Re: [Tinycc-devel] Newer tinycc repository?

[KHMan]

Okay, I'm breaking my word by replying, but I'll just put in a few
more comments. :-)

Peter Lund wrote:
> On Sat, 2007-10-27 at 19:14 +0800, KHMan wrote:
> 
>> I guess it depends on who becomes the next maintainer. Does the
>> community want "Rob II"?
> 
> I think we want "Rob without the Asperger's" ;)

I agree absolutely. I respect everyone developers' skills, but
sometimes management skills is also desirable. Just see the cygwin
or mingw/msys lists. :-) The way I see it, a whole lot of energy
is being pissed away all the time...

> What we will probably get is a loose band of people, some of whom
> sometimes find the time to work on tcc for a bit and then drift back to
> being passive members of the group again.

I agree absolutely. The community is mostly part-timers or
opportunistic participants.

> [snip]
> If we can find a good way to make that kind of involvement work, we
> should be able to make progress again without having an annointed
> maintainer.

So we need person(s) to accept patches and do basic quality
assurance. Either way (committers or maintainers), person(s) will
be at the focal points. If there is no one anointed maintainer,
then I guess we need multiple committers.

> [snip]
> Depends on how hard it is to include the contents of that repository
> into the shared pool of code.  Mercurial gives us a good chance of being
> able to opportunistically use people's locally hosted repositories while
> they are active without losing anything, without the sweat and hard
> work, and without a select group of CVS committers.

I was really happy when David Wheeler volunteered to sync the
repositories. Of course, such a scheme is doable. But the
repository's owner should take the bit of time and effort to care
for the community's interests. Say, for each Mercurial commit to
the main trunk, generate a patch automatically and mail it
automatically to a CVS committer. You know, something like that,
put in solutions rather than participate in endless debates.

> [snip]
> Did the official CVS repository work?

In the end, my personal opinion is that revision control systems
matters much less than having good maintainers (or even just
commit monkeys) that can unite the community, keep discussions
from blowing up, and set the correct tone (i.e. one that
encourages participation and diversity rather than insist on
strict obeisance to the maintainer's practices, tools and
philosophy.) But when it comes to tcc's next step, I would be
happy with anything that works with the minimal disruption, both
short-term and long-term. As I have opined, it all depends on who
wants to step up and seize the bull by its horns.

-- 
Cheers,
Kein-Hong Man (esq.)
Kuala Lumpur, Malaysia