On 11/19 07:51, Stuart Henderson wrote: > On 2010/11/18 15:49, Jeremy Evans wrote: > > Here's a new version of the port, after feedback from lan...@. The RUBY > > and RAKE variables are now used, REGRESS_DEPENDS has been fixed, and the > > port now installs to /usr/local/lib/rubinius. I've also copied a patch > > from lang/ruby/1.8 that makes the same change to FileUtils#mkdir_p so > > that you don't get systrace warnings when building ports. Additionally, > > I've precreated the gem directories and site_ruby directory, so that > > the existing PLISTs for gem ports will work. > > > > I've also made the changes to ruby.port.mk so you can build rbx > > versions of the gem and extconf ports, which I've tested > > successfully on about 10 different ports. The ruby.port.mk > > changes are inlined below. > > > > Tested on i386. Looking for OKs. > > Generally ok but a couple of questions, > > > +#.poison MODRUBY_BINREV > > +#.poison MODRUBY_PKGSPEC > > +#.poison MODRUBY_WANTLIB > > did you intend to add these commented-out? Yes. The same is currently done in the jruby block a few lines down, with a comment explaining why, but I didn't want to add a duplicate comment. It suppose it would make more sense for the comment to be moved into the rbx block, or maybe for the rbx block to just be added after the jruby block instead of before. I think moving it to after the jruby block is simpler, so I'll do that.
> Also for the rubinius port, do you think it would make sense to > add something like this and regenerate the PLIST? > > RAKE_VER = 0.8.7 > RAKE_COMPILER_VER = 0.6.0 > RDOC_VER = 2.5.1 > RUBY_DEBUG_VER = 0.10.47 > SUBST_VARS += RAKE_VER RAKE_COMPILER_VER RDOC_VER RUBY_DEBUG_VER That couldn't hurt and would make for less PLIST churn when they update the gem versions, so I'll do that. Thanks, Jeremy
