Code review comment for ~iconstantin/qa-regression-testing:master

Hey Alex,

I like the concept and have been playing around with it. From my understanding, the test-python-x.y.py scripts act as set-up/staging scripts that do two things:

- Specify QRT-Depends, QRT-Packages, and the version of Python being tested:

    # QRT-Depends: test-python.py
    # QRT-Packages: python3.10 netcat-openbsd w3m
    .
    .
    .
    PYTHON='python3.10'

- Runs 'test-python.py' (providing PYTHON as an arg) which is the script containing all the qa unit tests:

    subprocess.Popen([ 'python2', 'test-python.py', PYTHON]

When I want to get these over on a VM to test, I use:

    $ vm-qrt -p sec-impish-amd64 test-python3.10.py

At one point, vm-qrt runs:

    $ grep 'QRT-Packages' ./`basename $test_script`
    $ sudo apt-get install -y --force-yes <QRT-Packages>

This looks at the hard coded line in the source of test-python-x.y.py which specifies the Python version (and other package dependencies). This is where the implementation of an argv/single source script fails (for the current process) since the version would not be part of the source code when vm-qrt runs.