Emit justification info

[cam-schultz]

Why this should be merged

P-Chain validators attest that a validator is inactive and will never become active by signing a L1ValidatorRegistrationMessage with the registered field set to false. Since such validators are not held in the P-Chain's active state, P-Chain validators require an ACP-118 justification be provided along with the message to sign. At present, it's difficult to construct these justifications on-demand, since that information is not exposes on-chain or emitted by any event. This effectively requires validator owners to independently track this information.

This will require a PR to update the ACP-99 spec as well.

How this works

Updates InitiatedValidatorRegistration to emit the serialized bytes of the registerL1ValidatorMessage, and updates RegisteredInitialValidator to emit the subnet ID and validator's index in the ConvertSubnet message. These events can be queried by the indexed validationID to easily recover the information necessary to construct the justification for each case.

How this was tested

CI

How is this documented

NatSpec

[geoff-vball]

Does indexing the subnetID make querying events easier? (not a rhetorical question, I'm actually curious).

Let's say we have two different ValidatorManager contracts on the C-chain that manage two different subnets. Is it easier to query on an indexed field than it is on the address that emitted the event?

[cam-schultz]

I'm not sure. I added this here because there was an unused indexed slot, and because searching by subnetID seems like a common workflow.

[geoff-vball]

What's the eventual purpose of the index field here?

[cam-schultz]

It's used along with the subnetID to create the preimage of the validationID for initial validators. This is the justification for P-Chain validators to sign the L1ValidatorRegistrationMessage attesting that the validator has been removed. See here: https://github.com/ava-labs/avalanchego/blob/master/vms/platformvm/network/warp.go#L153