SI- Because rvm and rbenv are overkill
It's said that more than any other programming community Lispers tend to dislike working on problems that have already been solved. But this dogmatic avoidance of duplication doesn't seem to have permeated the rubyists' zeitgeist to the same degree.
I'm not in the habit of regularly compiling ruby interpreters for the hell of it- it's something I'd do maybe twice a year tops. I recently rebuilt rubinius and ruby 1.9 for Lion and that'll probably be it for 6 months. If you keep all your interpreters in userland and you aren't such an idiot that you can't install them yourself, how is it that the nebulous concept of a ruby version manager has come into vogue?
Switching interpreters shouldn't be any more complex than having a script to symlink the executable paths for you:
A couple of clarifications:
- I install all my rubies into ~/.ruby/
- The current ruby gets all its executables symlinked into ~/.ruby/current/bin
- I added ~/.ruby/current/bin onto the end of my $PATH
- That is all
I have it saved in /usr/local/bin/si. SI is short for 'switch interpreter'. A monkey can use it, although none have been known to do so.
The hash of interpreters is hard-coded and I can't even remember what else I did to get it working but that's not the point. It's a 15 minute job I tackled and forgot about 2 years ago. Let's use some initiative, stop writing version managers and do some bloody work.
