Redelegation transaction to remove leftover record state-bloat

[rigelrozanski]

Had a good discussion @cwgoes and @jaekwon

Delegators have only weak incentives to complete redelegations due to the fact that their records will not be slashed past the unbonding period - the only incentive delegators have to complete this redelegation is if they wish to further redelegate those records - at which point the old redelegation record must be cleared - which should happen as a part of the same redelegation transaction (ACTION ITEM)

Due to the fact that redelegation records are going to get left behind we need to make sure that, when slashing we iterate through these records by height, such that as soon as the first non-relevant record is fine, the active slashing mechanism for redelegations just stops (ACTION ITEM)

We should actually also probably do this for unbonding delegations too (ACTION ITEM) - just to be consistent - even though it's not anticipated that this will be a problem (I think).

Lastly we should probably remove the complete-redelegation command altogether because it's not anticipated to ever get used (ACTION ITEM)

[ValarDragon]

Per a call with @jaekwon @rigelrozanski @cwgoes @sunnya97 yesterday, we discussed that we want to go ahead and remove CompleteUnbonding and CompleteRedelegation, and have these automatically happen. This will be done by use a of a queue, similar to what is done in governance.

We discussed that this should be done after the staking refactor.

[rigelrozanski]

Is that the conclusion we came too? I didn't feel so certain that it was. but I mean we can do it, np, it's nice, and it superseded all the other items that were brought up originally in this issue - so if those original items were pre-launch, they sound like they might be even more work that just implementing (adapting) the queue - dope

[rigelrozanski]

Closing in favour of #2393 based on these discussions