Why mocking is still effective after existing a code in nosetests multiple tests?

Question

I have the following code test_A.py which mocks MyClass.mymethod:

from unittest import main
from mocker import Mocker, MockerTestCase
Class test_A(MockerTestCase):
  def setUp(self):
    self.m=Mock()
    MyClass.mymethod = self.m.mock()
    self.m.result(None)
    self.m.count(0,None)
    self.m.replay()

  def test_me(self):
    #Do something about MyClass.method

  def tearDown(self):
    self.m.restore()
    self.m.verify()

I also have another code test_B.py which DOES NOT mock MyClass.mymethod:

Class test_B(MockerTestCase):
  def setUp(self):
    pass

  def test_me(self):
    #Do something about MyClass.method

  def tearDown(self):
     pass

However, when I do "nosetests test_A.py test_B.py", it looks like after testing test_A.py and entering test_B.py, MyClass.mymethod is still mocked up. Not sure why and how to get around it. Thanks!

Answer 1

The line:

MyClass.mymethod = self.m.mock()

really does replace MyClass.mymethod() with a new object. All subsequent references to MyClass.mymethod will be to the mock object, even if those references are in a different class.

What you want is a way to replace MyClass.mymethod() that only works in class test_A. The simplest way to achieve this would be to restore the original mymethod in your tearDown method:

Class test_A():
    def setUp(self):
        self.originalMyMethod = MyClass.mymethod
        self.m=Mock()
        MyClass.mymethod = self.m.mock()
        self.m.result(None)
        self.m.count(0,None)
        self.m.replay()

    def test_me(self):
        # Do something about MyClass.mymethod

    def tearDown(self):
        self.m.restore()
        self.m.verify()
        MyClass.mymethod = self.originalMyMethod