sync_send_to causes deadlock when sending to calling module

[sjrct]

Ideally a self-send should function without causing a deadlock. Minimally, sync_send_to should throw an exception when there is a potential deadlock.

[richteer]

This deadlock problem is making me reconsider our concurrency model. I think it might be worth revisiting since the introduction of sync_send_to -- after doing some brief research into coroutines and such, there are cases where deadlocks can be formed due to improper use/waiting on futures.

Expect a crazy RFC PR in the future that might overhaul how message processing/handling works (again). I'm leaning towards a single event loop and proper use of await/yield/etc.

Side note: I'm merging the temporary fix into master, but leaving this issue open since we have a bigger fish to fry here.