Correct way to customize environment variables on the remote host

[fernandofloresg]

Hi All! I'm the Team Lead for Ansible Core for z/OS and this is our project https://github.com/ansible-collections/ibm_zos_core we are using pytest-ansible in our project to test our module development and have to say we love it, you've done a great work!

To use our collection, user needs to provide different environment variables (almost 10) and currently we have a hacky way of injecting them, I was not part of this team when the project started so don't know much about the decision making back then.

Last year a change in the way ansible quoted these environment variables made our test suite not compatible with 2.17+, and we had to rush to put a fix. We want to move away from this environment variable injection.

with pytest-ansible these options are now available:

pytest \
    [--inventory <path_to_inventory>] \
    [--extra-inventory <path_to_extra_inventory>] \
    [--host-pattern <host-pattern>] \
    [--connection <plugin>] \
    [--module-path <path_to_modules] \
    [--user <username>] \
    [--become] \
    [--become-user <username>] \
    [--become-method <method>] \
    [--ask-become-pass] \
    [--limit <limit>] \
    [--ansible-unit-inject-only] \
    [--molecule] \
    [--molecule-unavailable-driver] \
    [--skip-no-git-change] \
    [--check]

And we can use ansible_module and ansible_adhoc to test our functions but not sure what is the preferred way to declare environment variables to be used in those fixtures. Any help is appreciated.