Testing¶
Cloud-init has both unit tests and integration tests. Unit tests can
be found at tests/unittests. Integration tests can be found at
tests/integration_tests. Documentation specifically for integration
tests can be found on the Integration testing page, but
the guidelines specified below apply to both types of tests.
Cloud-init uses pytest to write and run its tests.
Note
While there are a subset of tests written as unittest.TestCase
sub-classes, this is due to historical reasons. Their use is discouraged and
they are tracked to be removed in #6427.
Guidelines¶
The following guidelines should be followed.
Test layout¶
For consistency, unit test files should have a matching name and directory location under
tests/unittests.E.g., the expected test file for code in
cloudinit/path/to/file.pyistests/unittests/path/to/test_file.py.
pytest guidelines¶
Use pytest fixtures to share functionality instead of inheritance.
Use bare
assertstatements, to take advantage ofpytest’s assertion introspection.Prefer
pytest’s parametrized tests over test repetition.
In-house fixtures¶
Before implementing your own fixture do search in */conftest.py files
as it could be already implemented. Another source to look for test helpers is
tests/*/helpers.py.
Relevant fixtures:
disable_subp_usage auto-disables call to subprocesses. See its documentation to disable it.
fake_filesystem makes tests run on a temporary filesystem.
paths provides an instance of cloudinit.helper.Paths pointing to a temporary filesystem.
Dependency versions¶
Cloud-init supports a range of versions for each of its test dependencies, as
well as runtime dependencies. If you are unsure whether a specific feature is
supported for a particular dependency, check the lowest-supported
environment in tox.ini. This can be run using tox -e lowest-supported.
This runs as a GitHub Actions job when a pull request is submitted or updated.
Mocking and assertions¶
Variables/parameter names for
MockorMagicMockinstances should start withm_to clearly distinguish them from non-mock variables. For example,m_readurl(which would be a mock forreadurl).The
assert_*methods that are available onMockandMagicMockobjects should be avoided, as typos in these method names may not raiseAttributeError(and so can cause tests to silently pass).An important exception: if a
Mockis autospecced then misspelled assertion methods will raise anAttributeError, so these assertion methods may be used on autospeccedMockobjects.
For a non-autospecced
Mock, these substitutions can be used (mis assumed to be aMock):m.assert_any_call(*args, **kwargs)=>assert mock.call(*args, **kwargs) in m.call_args_listm.assert_called()=>assert 0 != m.call_countm.assert_called_once()=>assert 1 == m.call_countm.assert_called_once_with(*args, **kwargs)=>assert [mock.call(*args, **kwargs)] == m.call_args_listm.assert_called_with(*args, **kwargs)=>assert mock.call(*args, **kwargs) == m.call_args_list[-1]m.assert_has_calls(call_list, any_order=True)=>for call in call_list: assert call in m.call_args_listm.assert_has_calls(...)andm.assert_has_calls(..., any_order=False)are not easily replicated in a single statement, so their use when appropriate is acceptable.
m.assert_not_called()=>assert 0 == m.call_count
When there are multiple patch calls in a test file for the module it is testing, it may be desirable to capture the shared string prefix for these patch calls in a module-level variable. If used, such variables should be named
M_PATHor, for datasource tests,DS_PATH.
Test argument ordering¶
Test arguments should be ordered as follows:
mock.patcharguments. When used as a decorator,mock.patchpartially applies its generatedMockobject as the first argument, so these arguments must go first.pytest.mark.parametrizearguments, in the order specified to theparametrizedecorator. These arguments are also provided by a decorator, so it’s natural that they sit next to themock.patcharguments.Fixture arguments, alphabetically. These are not provided by a decorator, so they are last, and their order has no defined meaning, so we default to alphabetical.
It follows from this ordering of test arguments (so that we retain the property that arguments left-to-right correspond to decorators bottom-to-top) that test decorators should be ordered as follows:
pytest.mark.parametrizemock.patch
Test pre-release cloud-init¶
After the cloud-init team creates an upstream release, cloud-init will be released in the -proposed APT repository for a period of testing which provides SRU updates to multiple Ubuntu releases. Users are encouraged to test their workloads on pending releases so bugs can be caught and fixed prior to becoming more broadly available via the -updates repository. This guide describes how to test the pre-release package on Ubuntu.
Add the -proposed repository pocket¶
The -proposed repository pocket will contain the cloud-init package to be tested prior to release in the -updates pocket.
echo "deb http://archive.ubuntu.com/ubuntu $(lsb_release -sc)-proposed main" >> /etc/apt/sources.list.d/proposed.list
apt update
Install the pre-release cloud-init package¶
apt install cloud-init
Test the package¶
Whatever workload you use cloud-init for in production is the best one to test. This ensures that you can discover and report any bugs that the cloud-init developers missed during testing before cloud-init gets released more broadly.
If issues are found during testing, please file a new cloud-init bug.
Remove the proposed repository¶
Do this to avoid unintentionally installing other unreleased packages.
rm /etc/apt/sources.list.d/proposed.list
apt update
Remove artifacts and reboot¶
This will cause cloud-init to rerun as if it is a first boot.
sudo cloud-init clean --logs --reboot