Test cases are executed within a virtual environment in order to have all python package dependencies installed. This virtual environment creation is managed by tox. Furthermore tox execution can be optionally wrapped with a docker environment to provide required operating system packages and configuration to pull dependencies.

Execution environment as simple diagram:

$ ./rfcli-docker <args> == [ docker [ tox [ virtualenv [ rfcli -> robotframework <args> ] ] ] ]

You can find here more details about how to execute in your machine here (windows specific, but it contains tips also if you have linux): windowsinstructions

Please use the following tags mentioned here: taggingpolicy and also consider our test case design guidelines: designguidelines

Docker + tox:

$ ./rfcli-docker -t path/to/target.ini -s smoke-tests testcases/

This is the most simple way to execute the cloudtaf2 tests. As docker is being used the created python virtual environment as well as the resulting files will be owned by the "root" and not the actual user. Additionally initial execution will take some time to build the image.

You may also run directly rfcli environment with tox:

$ tox -e rfcli -- -t path/to/target.ini -s smoke-tests testcases

This is more lightweight way to execute the tests but requires certain operating system packages to be installed first. Please see the "Dockerfile" for installed packages.

Tox only - more lightweight for those who know what they are doing:

$ tox -e rfcli -- -t path/to/target.ini -s smoke-tests testcases/

Python virtual environment is created with requirements.txt. All python packages must be set to certain versions in order to execute tests the same way with older builds as they have been initially executed.

To update the frozen versions to the latest, execute:

$ tox -e freeze

To add more packages, update the requirements-minimal.txt and execute:

$ tox -e freeze

You can add specific version requirements to the requirements-minimal.txt to make sure that they are used in the generated frozen requirements.txt.

The recommendation is that the frozen requirements.txt file is not edited directly but always via the requirements-minimal.txt changes and via this freeze generation.

In clear cases you can also update requirements.txt directly. However, in this case, please make sure that:

$ tox --recreate -e check-requirements

is succesful. This tool checks that the requirements-minimal.txt is consistent with the requirements.txt and that all requirements in requirements.txt can be really installed. installed.

Rebuild docker image manually e.g. when changing the Dockerfile contents:

$ ./rfcli-docker-build

Unit tests can be executed with:

$ tox

The 'dnf install' -command requires a proxy setting when Docker is running behind a proxy.

The Dockerfile writes the proxy information to /etc/dnf/dnf.conf -file when HTTP_PROXY argument is set as a --build-arg. For example:

# docker build --build-arg HTTP_PROXY=http://10.1.2.3:8080/