[][src]Crate merkletree

light Merkle Tree implementation.

Merkle tree (MT) implemented as a full binary tree allocated as a vec of statically sized hashes to give hashes more locality. MT specialized to the extent of hashing algorithm and hash item. [Hashable] trait is compatible to the std::hash::Hasher and supports custom hash algorithms. Implementation does not depend on any external crypto libraries, and tries to be as performant as possible.

This tree implementation uses encoding scheme as in Certificate Transparency by default. Encoding scheme for leafs and nodes can be overridden though. RFC 6962:

MTH({d(0)}) = ALG(0x00 || d(0)).
For n > 1, let k be the largest power of two smaller than n (i.e.,
k < n <= 2k).  The Merkle tree Hash of an n-element list D[n] is then
defined recursively as
MTH(D[n]) = ALG(0x01 || MTH(D[0:k]) || MTH(D[k:n])),


Implementation choices

Main idea is the whole code must obtain specialization at compile time with minimum allocations calls, hashes must be of fixed size arrays known at compile time, hash algorithm must be a trait and must not depend on any external cryptographic libraries and the lib itself must somehow mimic std Rust api.

Standard way in Rust is to hash objects with a std::hash::Hasher, and mainly that is the reason behind the choice of the abstractions:

Object : Hashable<H> -> Hasher + Algorithm <- Merkle Tree

Custom [merkle::hash::Hashable] trait allows implementations differ from [std::collection] related hashes, different implementations for different hashing algorithms / schemas and conforms object-safety trait rules.

[Algorithm] complements [Hasher] to be reusable and follows the idea that the result hash is a mapping of the data stream.

[Algorithm.hash] had to change its signature to be &mut self (&self) because most of the cryptographic digest algorithms breaks current state on finalization into unusable. ring libra tho contains interfaces incompatible to start-update-finish-reset lifecycle. It requires either cloning() its state on finalization, or Cell-ing via unsafe.

Turning back to having [Algorithm.write(&mut self, &[u8])] instead of write(T) allows to relax [Algorithm] trait [Hasher] constraint, even tho works together well still.


- build_tree (items) -> tree
- get_root -> hash
- gen_proof -> proof
- validate_proof (proof, leaf, root) -> bool


[test_sip.rs]: algorithm implementation example for std sip hasher, u64 hash items [test_xor128.rs]: custom hash example xor128 [test_cmh.rs]: custom merkle hasher implementation example [crypto_bitcoin_mt.rs]: bitcoin merkle tree using crypto lib [crypto_chaincore_mt.rs]: chain core merkle tree using crypto lib [ring_bitcoin_mt.rs]: bitcoin merkle tree using ring lib

Quick start

#[cfg(feature = "chaincore")]
extern crate crypto;
extern crate merkletree;

#[cfg(feature = "chaincore")]
mod example {
    use std::fmt;
    use std::hash::Hasher;
    use crypto::sha3::{Sha3, Sha3Mode};
    use crypto::digest::Digest;
    use merkletree::hash::{Algorithm, Hashable};

    pub struct ExampleAlgorithm(Sha3);

    impl ExampleAlgorithm {
        pub fn new() -> ExampleAlgorithm {

    impl Default for ExampleAlgorithm {
        fn default() -> ExampleAlgorithm {

    impl Hasher for ExampleAlgorithm {
        fn write(&mut self, msg: &[u8]) {

        fn finish(&self) -> u64 {

    impl Algorithm<[u8; 32]> for ExampleAlgorithm {
        fn hash(&mut self) -> [u8; 32] {
            let mut h = [0u8; 32];
            self.0.result(&mut h);

        fn reset(&mut self) {

fn main() {
#[cfg(feature = "chaincore")]
    use example::ExampleAlgorithm;
    use merkletree::merkle::MerkleTree;
    use merkletree::store::VecStore;

    let mut h1 = [0u8; 32];
    let mut h2 = [0u8; 32];
    let mut h3 = [0u8; 32];
    let mut h4 = [0u8; 32];
    h1[0] = 0x11;
    h2[0] = 0x22;
    h3[0] = 0x33;
    h4[0] = 0x44;

    let t: MerkleTree<[u8; 32], ExampleAlgorithm, VecStore<_>> = MerkleTree::try_from_iter(vec![h1, h2, h3, h4].into_iter().map(Ok)).unwrap();
    println!("{:?}", t.root());



Hash infrastructure for items in Merkle tree. Hash infrastructure for items in Merkle Tree.


Merkle tree abstractions, implementation and algorithms.


Merkle tree inclusion proof.


Store implementations.


Re-usable Testing primitives