On Thu, Jan 20, 2005 at 11:47:35AM -0500, Randy W. Sims wrote:
> John Peacock wrote:
> >Yitzchak Scott-Thoennes wrote:
> >
> >>If the goal is including M::B in 5.10, stick it in now and deal with
> >>what needs to be dealt with.
> >
> >
> >I have to agree. M::B would be a valuable addition to the core, and
> >doing so sooner rather than later is a Good Thing. I'm going through
> >and modifying all of my module distributions (both CPAN and private) to
> >use M::B, since it is much easier to use on Win32 and other odd
> >platforms, like OpenVMS, though I haven't actually installed M::B there
> >yet...
>
> Although I've previously disagreed with this, wanting to see a few other
> key features implemented in M::B before integration, I'm now starting to
> agree. M::B is very stable at this point. It is complete enough to allow
> pure perl distributions and most XS modules to build. The only really
> big feature missing at this point is recursive builds.
>
> There are a few loose ends that would be nice to tie up first, or at
> least during integration: 1) M::B prefers to use ExtUtils::ParseXS,
> which is a modularized version of a snapshot of xsubpp-we need to figure
> out what to do with that first: preferable update, integrate, and
> replace xsubpp with a call thru script. 2) The next release of M::B
> (0.27) factors out the compiler driver into a separate module,
> ExtUtils::CBuilder, along with some other minor refactoring. I think we
> should get that completed first, released, and make sure it's as solid
> as the 0.26 series now is; I think it's the version that should be
> integrated.
>
> However, I do think we should allow CPANPLUS to be completely integrated
> (with its dependencies) first and allow it time to be smoked and fit in.
> And then we can discuss it more fully.
I somehow missed this last paragraph earlier. I disagree; there's no
reason to delay working on integrating M::B just because the same is
happening for CPANPLUS. (All the more so now that CPANPLUS integration
has run into the rock of portable subprocess handling.)
Ken, Rafael (and others), I'm going to work on integration of
ExtUtils::{CBuilder,ParseXS}. So far, looks like the tests need some
tweaking to work in the perl core, and Ilya's xsubpp patch needs to
be added to ParseXS.pm. When I have things working, I'll first send
patches to the CPAN distributions for Ken to eyeball, and then a patch
to perl to add what should be identical files to the core.
As far as M::B itself goes, I'd say the 0.27 branch should get added to
core ASAP, even though there is ongoing work being done on it. Let's
get the ball rolling, here. There's no reason further changes can't
be added as they come.
Optionally, perhaps Ken would appreciate commit access for perl so he
can keep bleadperl up to date with M::B development periodically?
