Change empty leaf of the nullifier tree

[LHerskind]

Implement the successor merkle tree for nullifiers in the base with batch insertion. There are double-spend issues if using batch insertion as in normal merkle tree which need to be addressed in this separately.

The double-spend case is when the same predecessor is to be updated multiple times, such that one of the pending should have a predecessor in the pending set and not outside. Also need to address the 0 leaf insertions and not use empty leafs as base as it is possible to point an empty leaf to anything creating a "split" in the tree.

[iAmMichaelConnor]

Closing because moved to 'Done'