I don't understand the behavior of type with weakref?

Question

Python

Hello Habr.
I have always been sure that in Python for objects of new classes the attribute x.__class__and type(x)will have the same value. However, if you do this (tested with Python 2.7.3 and 3.3.0, and PyPy 2.0b1):

>>> import weakref
>>> x = set()
>>> y = weakref.proxy(x)
>>> x.__class__, type(x), isinstance(x, set)
(<class 'set'>, <class 'set'>, True)
>>> y.__class__, type(y), isinstance(y, set)
(<class 'set'>, <class 'weakproxy'>, True)

we'll see that for weakref.proxyy.__class__matches the wrapped type (I guess weakref.proxyjust spoofing an attribute to mimic the wrapped object), and even isinstancerecognizes it as a set. But at the same time , it typeindicates its true type - weakproxy. It turns out that it typedoes not use its attribute to determine the type of an argument __class__. It turns out that he typeuses some "more reliable" source to determine the type? Is it possible to access it?

Answer 1

The answer is this.
Inside the interpreter (CPython), each Python object is represented by a C structure PyObject(file object.h), in which one of the fields is ob_type, a pointer to a structure _typeobjectthat defines the "real" type of the object. type(x)and descriptor object.__class__refer to the type of this field. But __class__it can be overridden, which is what weakref.proxy. At the same time type(x), it will still continue to access ob_typeand the result of its determination cannot be changed.

Answer 2

read this book www.cafepy.com/article/python_types_and_objects/python_types_and_objects.html