Adding support of EIP-712 into Ethereum app #1568

xlab · 2021-04-13T21:59:10Z

Intro

Hello 👋! I'm Max from Injective Protocol. We're building a fully decentralized trading platform for derivatives. It's an L2 chain based on Cosmos SDK, and we rely heavily on EIP-712 typed data standard for interoperation with Ethereum and MetaMask. Basically, this allows us to sign arbitrary Cosmos transactions using Eth wallets like MetaMask or Ledger.

I know there are a lot of people trading/minting NFTs nowadays, many of which are minted through L2 solutions, almost all the time depending on EIP-712 as well. There are plenty of issues about EIP-712 support in hardware wallets like this one MetaMask/metamask-extension#10240

We have users with Trezor wallets too, so I decided to step in and help with the EIP-712 feature implementation for the firmware. My Pythonic code is based upon signTypedData_v4 reference implementation from MetaMask (see the TypeScript eth-sig-util code also signTypedData_v4 docs) since we need to maintain compatibility with current Web3 industry standard. It's also a fact that their *_v4 implementation has a bug that makes it not compliant to the official EIP-712 standard - see MetaMask/eth-sig-util#106, so one day it could be *_v5.

Meanwhile, we have to use Metamask's v4 everywhere and allow users to fall back to Geth's (which I'm not sure is used anywhere). There is a special flag --use-v4 in the reference client implementation that can be set to False.

$ trezorctl eth sign-typed-data --help
Usage: trezorctl eth sign-typed-data [OPTIONS] TYPED_DATA_JSON

  Sign typed data (EIP-712) with Ethereum address.

Options:
  -n, --address TEXT  BIP-32 path, e.g. m/44'/60'/0'/0/0  [required]
  --use-v4 BOOLEAN    Specify to use EIP-712 (v4) for typed data signing
  --help              Show this message and exit.

In this PR I'm submitting a working implementation of v4 typed data signatures, also a nice UI that in my opinion has the best balance between verbosity of the layout and the ability to introspect data on the device. This is not a "WIP" request, in my opinion, it's ready for being field-tested.

Click for the video👇

Data exchange flow

The exchange between client and device is defined by proto schema and consists of 3 messages:

EthereumSignTypedData - initial request for starting a signing flow
EthereumTypedDataRequest - request from device to send additional data
EthereumTypedDataAck - the fulfilled request with additional data types and values

In the final EthereumTypedDataRequest device tells signature of the data and address of the signer.

The general idea is that the device gets all the types from EIP712 struct first, and uses them as a stencil to get the rest of the data values. In that way, deeply nested structures are still a possibility, but the device cannot be "led by the nose" by the client. The device always knows the full layout before reading any values.

The flow on device can be summarized like this:

    domain_types = await collect_domain_types()
    domain_values = await collect_values("EIP712Domain", domain_types)
    primary_type, message_types = await collect_types()
    message_values = await collect_values(primary_type, message_types)

    domain_separator = hash_struct("EIP712Domain", domain_values, domain_types, use_v4)
    message_hash = hash_struct(primary_type, message_values, message_types, use_v4)

    return keccak256(b"\x19" + b"\x01" + domain_separator + message_hash)

Where collect_domain_types and collect_types are collecting the data schema, and collect_values is collecting the data values, communicating with the client. The rest is a 100% computational workload.

Ethereum ABI

Another important detail for working with EIP-712 typed data, and Ethereum in general, is the data encoding format called Eth ABI. EIP712 tells that there is a subset of types supported in the data definitions, but the encoding of the types really matches the standard ABI. Basically, ABI encoding allows us to represent any primitive data (no structs) as bytes, even an array of primitive types, like []uint256, so in the data flow described above the client transfers data on the device already ABI-encoded.

For the reference implementation in Python, I've picked eth-abi as it is the most sophisticated implementation out there and works nicely with native Python types already. I'm also using its eth_abi.packed variant since EIP-712 doesn't expect all those offset headers in the encoded values.

For the device-side ABI encoding and decoding, which is important (since we need to show all those arrays of data on the UI), I've decided to port JS version of ABI codec from ethereumjs-abi - it's relatively small, easy to understand, also we encode only primitive types for EIP712 sake. Also, it's used by the aforementioned eth-sig-util lib from MetaMask. Some helper functions for detecting type properties were useful for the client implementation as well.

UI/UX Flow

And the most important part of this implementation is the interface for easy data introspection. Based on standard Trezor Layout practics, it gradually briefs users about the data being signed and allows them to descend into the deepest depths of the schema, if the user wants that.

