Recovering from intermittent crashes

[amiller]

After crashing and restarting, can we resume the protocol in progress?
The easiest way would be to get a threshold signature on a most recent round, and just start there.
Getting an old signature could lead to equivocation, like sending messages in a round we've already participated in. This would be tolerated by the other nodes and not cause a failure, as long as not too many occurred at once.

[sbellem]

@amiller This issue was moved to initc3/HoneyBadgerBFT-Python#33