consortium-v2: new RLP encoder/decoder for HeaderExtraData

[andicrypt]

Currently, HeaderExtraData is encoded/decoded manually, which is error prone and hard to extend. Hence, this PR implements a new RLP encoder/decoder for HeaderExtraData and changes to use this new method.

[andicrypt]

Benchmark the length of encoded data (in bytes) in setting of 22 validators. The data is evaluated in term of the length of bytes.
Note: binary represents the presence of fields in mock extra data such that 0 (vacant), 1 (filled) from the right to left:
1st bit: (HasFinalityVotes, FinalityVotedValidators, AggregatedFinalityVotes)
2nd bit: CheckpointValidators
3rd bit: Seal

    consortium_test.go:650: binary: 000, original: 98; new: 69
    consortium_test.go:650: binary: 001, original: 202; new: 167
    consortium_test.go:650: binary: 010, original: 1594; new: 1657
    consortium_test.go:650: binary: 011, original: 1698; new: 1754
    consortium_test.go:650: binary: 100, original: 98; new: 69
    consortium_test.go:650: binary: 101, original: 202; new: 167
    consortium_test.go:650: binary: 110, original: 1594; new: 1657

[andicrypt]

Note: the smaller is better.
Benchmark the encoding performance, take performance average of 100k times of executions:

BenchmarkEncodeRLP-10             100000              6821 ns/op            6488 B/op         28 allocs/op
BenchmarkEncode-10                100000              6013 ns/op            7120 B/op         31 allocs/op

Benchmark the decoding performance, take performance average of 100k times of executions:

BenchmarkDecodeRLP-10             100000             81366 ns/op            8965 B/op         86 allocs/op
BenchmarkDecode-10                100000             76732 ns/op            3544 B/op         48 allocs/op