Consider: A) >>> class C(object): ... pass ... >>> def f(*args): ... print args ... >>> C.f = f >>> C.f <unbound method C.f> >>> c=C() >>> c.f() (<__main__.C object at 0x04A51170>,)
And B) >>> del c >>> C.f = types.MethodType(f, None, C) >>> C.f <unbound method C.f> >>> c = C() >>> c.f() (<__main__.C object at 0x04A51290>,) I don't understand A). It is my vague understanding, that methods are really properties that handle binding on attribute access, so B) should be the "right" way to add a method to a class after definition. Why does A show up as a method? Shouldn't it still just be a function? Certainly when you define a class, there is some magic in the __new__ method that turns functions in the initial dictionary into methods, but does this still happen for all setattr after that? Is is possible to set a class attribute equal to a regular (types.FunctionType) function? Any references that discuss these issues would be greatly appreciated. Thanks, Gerard -- http://mail.python.org/mailman/listinfo/python-list