A typical signing flow looks like this on the device:

1. Domain brief 👀

We present some brief about EIP712Domain, such as Name and Version data from it, so the user can know the vendor of the message immediately. User may choose to inspect the domain closely, a good thing it's a plain simple struct.

2. Domain inspection 👀

Structs are shown as paginated lists of fields. In the title, there's a struct name, also a pagination mark. The pagination mark shows the current field number and total field number in the struct. It's very simple for a static domain struct.

3. Message brief 👀

Once the user dealt with domain verification, we show him the message brief. It's impossible to print all fields from the message, so we decided to compose a maximum of 3 elements and note that there are more to inspect. Many times the user would skip checking the message thoroughly just because he trusts the domain already, and sees some of the fields of the message. The title contains the primary_type of the message.

4. Message inspection 👀

And this is where the rabbit hole begins 🐰. Since the combination of types, structs and arrays is completely arbitrary and depends only on the application, it's up to app developer to design messages in a clean way. Our Cosmos transactions converted into EIP-712 are not the best UX example, but at least it's a good durability test for the inspection.

👆This is a reference example of a typed data transaction shown in MetaMask

The inspection flow on Trezor consists of the following steps:

For a struct we ask user to inspect a field if it's also a struct, or array. In our example, we'd ask about field fee 3/6 (type Fee) and msgs 5/6 (type []Msgs)
For every internal struct we repeat 1.
After the user reviewed or skipped those fields, we show him the whole struct paginated.

Some tricky examples:

(1a)(1b)(2a)(2b)(3a)(3b)

1a) Asked to inspect 3/6 field named fee in primary struct (Tx). Struct has type Fee and has 2 fields.
1b) Looking at the first 1/2 field in the Fee struct named amount. It's array []Coin with only 1 element.
2a) Asked to inspect value field of Msg struct as the first element in msgs []Msg array field in the primary struct. The value struct has type MsgValue and has 4 fields.
2b) Asked to view the latest field (4/4) of MsgValue, which is called value.bridge_fee (gracefully trimmed), it's a TypeBridgeFee type and it has 2 fields inside.
3a) Looking at second (2/4) field of MsgValue in the paginated struct overview. It's a simple string field named eth_dest.
3b) Looking at the latest (6/6) field of the primary struct (Tx) paginated for final overview. At this point, the user has the option to approve the primary struct, and finalize the signing. Trezor emits a signature.

Short path:

Obviously, not everybody needs a full introspection of the typed data mazes. There's always a short path for typed data signing without the need to look into data at all. It takes only 2 clicks and 1 hold to sign a typed data hash.

Notes

Support for EIP-712 is not implemented for Trezor T1 model (Legacy) yet, but it's far more trivial to implement. The approach would be the same as for the Ledger app, where the device only accepts 2 hashes (domain separator, a message hash) and simply signs the EIP-712 digest. Maybe it will be implemented in the upcoming days.

This PR:

Resolves Add support for EIP-712 #1272
Updates Implement support for EIP-712 #131
Updates Implement Ethereum Sign/Verify message once EIP712 is widely adopted trezor-core#122
Supersedes WIP: Support for EIP-712 #148

bangjelkoski · 2021-04-22T23:22:10Z

Any update on this @matejcik @tsusanka?

xlab · 2021-04-25T18:07:52Z

PTAL @prusnak, is there something that holds from reviewing this feature implementation?

prusnak · 2021-04-25T18:11:55Z

PR looks rather fine. Thank you for that!

We are in the process of working on stuff that landed in our pipeline earlier than this PR. Once our engineers have more spare cycles they'll review the PR and provided more detailed feedback.

Howard3 · 2021-05-05T20:01:49Z

Heads up, without this we cannot sign for Uniswap v3 - this is a big deal and would prevent all trezor users from making use of their liquidity.

PSA10GEMMT · 2021-05-09T16:27:14Z

I'm new to github and trying to understand how I can check progress of tickets, in particular this one and when it will be complete. This is a pretty big deal. Thanks for all the work devs!

PRMomma · 2021-05-16T00:09:57Z

Any feedback for integrating this into the next firmware update?

tsusanka · 2021-05-20T12:36:04Z

Hi everyone, thanks a lot for the PR. We won't be able to include this in the upcoming June release. Thank you for your patience!

PRMomma · 2021-05-20T13:12:58Z

Tomas, that is quite unfortunate. Many of us on Injective Protocol Discord and Telegram have been discussing this issue. Many of us couldn't participate in the INJ Equinox Testnet and earn APY because we had our INJ tokens in the Trezor wallet. We have been patiently waiting for the firmware update for this.

