Skip to content
Branch: master
Find file Copy path
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
205 lines (148 sloc) 6.82 KB


Thank you for contributing to OpenShift Ansible. This document explains how the repository is organized, and how to submit contributions.

Table of Contents


Before submitting code changes, get familiarized with these documents:

Please consider opening an issue or discussing on an existing one if you are planning to work on something larger, to make sure your time investment is something that can be merged to the repository.

Submitting contributions

  1. Fork this repository and create a work branch in your fork.
  2. Go through the documents mentioned in the introduction.
  3. Make changes and commit. You may want to review your changes and run tests before pushing your branch.
  4. Open a Pull Request. Give it a meaningful title explaining the changes you are proposing, and then add further details in the description.

One of the repository maintainers will then review the PR and trigger tests, and possibly start a discussion that goes on until the PR is ready to be merged. This process is further explained in the Pull Request process document.

If you get no timely feedback from a project contributor / maintainer, sorry for the delay. You can help us speed up triaging, reviewing and eventually merging contributions by requesting a review or tagging in a comment someone who has worked on the files you're proposing changes to.

Note: during the review process, you may add new commits to address review comments or change existing commits. However, before getting your PR merged, please squash commits to a minimum set of meaningful commits.

If you've broken your work up into a set of sequential changes and each commit pass the tests on their own then that's fine. If you've got commits fixing typos or other problems introduced by previous commits in the same PR, then those should be squashed before merging.

If you are new to Git, these links might help:

Running tests and other verification tasks

We use tox to manage virtualenvs where tests and other verification tasks are run. We use pytest as our test runner.

Alternatively to tox, one can use detox for running verification tasks in parallel. Note that while detox may be useful in development to make use of multiple cores, it can be buggy at times and produce flakes, thus we do not use it in our CI jobs.

pip install tox

To run all tests and verification tasks:


Note: before running tox or detox, ensure that the only virtualenvs within the repository root are the ones managed by tox, those in a .tox subdirectory.

Use this command to list paths that are likely part of a virtualenv not managed by tox:

$ find . -path '*/bin/python' | grep -vF .tox

The reason for this recommendation is that extraneous virtualenvs cause tools such as pylint to take a very long time going through files that are part of the virtualenv, and test discovery to go through lots of irrelevant files and potentially fail.

Running only specific tasks

The tox configuration describes environments based on either Python 2 or Python 3. Each environment is associated with a command that is executed in the context of a virtualenv, with a specific version of Python, installed dependencies, environment variables and so on. To list the environments available:

tox -l

To run the command of a particular environment, e.g., flake8 on Python 2.7:

tox -e py27-flake8

To run the command of a particular environment in a clean virtualenv, e.g., pylint on Python 3.5:

tox -re py35-pylint

The -r flag recreates existing environments, useful to force dependencies to be reinstalled.



Here are some useful tips that might improve your workflow while working on this repository.

Activating a virtualenv managed by tox

If you want to enter a virtualenv created by tox to do additional debugging, you can activate it just like any other virtualenv (py27-flake8 environment in this example):

source .tox/py27-flake8/bin/activate

Limiting the unit tests that are run

During development, it might be useful to constantly run just a single test file or test method, or to pass custom arguments to pytest:

tox -e py27-unit -- path/to/test/

Anything after -- is passed directly to pytest. To learn more about what other flags you can use, try:

tox -e py27-unit -- -h

As a practical example, the snippet below shows how to list all tests in a certain file, and then execute only one test of interest:

$ tox -e py27-unit -- roles/lib_openshift/src/test/unit/ --collect-only --no-cov
collected 1 items
<Module 'roles/lib_openshift/src/test/unit/'>
  <UnitTestCase 'OCProjectTest'>
    <TestCaseFunction 'test_adding_a_project'>
$ tox -e py27-unit -- roles/lib_openshift/src/test/unit/ -k test_adding_a_project

Among other things, this can be used for instance to see the coverage levels of individual modules as we work on improving tests.

Finding unused Python code

If you are contributing with Python code, you can use the tool vulture to verify that you are not introducing unused code by accident.

This tool is not used in an automated form in CI nor otherwise because it may produce both false positives and false negatives. Still, it can be helpful to detect dead code that escapes our eyes.

You can’t perform that action at this time.