Make "unusable" IO port modes useful & add cell movement queue

[S4mpsa]

Has the following changes:

• IO Port on "Move to output when empty" and "Pull from network" modes will now move the cell to the output slot if the network was emptied and work was done, essentially checking if the network was emptied by the IO port.
• IO Port on "Move to output when full" and "Push to network" modes will now move the cell to the output slot if the network was filled and work was done, checking if the network was filled by the IO port.

These two new behaviors allow using the IO port in new ways more reliably.

Another change contained in the PR is a "move queue", where a cell that has finished its work, but couldn't be moved to the output slots, will no longer do work within the IO port and will try to move every time the IO port is updated. This change is required for the above two modes to work, but doesn't influence the behavior of any existing used modes.

[ivelieu]

This will break IO port automation that specifically relies on this functionality (notably, some of my vacuum nuke approaches which are used by a fair few people). Better to add this behavior as a fourth mode.

[Caedis]

drafted per request of Sampsa

[AnrDaemon]

Ain't there are already all modes necessary? What 4'th mode?
Move when done, move when empty, move when full.
What else do you need, and why?

[ivelieu]

Move when done some work (wait for work to start then stop), move when transfer finished (wait for work to stop), move when empty and move when full.

You may also read the relevant part of the GTNH official discord tips and trick thread "Universal Automation for any processing machine*" where this PR is linked. There sampsa explained instead changing this only for the modes "Move when cell full" when the IO port is on "Push to network" and "Move when cell empty" when IO port is on "Pull to cell" which are otherwise useless modes. I am indifferent both implementations seem reasonable.

[github-actions]

Warning: 2 uncommitted changes
#632