nose_test_generators.py

class TestActiveField(object):
    models_to_test = ['Model1', 'Model2', 'foo']
    
    def test_valid_choice(self):
        for model in self.models_to_test:
            yield self.check_valid_choice_for_active_field, model

    def check_valid_choice_for_active_field(self, model):
        # implement test logic here
        assert 'foo' != model, "Model was %s" % model

#
# results in:
#

# views.base.TestActiveField.test_valid_choice('Model1',) ... ok
# views.base.TestActiveField.test_valid_choice('Model2',) ... ok
# views.base.TestActiveField.test_valid_choice('foo',) ... FAIL
# 
# ======================================================================
# FAIL: views.base.TestActiveField.test_valid_choice('foo',)
# ----------------------------------------------------------------------
# Traceback (most recent call last):
#   File "/home/jlilly/.virtualenvs/dashboard/lib/python2.5/site-packages/nose-0.11.2-py2.5.egg/nose/case.py", line 183, in runTest
#     self.test(*self.arg)
#   File "/home/jlilly/src/dashboard/tests/views/base.py", line 13, in check_valid_choice_for_active_field
#     assert 'foo' != model, "Model was %s" % model
# AssertionError: Model was foo
# 
# ----------------------------------------------------------------------
# Ran 3 tests in 0.002 seconds
# FAILED (failures=1)

Why is that better than:

def test_valid_choice(self):
    for model in self.models_to_test:
        self.check_valid_choice_for_active_field(model)

You get every single case as a failure or success.