A few of the failures look bogus, but most look valid (visibility errors). Something's not right.

I'll take a look through them and reconcile with MRI.

There is definitely something weird going on here.

This spec is up to date but the results are wrong for MRI too. It's checking that initialize_clone doesn't appear in the public methods not that an exception is raised.

https://github.com/rubyspec/rubyspec/blob/master/core/kernel/initialize_clone_spec.rb

https://gist.github.com/tduehr/9144875

Is there a way to remove these methods from the methods listings without a special case in #methods for now?

@tduehr No simple way to remove those methods from the listings. I'm going to poke at this a bit.

I think we need to update master to latest rubyspec; many of these cases have been fixed as being bad tests.

It looks like other specs fail because mspec is not handling respond_to_missing? being private correctly.

Ok, I updated two specs (Kernel#initialize_clone and initialize_dup) and updated mspec and things appear to passing ok locally. We'll see how travis feels, but for now it's merged.

Any chance we can get a similar PR for missing private for Ruby 18/19 on jruby-1_7? I am betting we are missing a few there as well.

#1542