INJ will be launching mainnet very soon, anytime now within 2Q (before end of June 2021), and we were all hoping to use our Trezor wallets to start staking upon the INJ mainnet launch. Is there anything you can manage to get this included asap?

Otherwise, all of us would have to get a Ledger and migrate all of our INJ from Trezor to Ledger since Ledger supports EIP-712.

Please advise?

bangjelkoski · 2021-05-20T13:54:46Z

This PR concerns not only Injective Protocol users but Uniswap v3 users as well. The number of people that are restricted in using both Injective Protocol apps and Uniswaps V3 update is enormous. Please reconsider having the update in the next firmware release.

Kind regards.

prusnak · 2021-05-21T07:43:23Z

Please keep the discussion on-topic. Off-topic comments will be removed.

Even if the EIP-712 support was merged into the Firmware, there is still Connect part of things missing, so it's not like the support will be immediately usable after this PR is merged.

mmilata

Hi Max! Thank you for the pull request, I went through the changes and got some preliminary feedback. I think the overall design looks good, however there's couple of things we need before we can consider merging:

There has to be good test coverage - encoders/decoders like this one often have tricky corner cases.
CI pipeline needs to pass: https://gitlab.com/satoshilabs/trezor/trezor-firmware/-/pipelines/310943878 - the legacy jobs probably failed due to "gen prebuild". Let me know when you want to rerun it.

We'll also need the Product team to review the UI.

mmilata · 2021-05-27T22:02:12Z

core/src/apps/ethereum/abi.py

+    raw_type = parsed_type["raw_type"]
+
+    print("abi_decode_single", parsed_type, "data:", data, "offset:", offset)
+    print("type_name", type_name, "raw_type", raw_type)


No print()s please.

Good catch, leftovers from debug :)

Deleted in a05275a

mmilata · 2021-05-27T22:07:22Z

common/protob/messages-ethereum.proto

+ */
+message EthereumSignTypedData {
+    repeated uint32 address_n = 1;           // BIP-32 path to derive the key from master node
+    optional bool use_v4 = 2 [default=true]; // use EIP-712 (v4) for typed data signing


What exactly does false mean, that v3 is used? From OP and metamask docs it appears that uint or enum field would make sense here if there's a chance of future versions.

Metamask has its own implementation history, all v1-v2-v3 were not compliant with the latest standard and should not be considered.

Their implementation v4 is widely adopted now, but still, has a bug described there:
MetaMask/eth-sig-util#107 (comment)

So at one day, it might be a v5.

I agree that having version enumeration is better than trying to maintain only a single or two "true" versions. I'm not a big fan of maintaining that as part of HW, but we have no choice there.

mmilata · 2021-05-27T22:10:51Z

pyproject.toml

@@ -45,6 +45,7 @@ autoflake = "*"
 flake8-requirements = ">=1.3.2"

 # common
+eth_abi = "^2.1.1"


I think this dependency belongs to python/setup.py in extras_require, this pyproject.toml is just for the development/build environment.

Addressed in a05275a

mmilata · 2021-05-27T22:15:50Z

python/src/trezorlib/cli/ethereum.py

+@cli.command()
+@click.option("-n", "--address", required=True, help=PATH_HELP)
+@click.option("--use-v4", type=bool, default=True, required=False, help=TYPED_DATA_USE_V4)
+@click.argument("typed_data_json")


The jsons are quite large, would be nice to be able to read them from file. With @click.argument("file", type=click.File("r")) you can still use stdin as -.

Implemented in a0ce878

mmilata · 2021-05-27T22:17:48Z

python/src/trezorlib/messages/__init__.py

@@ -2,332 +2,668 @@
 # fmt: off

 from .Address import Address
+from .Address import Address


Something bad happened to this file:(

mmilata · 2021-05-27T22:40:02Z

python/src/trezorlib/ethereum.py

+
+    m = type_name_re.match(primary_type)
+    if m:
+        primary_type = m.string[m.start():m.end()]


Suggested change

primary_type = m.string[m.start():m.end()]

primary_type = m.group(0)

^should also work. Also what's the point of the match, strip [] from the end? Can primary type be an empty string?

If it's more canonical then OK. I'm from Go background, for us, the explicit code is better.
The goal is to strip also [][5] and whatever it might have on the end. We're interested in the Type name only.

Can primary type be an empty string?

The type name must be a valid identifier as a name in Solidity. So the answer is no.

Yeah, it's how it's usually written. If it can't be empty then let's change the regex to ^\w+?

