Revoked validators should be put into the unbonding state

[rigelrozanski]

When a validator is revoked it should begin unbonding, and eventually enter into the fully unbonded state - Should verify that this is explained in the spec and implemented properly in the codebase

[rigelrozanski]

We know this is not happening, so it needs to be programmed to meet this specification - also we should ensure this is documented in the spec

[rigelrozanski]

Related #1877

ALSO
As a part of this issue we need to ensure that the delegations that are either redelegating or undelegating from a validator in an unbonding state have a reduced unbonding period.

We will need to introduce unbonding_height and unbonding_time into the validator object for the above

[rigelrozanski]

related: #1879

because these are "undelegations" not always "unbonds"

[rigelrozanski]

This will likely have to change.
https://github.com/cosmos/cosmos-sdk/blob/develop/x/stake/keeper/validator.go#L227

[cwgoes]

Also when a validator is unbonded due to being kicked out (which can happen without being revoked).

[cwgoes]

Note that this means creating "unbonding delegation" objects, which we don't do presently.

edit: We just need the unbonding validator objects.

[rigelrozanski]

@cwgoes what do you mean? When a validator is unbonding - the delegator's unbonding tokens are lumped in with this new " unbonding valdiator" object - it's only if the delegators call "undelegate" (from a validator in any state) which an "unbonding delegation" can be created

[cwgoes]

@cwgoes what do you mean? When a validator is unbonding - the delegator's unbonding tokens are lumped in with this new " unbonding valdiator" object - it's only if the delegators call "undelegate" (from a validator in any state) which an "unbonding delegation" can be created

You're right, I stand corrected, that was just a note from simulation that the tokens were vanishing into nonexistence when a validator was forcibly unbonded. "Unbonding validator" objects make more sense.