From: jweirich@...
Date: 2001-08-20T06:15:27+09:00
Subject: [ruby-talk:19993] pose_as

I was reading someone commenting about Objective C yesterday.  They
said that Objective C was able to do 

    [NewClass poseAs: [OldClass class]];

(note Smalltalk like syntax).  Thereafter any time [OldClass new] was
executed, an object of class NewClass would be created instead.

This sounded like a useful thing to have when doing testing
(e.g. replacing the construction of a real object with a testable Mock
version), so I wondered if the same could be done in Ruby.  Here's the
result ...

----------------------------------------------------------------------
class Class
  def pose_as(other_class)
    if other_class.methods.include?("_unposed_new_")
      fail "Another class is already posing as #{other_class.name}"
    end
    other_class.class_eval <<-EOS
    class << self
      alias _unposed_new_ new
      @@use_original = false
      def new(*args)
        if @@use_original
          _unposed_new_(*args)
        else
          @@use_original = true
          result = #{name}.new(*args)
          @@use_original = false
          result
        end
      end
    end
    EOS
  end

  def unpose
    if not self.methods.include?("_unposed_new_")
      fail "No class is posing as #{self.name}"
    end
    self.class_eval <<-EOS
    class << self
      alias new _unposed_new_
      remove_method :_unposed_new_
    end
    EOS
  end
end
----------------------------------------------------------------------

Load the code and then saying

    MockThing.pose_as(Thing)

will cause all evaluations of Thing.new to return a MockThing.new
object instead.  Say:  Thing.unpose  to undo the effect.

A couple of limitations ...

o MockThing must have a compatible version of initialize().
o Only one level of posing is allowed, exception otherwise.
o The posing code is not thread safe (we use a class variable flag
  ... elegant solutions to this limitation is cheerfully solicited). 

I'm not suggesting this as a standard part of Ruby, but I do think it
is a tribute to the flexibility of the language to allow expressing
this kind of stuff within the language.

-- 
-- Jim Weirich     jweirich@one.net    http://w3.one.net/~jweirich
---------------------------------------------------------------------
"Beware of bugs in the above code; I have only proved it correct, 
not tried it." -- Donald Knuth (in a memo to Peter van Emde Boas)