Use override_settings rather than mucking with settings in tests

Second attempt to address possible test sequencing issue around tests that alter settings. (Failures in Travis tests not reproducible locally.)

Back-ports override_settings from Django 1.4 for compatibility with Django 1.3.

djrill/tests/mock_backend.py

@@ -1,19 +1,22 @@
 import json
 from mock import patch
 
-from django.conf import settings
 from django.test import TestCase
 
+from .utils import override_settings
 
+
+@override_settings(MANDRILL_API_KEY="FAKE_API_KEY_FOR_TESTING",
+                   EMAIL_BACKEND="djrill.mail.backends.djrill.DjrillBackend")
 class DjrillBackendMockAPITestCase(TestCase):
     """TestCase that uses Djrill EmailBackend with a mocked Mandrill API"""
 
     class MockResponse:
         """requests.post return value mock sufficient for DjrillBackend"""
-        def __init__(self, status_code=200, content="{}", json=['']):
+        def __init__(self, status_code=200, content="{}", json=None):
             self.status_code = status_code
             self.content = content
-            self._json = json
+            self._json = json if json is not None else ['']
 
         def json(self):
             return self._json
@@ -23,15 +26,8 @@ class DjrillBackendMockAPITestCase(TestCase):
         self.mock_post = self.patch.start()
         self.mock_post.return_value = self.MockResponse()
 
-        settings.MANDRILL_API_KEY = "FAKE_API_KEY_FOR_TESTING"
-
-        # Django TestCase sets up locmem EmailBackend; override it here
-        self.original_email_backend = settings.EMAIL_BACKEND
-        settings.EMAIL_BACKEND = "djrill.mail.backends.djrill.DjrillBackend"
-
     def tearDown(self):
         self.patch.stop()
-        settings.EMAIL_BACKEND = self.original_email_backend
 
     def assert_mandrill_called(self, endpoint):
         """Verifies the (mock) Mandrill API was called on endpoint.