From: "Chr. Rippel" <chr_news@...>
Date: 2002-01-12T09:45:07+09:00
Subject: Re: class name reported differently in different contexts

"ts" <decoux@moulon.inra.fr> wrote in
....
>  With this patch, classid is useless and even for "native" classes it can
>  exist only classpath.

Sorry Guy the first example from my previous post was a little bit
bogus. What I mend that to say if you consider a script like

------
class Module
  private
  def const_get_scoped sym
    if sym.is_a? (Symbol)
      Object.__const_get_scoped (sym.to_s)
    else
      Object.__const_get_scoped sym
    end
  end
  protected
  def __const_get_scoped sym
    if md = (/(::)/).match(sym)
      (const_get (md.pre_match)).__const_get_scoped md.post_match
    else
      const_get sym
    end
  end
end

class Klass
  def self.inherited(sub_klass)
    const_get_scoped (sub_klass.inspect)
  end
end

class X
  class Y
    begin
      class Z < Klass; end
    rescue NameError => mes
      puts mes
    end
  end
end
------

than the output would be something like

1) uninitialized constant Z
2) uninitialized constant Z at X::Y
3) wrong constant name #<Class:0xa046728>

depending on what strategy is implemented.  I am really not sure why you
think the first strategy is so much better. For better or for worse if an
exception occurs during the call to #inherited a Class object exists until
the garbage collector kicks in and there nothing you can do about this.


/Christoph