patch stake nominations bug

[4meta5]

What does it do?

CC @artkaseman Before, we were NOT using the snapshot of validator state at the time of validator selection for payout distribution. Instead, we were using the CandidateState at the time of the delayed payment to determine payout distribution. This means nominators that left before the payout distribution were not paid and, more dangerously, new nominators that joined during the delay were paid instead even though they weren't nominators at the time of the validator's block author reward.

Note that this PR also removes pallet-staking as a dependency. We may add this back if we choose to use the Exposure type when we implement slashing, but we don't need it now. I replaced it with ValidatorSnapshot which provides a snapshot of the state of the validator at the time they are selected for the round.

What important points reviewers should know?

Context

• validators are rewarded by a point system according to how many blocks they author in a round
• the issuance for the round is distributed among the validators based on relative point scores
• within each validator, the due issuance is distributed to validators first (commission + due_proportion) and then to nominators second (due_proportion)
  • due_proportion is actors_stake/total_validator_stake * total_due_issuance_for_the_validator

The bug used the current CandidateState to calculate distribution of rewards earned 2 rounds prior. This means nominators that left before the payout distribution were not paid and, more dangerously, new nominators that joined during the delay were paid instead even though they weren't nominators at the time of the validator's block author reward.

The fix snapshots CandidateState at the time the account is chosen for the round so it can weight the earnings distribution based on this snapshot instead of on the current CandidateState (which could easily have changed between the time the block author was recognized and the payout was executed).

Is there something left for follow-up PRs?

What alternative implementations were considered?

Are there relevant PRs or issues in other repositories (Substrate, Polkadot, Frontier, Cumulus)?

What value does it bring to the blockchain users?

Checklist

• Does it require a purge of the network? No
• You bumped the runtime version if there are breaking changes in the runtime ?
• Does it require changes in documentation/tutorials ? No

[JoshOrndorff]

Looks good to me. I like the added test coverage and the simpler ValidatorSnapshot type.