Run independent nose test suites together as one. This is particulary useful, if you are having suites with different package requirements.
In contrast to other Python tools it is not reasonable to put multisuite in virtualenvs, because its task is to manage virtualenvs for your tests. You can still install it into your system with pip, though:
$ sudo pip install multisuite
If that is not what you want, you can simply download the
file from the repository. It is one file
and currently contains everything that is needed to run multisuite.
Let's say you have two test suites with a set of requirements each. One contains MONK in version 0.1.1 and the other in version 0.1.2. Therefore it is not possible to run both test suites in the same virtualenv. Manually creating different virtualenvs for different test suites can become quite complex if you have more than two suites like these. Therefore you decide to use multisuite.
Your suites look like this:
root/ suite_1/ __init__.py requirements.txt # contains monk_tf==0.1.1 suite.py suite_2/ __init__.py requirements.txt # contains monk_tf==0.1.2 suite.py
The order is important here. Each suite can contain as many files as you want,
but it needs to contain at least these three files. The
file contains the requirements that this suite needs and the
contains the test case (or the reference to them). And it needs to contain an
__init__.py file because that's how
nosetests will recognize them.
If you want to make sure, that you have a correct suite, you can create them with multisuite as well:
$ multisuite makesuite suite_1 suite_2
You can check if your test suites can be found by multisuite:
$ multisuite list suite_1 suite_2
If a suite is not listed here it was not detected correctly. Please check again if all requirements are met. Now you can run a test suite by itself:
$ multisuite test suite_1 suite_2 ... suite suite_1 ok suite suite_2 ok
And you can simply run all tests together:
$ multisuite ... suite suite_1 ok suite suite_2 ok
You do not need to spell the prefix
suite_explicitly. Both the following commands are equal:$ multisuite makesuite suite_bugfixes suite_unittests $ multisuite makesuite bugfixes unittests
When developing on new test cases the
shellcommand might come in handy, it takes the user to the folder of a test suite, initializes the virtualenv of that suite, and then starts a python shell. Therefore the environment in the python shell is the same as for the test cases of that suite. For more details:$ multisuite shell -h