing code changes
and commit logs between cfengine releases. You setting up the CVS and
Mark writing to it is a very small, incremental step that nets huge wins
for everyone out in the trenches relying on cfengine every day.

--

 - Eric Sorenson - Explosive Networking - http://eric.explosive.net -