[Catalyst] Module::Build vs Module::Install

[Sam Vilain]

On Wed, 2005-11-23 at 16:12 +0000, Chisel Wright wrote:
> Hopefully I won't start any religious wars, but I'm interested to know
> the reasons for moving to Module::Install
> 
> I don't know either module intimately, and M::B seemed to do the job
> pretty well. What was it lacking that M::I provides?


17:48 < sungo> Module::Install++
17:48 < sungo> Module::Build sucks balls
17:49 < rjbs> sungo: what does MI get me over EUMM?
17:49 < rjbs> sungo: I don't to really interesting things with EUMM, at
              present.  Is MI going to help me?
17:49 < sungo> a very simply interface with a Makefile that works just
               like  Makemaker's shit
17:49 < sungo> go take a look at the Makefile.PL for
               POE::Component::Server::Syslog
17:49 < Alias> MI is sane and Just Works
17:50 < Alias> I cannot speak of it highly enough
17:50 < Alias> And it even reduces dependencies in some situations
17:50 < rjbs> Alias: how? because you can say build_requires, or?
17:50 < Alias> yep
17:50 < rjbs> ok.
17:50 < sungo> hell, it can bundle your dependencies
17:50 < Alias> only for test-time
17:51 < rjbs> I know it can do lots of neat things.
17:51 < rjbs> I just feel like they're all things I don't need.
17:51 < Alias> build_requires( 'Test::Module::PurePerl' ); include( 
               'Test::Module::PurePerl' );
17:51 < sungo> and it bundles itself so your users don't have to install
               yet another module just to install yours
17:51 < rjbs> I'll try repackaging Rubric tomorrow, maybe, and see how
              that goes.
17:51 < Alias> That last point being the most important
17:51 < rjbs> sungo: yeah, that part is nice.
17:51 < Alias> Because no matter how old and shitty the version of Perl,
               you can still get sane installation code
17:52 < sungo> it was the right solution. something easy to use that
               maintained makemaker compatibility
17:52 < sungo> unlike certain OTHER frameworks