Code Review: numeric types

tfpt review “/shelveset:num;REDMOND\sborde”

Comment :
Fixnum + Bignum should yield a Bignum, not a Float. This was causing
complex/denominator_spec to fail as Bignum was monkey-patched and not
Float. A different return type affects whether the monkey-patched method
is visible or not. Also changed fixnum/plus_spec to check for the result
BigDecimal coercing works a bit differently. So added a protocol for
Small change to the MSpec eql guard to show the type of actual and
expected values since the inspect string itself can be identical.

Tomas, I had added tests for Enumerable#each_cons each_slice. Let me
know if you want to grab them since you are working on the fix. Else, I
can check it in with tags (will need to move it from library/enumerator
to core/enumerable).

Looks good.