Hello, I have trouble understanding the need for the class << self idiom. For example, let's say I have this class: class My_Class def method_1 end end Now, I want to extend that class. I can do: class My_Class def method_2 end end or: class My_Class class << self def method_2 end end end I understand that in the second case we're extending the singleton class of My_Class, but why is the second form so often used in Ruby vs. the other form? Isn't the end result identical?
on 2007-06-20 23:40
on 2007-06-21 00:56
surge wrote: > class << self > def method_2 > end > end > end > > I understand that in the second case we're extending the singleton > class of My_Class, but why is the second form so often used in Ruby > vs. the other form? Isn't the end result identical? They are *not* identical. The first defines an *instance* method (My_Class.new.method_2); the second defines a *class* method (My_Class.method_2 Now, there are other ways of defining class methods. Which one you use depends a lot on personal preference but depends also on constants (which have a static scope) -> class Foo K = "foo" end #can access K class Foo class << self def foo1; K; end end def self.foo2; K; end def Foo.foo3; K; end end #must specify Foo::K class << Foo def foo4; Foo::K; end end def Foo.foo5; Foo::K; end Daniel
on 2007-06-21 01:21
You're completely right and I understand my mistake now. Thanks for the comprehensive answer. Surge