Add BatchMerkleProof + PoPow Interlink vector verification

[SethDusek]

Closes #536

[coveralls]

Pull Request Test Coverage Report for Build 2235024812

• 462 of 484 (95.45%) changed or added relevant lines in 14 files are covered.
• 9 unchanged lines in 2 files lost coverage.
• Overall coverage increased (+0.2%) to 87.894%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
ergo-chain-generation/src/chain_generation.rs	13	14	92.86%
ergo-chain-generation/src/lib.rs	4	5	80.0%
ergo-chain-types/src/extensioncandidate.rs	6	7	85.71%
ergo-merkle-tree/src/merkletree.rs	174	175	99.43%
ergo-merkle-tree/src/batchmerkleproof.rs	88	93	94.62%
ergo-nipopow/src/nipopow_algos.rs	72	78	92.31%
ergo-merkle-tree/src/json.rs	33	40	82.5%

Files with Coverage Reduction	New Missed Lines	%
ergotree-ir/src/serialization/types.rs	4	84.08%
ergotree-ir/src/chain/token.rs	5	71.67%

Totals
Change from base Build 2233555057:	0.2%
Covered Lines:	14905
Relevant Lines:	16958

---

💛 - Coveralls

[greenhat]

Sounds great! Good call! I appreciate you did benchmarks!

[greenhat]

Looking great! Just a few minor comments.

[SethDusek]

@greenhat Hi, do you happen to know why the Github workflow isn't running proptests for BatchMerkleProof? I noticed that they aren't included in the tests. While running proptests, I believe I've found a bug that leads to a stack overflow in both the Rust and Scala implementations of BatchMerkleProof::valid, it should be simple enough to fix however (If a/e are empty, BatchMerkleProof::validate will keep being called, a misbehaving node could send a bad proof for a header and cause a stack oveflow). I've attached an example that replicates the issue in Scala

val batch = BatchMerkleProof(Seq(), Seq((Digest32 @@ Array.fill[Byte](32)(0),MerkleProof.LeftSide)))
batch.valid(Digest32 @@ Array.fill[Byte](32)(0))

[greenhat]

@greenhat Hi, do you happen to know why the Github workflow isn't running proptests for BatchMerkleProof? I noticed that they aren't included in the tests.

Maybe because of #[cfg(feature = "arbitrary")] on mod test it's not included in the build?

While running proptests, I believe I've found a bug that leads to a stack overflow in both the Rust and Scala implementations of BatchMerkleProof::valid, it should be simple enough to fix however (If a/e are empty, BatchMerkleProof::validate will keep being called, a misbehaving node could send a bad proof for a header and cause a stack oveflow). I've attached an example that replicates the issue in Scala

val batch = BatchMerkleProof(Seq(), Seq((Digest32 @@ Array.fill[Byte](32)(0),MerkleProof.LeftSide)))
batch.valid(Digest32 @@ Array.fill[Byte](32)(0))

Great catch! Could you please make an issue in scrypto repo? Thanks.

[SethDusek]

Figured out how to get proptests to run by default (added the feature in ergo-lib dev-dependencies). Here's the scrypto issue input-output-hk/scrypto#96. I've fixed it in the Rust version

[greenhat]

Looking great! Thank you!