On Jun 14, 2006, at 1:03 AM, Adam Kennedy wrote:
As with several other iterations of this debate, though, this
consists of you pointing out some flaws, and someone else trying
to calm you down or ameliorate the debate. I do appreciate your
thoughts and analysis of the situation, but patches (or even
proposed solutions) speak a lot louder than speaking loudly.
Most often from my perspective it consists of being told the
problem either a) isn't a problem, or b) doesn't impact many people
at the moment, or c) We'll look into it.
Yeah, I can see how that would be frustrating. However, I think b)
and c) are indeed true. =)
Because I'm not a M:B user, and not really part of the "community"
I haven't felt it's really my place to make large changes to the
design of M:B.
Well, probably not to *make* them, but certainly your input is welcome.
If you are requesting design proposals, I am happy to suggest some.
Certainly I would welcome any proposals. There are about a million
different ways we could go, and it would be great to get a sense of
what some of people's ideas are.
Actually, just a document with a set of short, coherent statements
about what the problem(s) is(are) would be a good start. You'd be a
good person to take a first crack at that, I think?
As a starting point, I think that the idea of CPANPLUS
instantiating Module::Build objects directly has to go, and favour
of a CPAN.pm style loosely coupled execution approach, as others
have mentioned.
Yes, I think that would give us the biggest bang for the buck at this
point. Clearly we'll still have work to do afterwards, but it's a
good first step.
The CP+ people turned over the CP+::Dist::Build module to us, so if
someone wants to look into this and submit patches we could get them
tested and released without waiting for those guys to act.
-Ken
