How to handle more complex commands in RF?

[macpijan]

Sometimes there is no single program we can run and inspect it's output.
We need to post-process the ouput in a ways beyond capabilities of RF, or create more complex scenarios as shell scripts, which becomes unreadable in RF.

One example of such is in this thread: #555 (comment)

How do we want to handle it in the future?

The discsussed options include:

• leave it like this,
• store shell scripts in this repo, and send them to DUT prior execution?
• implement more complex kwds in Python instrad?
• ???

[pietrushnic]

Someone could inspect other RF public codes (OpenBMC?) or look for (reach out for) RF community advice.