TST: clean up pytest spam, warnings

[ZLLentz]

Description

• Add shutdown method to ActionNode to clean up the worker on shutdown. py_trees describes this as something like "the opposite of setup" so since we set the worker up in setup, this seems like the canonical place to clean up the worker.
• Tweak ActionNode's logic so it only cleans up the worker once.
• Create a helper for the test suite that can be used to register behavior trees and nodes for cleanup at the end of a unit test. This will effectively stop the workers between tests.
• Apply this helper to all the tests that need it.
• Register a dummy "flaky" mark

Motivation and Context

• Stopping work with the atexit hook is correct and necessary, but causes issues (spam) in pytest due to some sort of cleanup ordering that I wasn't able to pin down.
• In other cases (qtbot) there is precedence for making sure each test cleans itself up.
• The flaky mark is used in caproto and comes from pytest-rerunfailures. Since we don't want to use flaky tests, instead I opted to register the mark so pytest would stop complaining about it existing.

How Has This Been Tested?

Unit tests still work
Unit tests no longer spam on exit

Where Has This Been Documented?

Release notes

Pre-merge checklist

• Code works interactively
• Code follows the style guide
• Code contains descriptive docstrings, including context and API
• New/changed functions and methods are covered in the test suite where possible
• Test suite passes locally
• Test suite passes on GitHub Actions
• Ran docs/pre-release-notes.sh and created a pre-release documentation page
• Pre-release docs include context, functional descriptions, and contributors as appropriate

[tangkong]

Can't complain with the results, this works well for me.

I originally did some messing with caplog in the conftest.central_logging_setup as an attempt to also avoid the spam, but this is a much more robust way to handle things. We could potentially remove that call, but I didn't see any change in behavior when I tried removing it (invoking pytest -s)

[tangkong]

Also I dig the fixture. I wondered if it was possible to register nodes more auto-magically, but I think that's more work than it's worth (and also I'm already used to qtbot so I can't complain)

[joshc-slac]

Very good work. I could imagine this notion of BTCleaner being reused, when this comes up we can remove it from conftest

[ZLLentz]

I wondered if it was possible to register nodes more auto-magically

I certainly tried! But it's not so easy unless we start registering them at init time.