Fix #74

[Moomboh]

Fixes #74

[Moomboh]

@uranusjr After digging into this again I went the extra mile and also did the refactoring you mentioned here: #55 (comment). This means it fixes #55 and should also fix #35. Additionally, since I wanted to make the ps fallback also tty independent, I thought I might just as well change the ps call to berkeley style for better compatibility and try to address #21 with this as well.
I could not verify if this actually also fixes #21 and #35 though.

[uranusjr]

Let’s not be greedy, and only fix one issue at a time instead.

[Moomboh]

I reverted the changes to the ps call and only made it fit the refactor. This PR now only focuses on fixing #74 and #55 and should not change other behaviour.
@uranusjr Thank you very much for your feedback, so far!

[uranusjr]

I pushed a commit to refactor the get_process_parents into generator functions and slightly reworded a comment. Can you take a look to ensure I haven’t broken anything at your end?

[Moomboh]

Looks good to me. Still correctly detects shell inside docker container on running on M1 MacBook Pro/osx-arm64 both with rosetta and qemu amd64 emulation.