Changed in a05275a

mmilata · 2021-05-27T22:41:22Z

python/src/trezorlib/ethereum.py

+    for key in allowed_typed_data_properties:
+        val = data.get(key)
+        if val is None:
+            continue


Looking at the spec it seems that the 4 data properties are in fact required?

Good catch. I was following this code https://github.com/MetaMask/eth-sig-util/blob/41a62c8255ce061130f85876be178e65e6ae99a2/src/index.ts#L270-L289

But didn't notice they actually check the result using TypeScript's type matching Required<TypedMessage<T>>.

I think we need to require every key here.

Addressed in a05275a

mmilata · 2021-05-27T22:42:18Z

python/src/trezorlib/ethereum.py

+        sanitized_data[key] = val
+
+    if "types" in sanitized_data:
+        sanitized_data["types"] = { "EIP712Domain": [], **sanitized_data["types"] }


Why not sanitized_data["types"]["EIP712Domain"] = []?

I don't get the problem here. This line was a carbon copy of this one:
https://github.com/MetaMask/eth-sig-util/blob/41a62c8255ce061130f85876be178e65e6ae99a2/src/index.ts#L286

I tried to stay close to that implementation. Isn't this a correct translation?

I'm not familiar with TypeScript, the translation is probably alright, just very unusual to write it like this in python.

that might be idiomatic typescript (presumably? i'm not fluent) but it's decidedly not idiomatic Python. let's go with @mmilata's suggestion instead

It seems this line has a meaning in the way it is written - we want to make sure the EIP712Domain key is always there and defaulting to an empty list. When it is already present in sanitized_data["types"], it will then overwrite this empty list with current value (therefore the dict unpacking with **).

What is the best way of handling this situation? We could do an explicit check for this:

if "EIP712Domain" not in sanitized_data["types"]: sanitized_data["types"]["EIP712Domain"] = []

When using only sanitized_data["types"]["EIP712Domain"] = [], we received FirmwareError (99) - most probably is complaining about the EIP712Domain being empty (it probably shouldn't though)

Changed in a05275a

mmilata · 2021-05-27T22:43:43Z

python/src/trezorlib/ethereum.py

+
+    return result, result_indexed
+
+allowed_typed_data_properties = ["types", "primaryType", "domain", "message"]


Constant sequences are usually defined as tuple and named in SCREAMING_SNAKE_CASE.

Changed in a05275a

mmilata · 2021-05-27T22:44:02Z

python/src/trezorlib/ethereum.py

+
+def is_array(type_name: str) -> bool:
+    if type_name:
+        return type_name[len(type_name) - 1] == ']'


Suggested change

return type_name[len(type_name) - 1] == ']'

return type_name[-1] == ']'

Changed in a05275a

xlab · 2021-05-28T14:02:44Z

Hey, thanks for getting some time to review this. I really appreciate this is moving :)

Yeah, I agree that some tests are required, especially for the encoders and decoders I implemented. I was waiting until the logic itself is reviewed.

So my next steps would be:

I'll revisit the code and fix code at your comments, also will resolve conflicts accumulated.
Get some tests for encoding.
I'll try to see what needs to be done on the "Connect" client part. The client-side requires some logic as well, currently such logic exists only for CLI in Python.

mmilata · 2021-05-31T11:33:31Z

Yep, sounds good! As part of testing it would also be interesting to find out what's the largest structure the device can sign before running out of memory, or at least make sure that it doesn't happen with structures that have "normal" size.

tsusanka · 2021-06-11T10:56:32Z

Hello everyone. Just to inform you all about the state of things, we will need our Product team to look into this as we are not sure showing such advanced structures on Trezor is a good idea. We are really trying to push usability of Trezor the right direction and we are worried users will just blindly confirm whatever structure is presented before them which is something we do not want to encourage. We will get back to you soon on this.

xlab · 2021-06-11T11:17:26Z

showing such advanced structures

Most of EIP712 use cases are rather short. Just domain confirmation and then a small message of 1-2 fields. The alternative approaches would be to blindly approve hash (Ledger's approach) or just don't support EIP 712 and keep user's funds locked.

The approach proposed here allows skipping message verification for known domains. Or skip everything, just approving the hash - in this way it falls back to the same flow as Ledger already has.

For people who decided to actually study the data they signing, I was very glad that Trezor can bootstrap a nice UI to walk through it.

My example with our message is an extreme case that aims to validate the approach on big payloads. It's easy to implement bad UI when use cases just contain a few strings.

blueblack2 · 2021-06-12T03:18:27Z

Hello everyone. Just to inform you all about the state of things, we will need our Product team to look into this as we are not sure showing such advanced structures on Trezor is a good idea. We are really trying to push usability of Trezor the right direction and we are worried users will just blindly confirm whatever structure is presented before them which is something we do not want to encourage. We will get back to you soon on this.

