B-2 (covariant returns) and vtable order

[Jim Dehnert]

I've updated the layout document (which now contains all the changes
that should have been there last night, too).  What I did today was
to combine the vtable content and order sections, and attempt to
clarify the ordering questions that have arisen.  Look for red, as
usual.

I specified new entries for covariant returns in the optimal fashion
suggested by Jason, i.e. no new entry if overriding a vfunc from the
primary base that requires no return adjustment.  (Note: that is not
the same as requiring the same type.)  If, after looking this over,
we can't quickly agree to this solution (or a same type condition),
I'll reopen B-2.

Send your notes...

Jim
-		Jim Dehnert  x3-4272