Avoid cloning of in-memory structs on apply_block

[bobbinth]

Currently, we clone in-memory data structures before we apply changes to them (e.g., see here). This is done to make sure that the received changes result in a valid update before applying them.

While this works, it will probably get pretty expensive once we have more data in the chain. A better approach would be to compute the updates without cloning the data, then validate the changes, and then apply the updates. To enable this, 0xPolygonMiden/crypto#222 will need to be implemented first.

[bobbinth]

Once #528 is merged, we should be able to address this issue by using the new compute_mutations() and apply_mutations() method of the Sparse Merkle trees.

[polydez]

Resolved by #532