[Asterisk-Dev] benevolent dictatorship, or inclusive developper community?

Martin Kihlgren zond at troja.ath.cx
Tue Jan 6 19:46:24 MST 2004


<IMHO>

On Tue, Jan 06, 2004 at 09:28:20PM -0500, asterisk at lists.styx.org wrote:
> Some minimal amount of testing must be done before the changes go into
> CVS. But, it seems to me that CVS should be, explicitly, caveat utilisator.
> It should not be guaranteed to compile or run or do anything useful other
> than to have people testing new stuff.
> 
> CVS code /is/ for testing.
> 
> It emphatically should not be the rule to be using CVS code in production.
> This is why releases should be made from time to time, as often as
> possible, of known or thought good code. People that run this stuff
> in production should run releases. CVS is for developers only.

I agree totally, except for the mentioning of compile. The cvs _should_
compile_, at least. 

> Yes, I agree, Brian has done a lot of very good work and is a great
> boon to the Asterisk community. But from my experience, what I can see
> happening is, someone writes a patch and has to spend so long pestering
> you to get the patch imported into CVS that repeating the experience
> begins to seem a waste of time. So they lose interest.
> 
> As well, if CVS code comes with no guarantees, and people are strongly
> warned to run the stable snapshots in production, letting more than one
> single person commit changes becomes less of a problem. If someone commits
> something that's wrong, just take it out. No harm done. It moves the
> dialogue of the development team into the source code where it should be
> and out of IRC, which has its place but is not where everyone wants to
> spend their time.

Yes, multiple persons with separate responsibilities having commit-rights is
the only proper way of doing open-source development.

That would 1) alleviate the workload of the project leader 2) make it easier
to get a dialogue with the person responsible for (and knowledgeable of) the
part you want to patch (to improve the patch and understand possible
problems with it) 3) speed up development 4) broaden the feature-set (not
just the features one person appreciates will get commited).

All good things, no?

Of course, it would also let through some inferior patches, but that is what
version-management is meant to fix, right?

</IMHO>

//Martin

-- 
###################################################################
My idea of roughing it is when room service is late.
###################################################################
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 827 bytes
Desc: not available
Url : http://lists.digium.com/pipermail/asterisk-dev/attachments/20040106/c6709a41/attachment.pgp


More information about the asterisk-dev mailing list