@d3r3kk
Please review, you reviewed the previous PR.
The logic had to change.
One can debug using either Python Executable or the PySpark-submit executable.
To validate the executable, we were using the command python -c print('1').
However that same mechanism doesn't work for pysparm-submit.
I thought using python one.py would work. However when you execute a python file with spark-submit it outputs other stuff as well.

So the new logic is as follows:

• If either one of the following is true, then assume the executable exits:
• Check if python -c print("1") returns the expected output
• Check if the executable file exists

I've created a separate file with targeted tests for spark-submit so hopefully the logic doesn't change without considering the impact on debugging with spark-submit

@d3r3kk Please re-review