Fix ActiveState.chain maintenance and FFG rewards

[hwwhww]

1. Filled the missing active_state.chain maintenance during processing block
2. voter_indices = block_vote_cache[block_hash]['total_voter_deposits'] -> voter_indices = block_vote_cache[block_hash]['voter_indices']
3. Keep the balance nonnegative now until we have clear rule of forced exit.
4. Fix FFG reward slot iteration: range(last_state_recalc - config['cycle_length'], last_state_recalc) -> range(max(last_state_recalc - config['cycle_length'], 0), last_state_recalc)
5. Fix formula: quadratic_penalty_quotient = int(sqrt(config['sqrt_e_drop_time'] / config['slot_duration'])) -> quadratic_penalty_quotient = (config['sqrt_e_drop_time'] / config['slot_duration']) ** 2

[hwwhww]

The balance B is 3.2E+19, reward_quotient is 1146880.

1. When slot = config['cycle_length'] * 3, then time_since_finality <= 3 * config['cycle_length']:

Slot	rewards_and_penalties	cumulative rewards_and_penalties	B + cumulative rewards_and_penalties
0	-2.7901785714E+13	-2.7901785714E+13	3.1999972098E+19
1	-5.5803571429E+13	-8.3705357143E+13	3.1999916295E+19
2	-8.3705357143E+13	-1.6741071429E+14	3.1999832589E+19
3	-1.1160714286E+14	-2.7901785714E+14	3.1999720982E+19
4	-1.3950892857E+14	-4.1852678571E+14	3.1999581473E+19

2. When slot = config['cycle_length'] * 4 then time_since_finality > 3 * config['cycle_length']:
   • time_since_finality: 256
   • quadratic_penalty_quotient: 17179869184
   • The difference validators[index].balance * time_since_finality // quadratic_penalty_quotient: 4.76837E+11

Slot	rewards_and_penalties	cumulative rewards_and_penalties	B + cumulative rewards_and_penalties
0	-2.8378622872E+13	-2.8378622872E+13	3.1999971621E+19
1	-5.6757245745E+13	-8.5135868617E+13	3.1999914864E+19
2	-8.5135868617E+13	-1.7027173723E+14	3.1999829728E+19
3	-1.1351449149E+14	-2.8378622872E+14	3.1999716214E+19
4	-1.4189311436E+14	-4.2567934309E+14	3.1999574321E+19

[hwwhww]

@djrtwo this one is also ready to review! 🙂

[djrtwo]

lgtm

[djrtwo]

mergginnnggggggg