From: Paul Brannan <pbrannan@...>
Date: 2002-04-13T00:54:34+09:00
Subject: Re: Idea for a new shorthand

On Fri, Apr 12, 2002 at 03:59:51PM +0900, Hal E. Fulton wrote:
> What do you think?
> 
> Hal Fulton

Does this mean that when a method is redefined, the old definition must
ALWAYS stick around indefinitely, just in case I do something like:
  eval("prior")

[ruby-talk:29122] does propose a solution that gives you the
functionality you propose, though it uses super() and not prior().
There is no way to call the real super(), though, unless the "parent
class" is really a module.

(BTW, is it a bug in UnboundMethod#bind that the following doesn't work?

  class Base; def foo; puts "Base#foo"; end; end;
  class Derived < Base; end
  d = Derived.new
  Base.instance_method(:foo).bind(d).call()

If I change line 6709 of eval.c from:

  if (!((TYPE(data->oklass) == T_MODULE) ?
        rb_obj_is_kind_of(recv, data->oklass) :
        rb_obj_is_instance_of(recv, data->oklass))) {

to:

  if(!rb_obj_is_kind_of(recv, data->oklass)) {

then the above works, and doesn't appear to have any odd behavior.

)

Paul