Handle meridian flipping better

[wtgee]

On PAN001 last night it tracked Wasp-35 for 55/60 images before it stopped. It was about 30° past the meridian, which is how far it is set to go. However, the scheduler should not have selected the target unless it can get through the full 60 images. Furthermore, the image timed out while trying to do a tracking correction whereas it should have seen that the mount status was no longer Tracking and moved to scheduling instead.

The tracking correction failed, which sends the mount to park. If the mount fails in the night when conditions are still good it will try to start up again. Feasibly it could do this without going all the way to park position first although that is probably safest.

So:

• [ ] Figure out why scheduler didn't know target would hit limits past meridian.
• [ ] Figure out what mount status says when stopped because of meridian.
• [x] Decide if error in night with safe conditions should go home and try again or park and try again. (Or not try again). EDITED: #407 handles this behavior better. We still park but skip the housekeeping and sleeping states. See #407 for details.

[wtgee]

I've set PAN001 to record the mount status tonight so can see what it reports when stopped at meridian. I suspect it just reads 'Stopped', which we could easily handle.