Urabe Shyouhei wrote:
> Ruby based program to check ruby itself?  Does that really reliable?
>   
I think so. Stephen's RubyBuild can build a Ruby interpreter and my 
RubyCheck can run a battery of tests against it. If either of these 
simple Ruby programs fails, there's something wrong with the 
interpreter, which is what we're trying to catch anyway. Is this 
reasonable or is there a need for a shell-based solution?

> Rake is a great software but we are standing at the point where MRI is
> suspicious to be great, when we run tests for it.
>   
My Ruby version of RubyChecker uses only the core and standard 
libraries. I would have liked to use Rake, but it's not available 
because my code is the part that's responsible for conditionally 
installing RubyGems and Rake into the interpreter under test.

-igal