Add hypothesis based tests to test Clique RLP types

[cburgdorf]

What is wrong?

The following test would benefit from some hypothesis based tests to catch more edge cases such as empty data etc.

py-evm/tests/core/consensus/test_clique_encoding.py

Lines 56 to 72 in 3c3fd7e

	@pytest.mark.parametrize(
	'val, meant_to_be, encode_fn, decode_fn',
	(
	(ADDRESS_TALLY_PAIRS[0], tuple, encode_address_tally_pair, decode_address_tally_pair),
	(DUMMY_VOTE_1, Vote, encode_vote, decode_vote),
	(DUMMY_VOTE_2, Vote, encode_vote, decode_vote),
	(SNAPSHOT_1, Snapshot, encode_snapshot, decode_snapshot),
	(TRUMP_TALLY, Tally, encode_tally, decode_tally),
	(YANG_TALLY, Tally, encode_tally, decode_tally)
	)
	)
	def test_encoding_decoding(val, meant_to_be, encode_fn, decode_fn):
	assert type(val) is meant_to_be
	binary = encode_fn(val)
	assert type(binary) is bytes
	revived = decode_fn(binary)
	assert revived == val

How can it be fixed

Add some hypothesis based test cases