Improve timing of submitting current position to data stream in set_current_position() of Thorlabs cube motor

[ivalaginja]

Ok this one is tricky. If we do not wait for completion in the set_current_position(), it intercepts the first command and executes the second command instead, and correctly. So I am fine with this behavior and don't want to add the wait_for_completion() there.

However, the method get_current_position() gets called too early and puts an intermediate position on the current_position data stream, which the device reads while it is moving. Which is obviously not at all what we want.

We then tried adding the wait_for_completion() anyway, to fix the wrong values on the data stream, but in this case it just finishes running the first command. Data stream value is still wrong. And it puts command 2 in a cache apparently and runs it before the next command that is sent to the motor... also not the behavior we want, so I still vote against adding the wait_for_completion() in set_current_position().

Since this was so weird, we checked the data stream values after calling home() a couple of times and this is always fine, so it's something about the command data stream monitoring and getting the real device position before the device stops moving.

I don't actually understand what wait_for_completion() does. @a-sevin, maybe you can help out with this conundrum? What does it do and what are the values for message_id and message_type it checks against?

Originally posted by @ivalaginja in #200 (comment)

[ivalaginja]

@a-sevin even if I assigned you, this is not a high-priority issue at this time so don't sweat it if you can't look into this anytime soon.