Fast sync

[chancehudson]

A client should be able to fast sync by downloading all the merkle tree leaves and then verifying that the resulting hash exists on the blockchain. This would happen in the following steps:

1. Client asks coordinator for leaves
2. Coordinator provides leaves, and the latest block header hash
3. Client downloads leaves, adds them to the database, then calculates root hashes
4. Client verifies that block header exists, then matches nullifier, utxo, and withdrawal roots
5. Client synchronizes blocks as usual, skipping grove patches until it reaches the header hash provided by the coordinator
6. Client continue synchronizes as usual, calculating and applying grove patches

Considerations:

• Coordinator should store a compressed fast sync payload. This payload may take a while to build so it should be done no more than once every hundred blocks (for example)

[chancehudson]

Closing in favor of #358 #357 and #348