On Mon, Apr 05, 2010 at 09:54:32PM +0200, Aristotle Pagaltzis wrote: > * Ævar Arnfjörð Bjarmason <avarab@gmail.com> [2010-04-05 21:35]: > > But of course having all that stuff in core means maintaining > > it indefinitely, which means more strain on the porters. > > I wouldn't suggest something like "let's drop the top 100 CPAN > > module into core" without *very* widespread consensus on that > > point. > > A while ago I suggested on #p5p that it would be cool if the > whole dual-life tree in the core was reduced down to dropping > actual tarballs from CPAN ??? not even unpacked. That would not > only make it trivial to keep the dual-life stuff in core up to > date with their CPAN upstreams, it would also make it very easy > for anyone to build a batteries-included version of the core. > The suggestion was well received on #p5p, at least. Given that most of the toolchain modules are dual lifed, including the ones needed to unpack these tarballs, it strikes me as tricky to bootstrap this. So I can't see how the whole tree can be done this way. Given that DynaLoader needs XS compilation, and DynaLoader.o needs to be built to build perl, I'm not sure that any part of the current toolchain code could be removed into tarballs, hence I can't envisage any area where this simplifies the existing extension building codebase. (Whilst clearly it adds complexity to support tarballs) Nicholas ClarkThread Previous | Thread Next