On Fri, Mar 19, 2010 at 12:08 PM, Xavier Noria <fxn / hashref.com> wrote:

> Given:
>
>  class C
>   def self.m; end
>  end
>  class D < C; end
>
> the superclass of the singleton class of D is the singleton class of C.

Just a remark here. That's what Flanagan & Matz says, but the draft of
the spec says that the superclass of an eigenclass is implementation
defined, and that the singleton class of D has the singleton class of
C among its superclasses.

So the question still holds, either the spec is wrong, or there's a
bug I believe. Perhaps I should ask in -core.