In the spirit of the library's :doc:`design_goals`, the aim is to have thorough tests for the entire library. This page explains the testing philosophy and shows how to run the tests locally.
Our testing goals are to:
- Write one set of tests that all provider implementations must pass.
- Make that set of tests a 'conformance' test suite, which validates that each implementation correctly implements the CloudBridge specification.
- Make the test suite comprehensive enough that a provider which passes all the tests can be used safely by an application with no additional testing. In other words, the CloudBridge specification and accompanying test suite must be comprehensive enough that no provider specific workarounds, code or testing is required.
- For development, mock providers may be used to speed up the feedback cycle, but providers must also pass the full suite of tests when run against actual cloud infrastructure to ensure that we are not testing against an idealised or imagined environment.
- Aim for 100% code coverage.
- To run the test suite locally:
- Install tox with
pip install tox
- Export all environment variables listed in
- Install tox with
This will run all the tests for all the environments defined in file
Specific environment and infrastructure
If you’d like to run the tests on a specific environment only, say Python 2.7,
against a specific infrastructure, say aws, use a command like this:
tox -e py27-aws. The available provider names are listed in the
ProviderList class (e.g.,
Specific test cases
You can run a specific test case, as follows:
tox -- tests/test_image_service.py:CloudImageServiceTestCase.test_create_and_list_imag
It can also be restricted to a particular environment as follows:
tox -e "py27-aws" -- tests/test_cloud_factory.py:CloudFactoryTestCase
See nosetest documentation for other parameters that can be passed in.
Using unittest directly
You can also run the tests against your active virtual environment directly
python setup.py test. You will need to set the
environment variable prior to running the tests, or they will default to
You can also run a specific test case, as follows:
python setup.py test -s tests.test_cloud_factory.CloudFactoryTestCase
Using a mock provider
Note that running the tests may create various cloud resources, for which you
may incur costs. For the AWS cloud, there is also a mock provider (moto) that
will simulate AWS resources. You can use
CB_TEST_PROVIDER=mock to run tests
against the mock provider only, which will provide faster feedback times.
Alternatively you can run the mock tests through tox.
tox -e "py27-mock"