Getting Python's unittest results in a tearDown() method (my SO answer, question 4414234)

results_in_teardown.py

"""Getting Python's unittest results in a tearDown() method
https://stackoverflow.com/questions/4414234/getting-pythons-unittest-results-in-a-teardown-method#39606065
"""
import unittest


class MyTest(unittest.TestCase):
    def tearDown(self):
        if hasattr(self._outcome, 'errors'):
            # Python 3.4 - 3.10  (These two methods have no side effects)
            result = self.defaultTestResult()
            self._feedErrorsToResult(result, self._outcome.errors)
        else:
            # Python 3.11+
            result = self._outcome.result
        ok = all(test != self for test, text in result.errors + result.failures)

        # Demo output:  (print short info immediately - not important)
        if ok:
            print('\nOK: %s' % (self.id(),))
        for typ, errors in (('ERROR', result.errors), ('FAIL', result.failures)):
            for test, text in errors:
                if test is self:
                    # the full traceback is in the variable `text`
                    msg = [x for x in text.split('\n')[1:]
                           if not x.startswith(' ')][0]
                    print("\n\n%s: %s\n     %s" % (typ, self.id(), msg))
                    
    def list2reason(self, exc_list):
        if exc_list and exc_list[-1][0] is self:
            return exc_list[-1][1]

    # demo tests
    def test_success(self):
        self.assertEqual(1, 1)

    def test_fail(self):
        self.assertEqual(2, 1)

    def test_error(self):
        self.assertEqual(1 / 0, 1)

    @unittest.expectedFailure
    def test_expected_fail(self):
        self.assertEqual(2, 1)

    @unittest.expectedFailure
    def test_expected_error(self):
        self.assertEqual(1 / 0, 1)

    @unittest.expectedFailure
    def test_unexpected_success(self):
        self.assertEqual(1, 1)

@hynekcer this was very helpful

I hope you don't mind me borrowing this 💎

I've changed a bit with extra verbosity, since i'm doing a system test, i.e. I'm running one test at a time, and my tearDowns are too long to get the answers I need:

        if not ok:
            typ, text = ('ERROR', error) if error else ('FAIL', failure)
            TEST_LOG.error("="*70)
            TEST_LOG.error("%s: %s" % (typ, self.id()))
            TEST_LOG.error("-" * 70)
            TEST_LOG.error(text)
            TEST_LOG.error("-" * 70)

Yes I agree with every use of this snippet or my answer on Stackoverflow. This solution can be even extended to extract a traceback (are you interested?), but it was not was not a subject of the question and a short clear code is usually preferred on SO.

Thanks, mentioned my extra verbosity here and not in SO, for the exact reason, whom ever find this (since it's not linked from SO) might be interested :)

the one downside of this change, it's make my argument of switch to pytest right now a bit less strong 😞
(Since I know how to get this working in pytest, with a few lines in conftest.py)

Updated for Python 3.11

Hi, I am really interested in extending this to extract a traceback when a test fails

@SissiWei123 It is easy: The full traceback is in the variable text. I added it as a comment to the code

Note that the Python 3.11 approach doesn't work if the tests are run with pytest and the pytest developers have no no intention of making it work, since the _Outcome is a private attribute. The 3.4 to 3.10 code works fine when pytest is used to run unittest based tests.

Note that the Python 3.11 approach doesn't work if the tests are run with pytest and the pytest developers have no no intention of making it work, since the _Outcome is a private attribute. The 3.4 to 3.10 code works fine when pytest is used to run unittest based tests.

Why do you need this trickery when using pytest as a runner ? pytest offer much better alternatives to this hack

Why do you need this trickery when using pytest as a runner ? pytest offer much better alternatives to this hack

We don't need it, but with Python 3.4-3.10 the original hack works the same both with and without using pytest.

In this way, the results of pass, fail, and error can be obtained, but it is impossible to know whether the use case is skipped. is there any way to know