ivanilves, thank you for writing this up and sharing the code & the video of the end result, it is a fun project!

As a thought experiment I've had a play to see what the state machine logic might look like if we try to replace the bool state variables with a set of explicit finite states, and to push the state transition logic into a pure function. Results are here: https://gist.github.com/fcostin/851c1b4d1e3cb75ba972408151f1...

It is more verbose, but one advantage of structuring it like this (that i've only partially succeeded at) is that state transitions only happen at most once per each call to advance_state. Might make it easier to read through and follow the logic. It also makes the state transition logic easy to unit test, since it is pure-functional instead of being mixed together with sensing and doing actions (unsure how much of a concern or possibility unit testing is with ino files, I've never done any arduino dev myself).

Not sure if this is anything like what SomeoneFromCA was thinking.

Thanks for your code, looks like this can help improving my state machine indeed. I'll try to apply that approach when I get my next slice of free time ;)

Sure, no problems, will give you feedback later, but here is one recommendation: combine group of booleans frequently used together in conditions into a new separate boolean variable. Say, for example, bool isMoving = (isOpening or isClosing).

boolean grouping sounds neat, thanks!