- Drop support for Python 2.
- #40: add logging.
- #50, #51: test suite incompatibility with pytest 6.2.0.
- #58: declare the type of automark_dependency ini-option correctly as bool.
- #75: review build tool chain.
- Fix failing documentation build.
- Require pytest version 3.7.0 or newer.
- #34: failing test with pytest 4.2.0 and newer.
- Use setuptools_scm to manage the version number.
- Require pytest version 3.6.0 or newer. This implicitly drops support for Python 2.6 and for Python 3.3 and older.
- #24, #25: get_marker no longer available in pytest 4.0.0. (Thanks to Rogdham!)
- #28: Applying markers directly in parametrize is no longer available in 4.0.
- #5: properly register the dependency marker.
- Do not add the documentation to the source distribution.
- #17: Move the online documentation to Read the Docs.
- Some improvements in the documentation.
- #7: Add a configuration switch to implicitly mark all tests.
- #10: Add an option to ignore unknown dependencies.
Prepend the class name to the default test name for test class methods. This fixes a potential name conflict, see #6.
If your code uses test classes and you reference test methods by their default name, you must add the class name. E.g. if you have something like:
class TestClass(object): @pytest.mark.dependency() def test_a(): pass @pytest.mark.dependency(depends=["test_a"]) def test_b(): pass
you need to change this to:
class TestClass(object): @pytest.mark.dependency() def test_a(): pass @pytest.mark.dependency(depends=["TestClass::test_a"]) def test_b(): pass
If you override the test name in the :func:`pytest.mark.dependency` marker, nothing need to be changed.
- #11: show the name of the skipped test. (Thanks asteriogonzalez!)
- #13: Do not import pytest in setup.py to make it compatible with pipenv.
- #15: tests fail with pytest 3.3.0.
- #8: document incompatibility with parallelization in pytest-xdist.
- Clarify in the documentation that Python 3.1 is not officially supported because pytest 2.8 does not support it. There is no known issue with Python 3.1 though.
Initial release as an independent Python module.
This code was first developed as part of a larger package, python-icat, at Helmholtz-Zentrum Berlin für Materialien und Energie.