Update to latest merkletreejs version will break a unit test in MerkleProof.test.js

[pcaversaccio]

I just wanted to let you know that if you update to the latest merkletreejs version (0.3.2), the following line will break (at least it did for me :-D):

const badProof = badMerkleTree.getHexProof(badElements[0]);

I realised it since I used the same test setup for my Vyper Snekmake foundry tests. You can fix it by adding the index of 2. See my commit here: pcaversaccio/snekmate@0f66c3e:

const badProof = badMerkleTree.getHexProof(badElements[0], 2);
// or the following will produce the same result as well
const badProof = badMerkleTree.getHexProof(badElements[2], 2);

[frangio]

Do you know what changed? We should change our tests to use our own merkle-tree library now.

[pcaversaccio]

Tbh I didn't do a deep-dive since there are no release notes which pisses me somehow off. I think that PR introduced the change: merkletreejs/merkletreejs@8af3c3d, but might be wrong. Tagging @miguelmota for help.

I actually assumed you wanted to switch, therefore I didn't open a PR :).

[pcaversaccio]

For the error logs, you can see a failed action here:

2022-11-14T02:10:50.649702Z ERROR foundry_evm::executor::inspector::cheatcodes::ext: stderr err="/home/runner/work/snekmate/snekmate/node_modules/@ethersproject/logger/lib/index.js:247\n        throw
this.makeError(message, code, params);\n        ^\n\nError: types/values length mismatch (count={\"types\":1,\"values\":2},
value={\"types\":[\"bytes32\"],\"values\":[\"0x\",\"0x\"]}, code=INVALID_ARGUMENT, version=abi/5.7.0)\n    at Logger.makeError
(/home/runner/work/snekmate/snekmate/node_modules/@ethersproject/logger/lib/index.js:238:21)\n    at Logger.throwError
(/home/runner/work/snekmate/snekmate/node_modules/@ethersproject/logger/lib/index.js:247:20)\n    at AbiCoder.encode
(/home/runner/work/snekmate/snekmate/node_modules/@ethersproject/abi/lib/abi-coder.js:83:20)\n    at Object
<anonymous> (/home/runner/work/snekmate/snekmate/test/utils/scripts/generate-bad-proof.js:11:32)\n    at
Module._compile (node:internal/modules/cjs/loader:1155:14)\n    at Object.Module._extensions..js
(node:internal/modules/cjs/loader:1209:10)\n    at Module.load (node:internal/modules/cjs/loader:1033:32)\n    at
Function.Module._load (node:internal/modules/cjs/loader:868:12)\n    at Function.executeUserEntryPoint [as runMain
(node:internal/modules/run_main:81:12)\n    at node:internal/main/run_main_module:22:47 {\n  reason: 'types/values length
mismatch',\n  code: 'INVALID_ARGUMENT',\n  count: { types: 1, values: 2 },\n  value: { types: [ 'bytes32' ], values: [ '0x', '0x' ]
}\n}\n"

[pcaversaccio]

I actually just realised that the latest version 0.3.2 is not even committed to the repo as the latest version bump is to 0.3.1. merkletreejs/merkletreejs@68566c5.

[miguelmota]

Can you provide a reproducible snippet? there hasn't been any functional changes; only addition of new methods and updated return type as you pointed out in commit, and all the lib tests still pass, so I wouldn't expect there to be any breaking changes but I could be missing something

[pcaversaccio]

sure - the following snippet will run with version 0.3.0 but not with 0.3.2:

const { MerkleTree } = require("merkletreejs");
const ethers = require("ethers");

const badElements = ["d", "e", "f"];
const badMerkleTree = new MerkleTree(badElements);

const badProof = badMerkleTree.getHexProof(badElements[0]);
// to make it work with version `0.3.2` I had to add an index value of 2
// const badProof = badMerkleTree.getHexProof(badElements[0], 2);

// eslint-disable-next-line no-undef
process.stdout.write(
  ethers.utils.defaultAbiCoder.encode(["bytes32"], badProof)
);

[miguelmota]

Thanks @pcaversaccio. The latest version should be fixed now. There were some optimization updates recently that I had forgotten went into a release but the tests didn't catch that there was a breaking change so it's been updated now to undo the line that caused the breaking change and additional test has been added.

[miguelmota]

As frangio pointed out, the oz lib @openzeppelin/merkle-tree will supersede merkletreejs since it'll be more well maintained in the longer term

[pcaversaccio]

@miguelmota awesome - just upgraded and can confirm it runs smoothly. Ok, I will upgrade my Merkle proof verification tests to @openzeppelin/merkle-tree once my time permits. Closing the issue now as it's resolved.