[...] consistency with the existing specs I ended up putting them in kernel/frozen_spec.rb. Was this the best choice?

Looks fine to me.

The expectation for the new quoted symbol syntax uses an eval because otherwise it would break the whole file (the parser would throw a syntax error and no examples would be run).

We'll probably need to re-generate the parser first, then we should be able to remove the use of eval.

Hey @yorickpeterse, thanks for the feedback. I'll probably write a few more language specs in the next couple of days. I'll update the PR when I do.

Hi everyone,

• Added block and lambda literal specs for MRI issue #9593 (optional argument accesible in its default value expression)
• Updated hash_spec according to MRI issue #10315 (ChangeLog)
• Removed assertion checking that U+180E matches /[[:blank:]]/, as it's no longer considered a whitespace character

At this point I still see 3 failed specs with MRI 2.2.1 in spec/ruby/language, but I suspect one or two of those to be MRI bugs. I opened this issue, and I'm still waiting for feedback.

@ruipserra Any feedback on the Redmine ticket you submitted? Sometimes it can be a bit hard to get a hold of Matz & co, in that case replying to the ticket (and specifically mentioning somebody like Koichi or Nobu) sometimes can do the trick.

Unless this particular Redmine ticket is blocking I'll take a look at this PR this weekend.

Fail, I didn't see this PR until now.
Many of these commits are already on the 2.2 branch.
How should we proceed?

I think this needs a rebase as there are a bunch of commits that aren't yet on our 2.2 branch.

The easiest way to do this is probably cherry-pick the commits from here to the 2.2 branch so we can close this and avoid the confusion. If @ruipserra, @yorickpeterse, or @tak1n can't do this, I'll do it in a day or two.

Commits rebased into the 2.2 branch, thanks!

Awesome thanks :)