Improve rewards withdrawal process

[maurolacy]

The current rewards withdrawal process would break if a validator was slashed or fell out of the validator set.

See

mesh-security/contracts/consumer/virtual-staking/src/contract.rs

Lines 78 to 97 in 8e520df

	/**
	* This is called once per epoch to withdraw all rewards and rebalance the bonded tokens.
	* Note: the current implementation may (repeatedly) fail if any validator was slashed or fell out
	* of the active set.
	*
	* The basic logic for calculating rebalance is:
	* 1. Get all bond requests
	* 2. Sum the total amount
	* 3. If the sum <= max_cap then use collected requests as is
	* 4. If the sum > max_cap,
	* a. calculate multiplier Decimal(max_cap / sum)
	* b. multiply every element of the collected requests in place.
	* 5. Find diff between collected (normalized) requests and last bonding amounts (which go up, which down).
	* 6. Transform diff into unbond and bond requests, sorting so all unbond happen first
	*/
	fn handle_epoch(
	&self,
	mut deps: DepsMut<VirtualStakeCustomQuery>,
	env: Env,
	) -> Result<Response<VirtualStakeCustomMsg>, ContractError> {

And this comment

mesh-security/contracts/consumer/virtual-staking/src/contract.rs

Lines 138 to 150 in 8e520df

	/**
	* This is called every time there's a change of the active validator set.
	*
	*/
	fn handle_valset_update(
	&self,
	deps: DepsMut<VirtualStakeCustomQuery>,
	additions: &[Validator],
	removals: &[Validator],
	tombstones: &[Validator],
	) -> Result<Response<VirtualStakeCustomMsg>, ContractError> {
	// TODO: Store/process removals (and additions) locally, so that they are filtered out from
	// the `bonded` list

Fixing this requiring keeping track of the active validator set on the Consumer (i.e. on the virtual-staking contract).

This is similar to the way the valset is being kept up-to-date in Tgrade. You can take a look at tgrade-valset for inspiration / ideas.

[uint]

I think this was dealt with in #137. @maurolacy Do we close this, or is there more to do?

[maurolacy]

There's more to do (FIXMEs / TODOs in the virtual-staking code / tests), but will be dealt with during full valset state implementation (#118). Closing.