This is my own hacky micro test harness for Python unittests. It is currently in use in a number of projects with large test cases. However, you might want to checkout "nose" or "unittest2" first. :) Some features include:
- small, you can just put this one file in your "test" dir
- no external dependencies other than the stdlib
- tagging of individual and groups of tests to allow easily running test subsets (can be very helpful in huge doc sets)
- support for "TestSkipped" as a return from a test (TODO: has modern unittest.py added this now?). This is helpful for tests of platform-dependent or optional features.
A typical Python project setup (at least for me) is:
foo/ # project foo README.md setup.py lib/ foo.py # the main code of the project test/ testlib.py # the testlib.py from *this project* test.py # a small stub driver test_foo.py # an actual test file with TestCase's test_bar.py # another test file
Minimally the "test.py" driver is:
import sys, os import testlib # Put the 'lib' dir on sys.path. sys.path.insert(0, join(dirname(dirname(abspath(__file__))), "lib")) sys.exit(testlib.harness())
Then you run:
cd test python test.py
The test harness will find all TestCase classes in all "test_*.py" files and run them. Tagging support allows you to run subsets of the full test suite:
python test.py foo # Just run tests in "test_foo.py" python test.py bar # ... just in "test_bar.py"
If "test_bar.py" looked something like this:
import unittest from testlib import tag class BlahTestCase(unittest.TestCase): @tag("question") def test_whozit(self): ... @tag("question") def test_whatzit(self): ... def test_thatzit(self): ...
The then following would be possible:
python test.py blah # run all tests in `BlahTestCase` python test.py question # run all tests tagged "question" python test.py -- -question # run all test *except* those tagged "question" python test.py whozit # run just `BlahTestCase.test_whozit`
See "Naming and Tagging" below for more details.
Naming and Tagging
Gathering from Multiple Test Directories