On Sun, Sep 21, 2014 at 04:15:22PM -0700, Jeremy Evans wrote: > On Sun, Sep 21, 2014 at 12:09 PM, Marc Espie <es...@nerim.net> wrote: > > > On Sun, Sep 21, 2014 at 08:38:44PM +0200, Landry Breuil wrote: > > > On Sat, Sep 20, 2014 at 02:11:04PM -0700, Jeremy Evans wrote: > > > > This makes ruby 2.1 the default ruby version. Now that ruby 2.1.3 has > > > > been released, it makes sense to switch the default from ruby 2.0 to > > > > ruby 2.1. > > > > > > Thinking out loud, but is there still a point in having 4 different > > > versions in the tree ? We tried to reduce the number of pythons and > > > gccs... > > > > > > Landry > > > > I hope ruby 1.8, at least, is on the way out. This is getting > > ridiculous. > > > > The majority of ports that embed ruby still use ruby 1.8. Some use ruby > 1.9, 2.0, or 2.1. I'm fine with removing old ruby versions, assuming that > nothing else in the tree depends on them. Unfortunately, that doesn't > happen to be the case currently. Now, nothing in the tree embeds rubinius > (or can, since rubinius doesn't use a shared library), and rubinius takes a > lot longer to build and is less useful since the version we ship doesn't > use JIT most of the time (since it rarely supports the ports version of > LLVM). If the thought is we just have too many ruby interpreters, I'd vote > to remove that one. FWIW, I think all ruby interpreters and libraries in > ports take less time to build than pypy. :)
Number of ports is important as well, unfortunately. There is a huge chunk of time spent waiting for dependencies to install and for the disk to unpack/repack stuff. Death of a thousand cuts.