Expose BLS public keys from platform.getValidatorsAt

[StephenButtolph]

Why this should be merged

This is the last piece to be able to implement an adapter for the validators.State interface from the platformvm.Client interface.

Resolves #1634

Split out of #1611

How this works

Changes the platform.getValidatorsAt API return type to include BLS public keys. Note - this is a breaking change.

validators.State interface can be implemented by:

• GetMinimumHeight -> GetHeight
• GetCurrentHeight -> GetHeight
• GetSubnetID -> ValidatedBy
• GetValidatorSet -> GetValidatorsAt

How this was tested

Calling:

curl --location 'http://localhost:9650/ext/bc/P' \
--header 'Content-Type: application/json' \
--data '{
    "jsonrpc":"2.0",
    "id"     :1,
    "method" :"platform.getValidatorsAt",
    "params" :{
        "height": 109466,
        "subnetID": "11111111111111111111111111111111LpoYY"
    }
}'

against a synced Fuji node.

Previous result
New result

[darioush]

should we add a test that un-marshaling the marshaled json returns the original object?
(Also it seems the weight is returned as a string in the new format vs a number in the old format, not sure if this change is intentional)

[StephenButtolph]

Also it seems the weight is returned as a string in the new format vs a number in the old format, not sure if this change is intentional

We typically return weights as strings because javascript represents all numbers as floats. Which can cause precision errors with the sizes of the numbers we are returning. So we explicitly return strings so that javascript can parse them into big nums correctly.

The prior implementation was an oversight.

[StephenButtolph]

should we add a test that un-marshaling the marshaled json returns the original object?

Added

[cam-schultz]

One sanity check question, but LGTM

[cam-schultz]

Do we need to add MarshalJSON to validators.Validator like the comment describes?

[StephenButtolph]

We do not. I decided to do it this way because it felt weird not to include the NodeID in the struct json definition... But it was duplicate information in the map (which I didn't want to expose over the json rpc)