Js sign typed data comparison testing

[pacrob]

What was wrong?

sign_typed_data has been implemented with the intention of producing similar results as metamask and ethers-js, but the current test cases only compare with a static value (hash or sig)

How was it fixed?

• Built on the existing javascript mnemonic testing, verify messages encode to the same sig where expected. Test cases explain why where they don't match.
• Fixed a bug in encoding where the order of encoding the domain data was determined by the data (a json object, which is unordered) instead of the standard domain types (an array, which is ordered).

Todo:

• Clean up commit history

• Add or update documentation related to these changes

• Add entry to the release notes

Cute Animal Picture

[reedsa]

Nit over the naming, but overall lgtm!

[reedsa]

This name seems redundant since it is located in the integration directory. I don't have context about the ethers tests or what that name was about. Do the tests encompass tests for any js client? Would be curious if there is a name that aligns with what the tests do?

[pacrob]

Ethers is to JS as web3py is to python - https://github.com/ethers-io/ethers.js/tree/e454afb2fa3612c273a09b154c36b35fafab17b1

Prior to this PR, the integration tests there only used a specific set of functions from ethers, so the name was appropriate. Since I've added tests that use both ethers and a metamask library, I figured a more general name was in order.

[pacrob]

Agree on the redundancy - I'll clean that up.

[pacrob]

Okey, shortened to js-scripts

[fselmo]

First pass looks great to me 🔥

I had a question on one of the validations we do that matches ethers.js but not metamask. I'm going to run the tests now and make sure I understand how they are set up on the JS side as my JS is rusty 😅 ... but the rest looks good to me 👍🏼

[fselmo]

lgtm 👍🏼