Pipeline node properties does not recognize new environment variables

Create a notebook with some environment variables in a cell:

print(os.getenv('USER', default_value))
print(os.getenv('LOGNAME', default_value))

Add node to the pipeline and then go to node properties

Add more environment variables to the same or another cell:

print(os.getenv('USER', default_value))
print(os.getenv('LOGNAME', default_value))
print(os.getenv('CONDA_DEFAULT_ENV', default_value))

Save the notebook, reopen the pipeline and the node properties

No changes, this is because we are reading/parsing the env_vars values during node add.

This should be fixed by implementing #957 and then integrating on the UI.

[lresende]

@kevin-bates you are correct. The implementation does not resonate this today because they were originally created as two distinct extensions without the ability to share code. Now that we managed to merge them as one extension, we should cleanup the implementation.

[lresende]

Also, we should always keep an eye on submitting notebook versus pipelines as they have similar semantics and dialogs but completely different implementations, and this is one behavioral difference.