Enum encoding is different from the SNIP-12 spec

[thiagodeev]

Hi. I want to talk about this line of code:

starknet.js/src/utils/typedData.ts

Line 375 in 01171b0

revisionConfiguration[revision].hashMethod([variantIndex, ...encodedSubtypes]),

The code only hashes the enum index + the enum parameters but the SNIP-12 says the enum hash is:
Enc[enum] = hash_array(type_hash(enum), variant_index, Enc[chosen_variant_parameter1],..., Enc[chosen_variant_parameterN])
(ref: https://github.com/starknet-io/SNIPs/blob/main/SNIPS/snip-12.md#encoding-4).

AFAIU, it should also hashes the 'type hash' with the index and parameters.

Look at these prints taken when debugging while starknet.js encodes the enum_example:

Here we are encoding EnumA, and the data is [2, [0, 1]]. The code encodes the data to a new array containing the encoded data and passes it with the variantIndex variable (value 1).

Inside the hash function, we can see that only these three values will be hashed, with no mention of the type hash value quoted by the spec. The returned value is '0x583136381661e3eb69d103f2d6f40e6aa190005c3d5c8f0351b080c69602dcb' to the EnumA type.

This is the result of EnumB (value: '0x1fb7169b936dd880cb7ebc50e932a495a60e0084cdab94a681040cb4006e1a0'):

At the end, this will be the result of calling the getStructHash function to calculate the hash of the Example type.

• '0x8eb4aeac64b707f3e843284c4258df6df1f0f7fd38dcffdd8a153a495cd351' is the type hash of Example
• '0x583136381661e3eb69d103f2d6f40e6aa190005c3d5c8f0351b080c69602dcb' is the hash of EnumA
• '0x1fb7169b936dd880cb7ebc50e932a495a60e0084cdab94a681040cb4006e1a0' is the hash of EnumB.

This is correct for the Example type, since the spec says that this is the default for object types (Enc[x] = hash_array(type_hash(MyObject), Enc[param1], Enc[param2], ..., Enc[paramN])), but the values of the EnumA and EnumB encoding are theoretically wrong since they were calculated without the type hash value. So, it's a wrong result after all.

What do you think guys? Please let me know if I misunderstood something.

[penovicp]

@thiagodeev Good catch, thank you for reporting it!

I opened PR #1281 that should address the issue, I would appreciate it if you could take a look.

[thiagodeev]

Thanks!
LGTM. I compared your test result with the one made on starknet.go and they are the same

[github-actions]

🎉 This issue has been resolved in version 6.21.0 🎉

The release is available on:

• npm package (@next dist-tag)
• GitHub release

Your semantic-release bot 📦🚀