ZIP-2 Create another court voting round after unclear court decision

[Chralt98]

zeitgeist/zrml/court/src/lib.rs

Lines 958 to 962 in 51b2e04

	// TODO this should never happen, we should have another vote round for it
	// TODO: the appeal round should be repeated
	// TODO: right after each aggregation period `on_initialize` should check if we have a clear winner (one outcome with plurality of votes) and at least one revealed vote
	// TODO: if there is no clear winner, the appeal should be repeated (same appeal number)
	return Err(Error::<T>::NoVotes.into());

[Chralt98]

For unclear court cases @lsaether and @maltekliemann are in favour of reverting back to the oracle report outcome.

Two possible scenarios:

1. the highest two or more outcomes have the same amount of votes
2. no juror did correctly reveal the vote (all denounced or did not reveal or did not vote)

I see a possible flaw when we revert back to the oracle report.

Assume we are in the last appeal round. All court appeal rounds came to a clear decision, but not the last appeal.

In this case there was a huge amount of time effort for the court system. Now assume that scenario 1 hits.

Say we have 2^5 * 5 + 2^5 - 1 = 191 jurors randomly selected. The vote distribution is the following:

Outcome A: 95 votes
Outcome B: 95 votes
Outcome C: 1 vote
Oracle Report Outcome: 0 votes

The court clearly decided not for the oracle report outcome. But if we revert to this, the court was a time waste, even if previous appeal rounds had clear winners.

Suggestions:

For 2.) I would revert back on the oracle report. For 1.) I would either start a new juror voting round or trigger a global dispute. Keep in mind that a global dispute is storage expensive. So I lean towards a new voting round.

[Chralt98]

UPDATE: for 1.) we resolve now on the last appeal result if present. If the last appeal is not present we resolve on the oracle report.

Everyone has still the chance to appeal again if we end up with an unclear decision like pointed out in 1.)

[Chralt98]

Fixed by #976