Tests

This page gives an overview of our testing system.

Continuous Integration

Our continuous integration server can be found at: http://ci.unknown-horizons.org

Manually running tests

Before you can run the tests, the following python modules have to be installed:

• nose (http://readthedocs.org/docs/nose/en/latest/)
• mock (http://www.voidspace.org.uk/python/mock/)
• greenlet (https://pypi.python.org/pypi/greenlet)

We're using nose as test runner, but you have to use the run_tests.py script instead of nosetests that comes with nose. However run_tests.py exposes the same command line interface as nosetests, therefore both commands should behave the same.

python run_tests.py

This command will run all tests in the tests/ directory.

Run a subset of tests

To run only specific tests, you may either pass a directory path or a python package path.

python run_tests.py tests/unittests/                                                      # run every test in this directory (including sub-directories)
python run_tests.py tests.unittests.world                                                 # run every test in this package/module
python run_tests.py tests.unittests.world.test_storage:TestGenericStorage                 # run a specific test case
python run_tests.py tests/unittests/world/test_storage.py:TestGenericStorage.test_alter   # run a specific test in the test case

For more information, check out the documentation on nose.

Hide unnecessary logging output

Sometimes failing tests produce a lot of output to the console which makes relevant information harder to read. In order to hide it, run tests with additional parameters.

Hide logging:

python run_tests.py --nologcapture

Disable printing:

python run_tests.py -s > /dev/null

Running both, e.g.

python run_tests.py --nologcapture -s > /dev/null

Should leave you only with relevant unit test output.

Test types

Unit tests

TODO: Add explanation how to write one

AI tests

AI tests take longer to run, therefore by default they will not run. They are tagged with the attribute 'long', using -a long will include them in the test discovery:

python run_tests.py -a long

TODO: Add explanation how to write one

GUI tests

GUI tests work differently than unit/ai tests. For each test, a new game will be started and the test interacts with the game. This takes some time, therefore they are disabled by default.

python run_tests.py -a gui

How to write: Writing GUI tests