From: Charles Thornton <ceo@...>
Date: 2007-03-23T07:28:51+09:00
Subject: MyBad ==> Re: [ ruby-Bugs-9462 ] BUGS in metaclasses inheritance

Charles Thornton wrote:
> Rick DeNatale wrote:
>> On 3/22/07, Charles Thornton <ceo@hawthorne-press.com> wrote:
>>
>>> This chart is essentially correct,  It simply leaves out a few
>>> implementation
>>> details.
>>>
>>> (1)  "Object" super entry includes the ObjectSpace Module
>>
>> Where did that come from?
>>
>> Object includes Kernel, not ObjectSpace.  So that rb_cObject's
>> superclass chain should look like:
>>
>> rb_cObject --> (T_Iclass:Kernel) ---> nil
>>
>> where T_ICLASS:Kernel has:
>>   klass = rb_cClass
>>   iv_tbl = rb_mKernel.iv_tbl
>>   m_tbl = rb_mKernel.iv_tbl
>>
>>
>> Note also that Object.superclass doesn't always return the value of
>> the klass field. It walks the superclass chain and skips past classes
>> marked at T_ICLASS which are proxies for included modules.
>>
> I will look at my dumps again -- But the chart I submitted was
> automatically generated by walking through all objects in
> ObjectSpace and recording the links as they are implemented
> in the Objects that are ACTUALLY there!
>
> Obviously,  while this snapshot was taken after the ruby program
> had run (the call to ObjectSpace.my_short_object), It does reflect
> the actual state of  ObjectSpace at that time.
>
> I include the program and the ObjectSpace Diagram
>
JUST A NOTE -- I took out the "test_object" Instance and singleton for 
the chart --
>
> ------------------------------------------------------------------------
>
> #
> # Simple Test Case for Plot
> #
> include ObjectSpace
> #
> # Create Test Class
> #
> class TestCase
>   def initialize
>   end
> end
> #
> # Define a Singleton Class for 'TestCase'
> #
> #test_object = TestCase.new()
>
> #def test_object.test_singleton_method()
> #  puts ("TestCase Singleton Method")
> #end
> #
> # Dump Object Space (ALL Objects - Including Singletons!)
> #
> ObjectSpace.my_short_object
>