Any approximate timeline on how long this will take? As earlier posters have pointed out it sounds like this is needed for Uniswap V.3 as well as for Injective staking. If you foresee this taking several months longer please let us know so we can look into other options like a Ledger. Thank you.

0xGitGuy · 2021-09-16T18:13:38Z

If we get the Trezor One or Trezor T now, can we sign eip 712 signTypedData_v4 transactions as of now? After reading all the comments, it's unclear and with all the activities happening on Ethereum, this is holding trezor users back.

ernest22 · 2021-09-24T18:17:39Z

This is stoping people to use their trezor for opensea on matic, can we speed this up please for the users and community

AarbozT · 2021-09-28T08:34:15Z

Sorry guys for been recursive as I replied on #1272 , but this issue is quite critical.

Yes guys, we need this to be implemented! Signing in OpenSea doesn't even works with a Trezor One (with latest firmware v1.10.3) connected to MetaMask.
This issue must be critical as OpenSea is used by millions of people.
Screenshots:

You guys can do it!
Thanks!!!

gewure · 2021-10-22T16:30:25Z

We are also blocked in our development by this issue.

Thanks for solving in advance.

matejcik · 2021-10-26T11:01:45Z

Thank you @xlab for this contribution. We have overhauled your implementation and are tracking the modified version as #1835.

remixie · 2021-12-01T15:13:57Z

@matejcik has these changes gone live yet as a firmware update? I'm here losing a ton of money over here. I want my NFTs out of trezor already. This is insanely unacceptable.

roshhhyy · 2021-12-01T15:53:55Z

trezor has completely abandoned their original flagship product, the trezor one, and my nfts stuck in there will be forever. They’re also actively false advertising by saying the trezor one supports ETH yet it doesn’t support one of the most critical EIPs in existence. Even worse is there are many people with NFTs trapped in trezor ones and we are all collectively ignored and told to extract our seeds to software wallets or the model T. Ledger supported ETH fully on the Nano S a device with less memory than a flip phone from 1996, trezor should be embarrassed

mohamedmansour · 2021-12-01T16:05:44Z

@roshhhyy @remixie since they don't support new transactions, you can rollback to a year old metamask extension to support older transactions, you will pay more since you cannot control basefee and priorityfee, but hopefully you will be able to do legacy transactions. Regarding signing newer transactions, you are unfortunately stuck until metamask implements it, since trezor was very very late to implementing EIPs.

But yea, Trezor is unfortunately a Bitcoin first company, but majority of its users are Ethereum.

remixie · 2021-12-01T16:33:57Z

@roshhhyy @remixie since they don't support new transactions, you can rollback to a year old metamask extension to support older transactions, you will pay more since you cannot control basefee and priorityfee, but hopefully you will be able to do legacy transactions. Regarding signing newer transactions, you are unfortunately stuck until metamask implements it, since trezor was very very late to implementing EIPs.

But yea, Trezor is unfortunately a Bitcoin first company, but majority of its users are Ethereum.

@mohamedmansour I have tried the tactic of using metamask 9.8.4. I get the same errors on opensea. I have tried multiple computers, reconnecting my trezor, reinstalling metamask on a clean profile for both browsers, everything but I still get stopped by trezor's lack of support rn

prusnak · 2021-12-06T12:25:33Z

EIP-712 for Trezor One is being developed here and will be hopefully released in January: #1970

AarbozT · 2021-12-07T00:00:58Z

EIP-712 for Trezor One is being developed here and will be hopefully released in January: #1970

Wow! Thanks! That's what I'm talking about! Being unstoppable and decided guys! I'm going to remove my complaint tweet now regarding the other ticket! I hope everything goes well with that and that you comply with it!

roshhhyy · 2021-12-07T00:02:34Z

EIP-712 for Trezor One is being developed here and will be hopefully released in January: #1970

thank you this means a lot

AarbozT · 2021-12-07T00:17:03Z

EIP-712 for Trezor One is being developed here and will be hopefully released in January: #1970

Wow! Thanks! That's what I'm talking about! Being unstoppable and decided guys! I'm going to remove my complaint tweet now regarding the other ticket! I hope everything goes well with that and that you comply with it!

Done, I deleted, but sad, no idea you blocked me on twiter. I run out of hard feelings. So i will be alone with my Trezor and other people I recommended months ago!

PRMomma · 2021-12-07T00:29:23Z

So what is the timeframe for Trezor Model T?

remixie · 2021-12-07T00:38:27Z

After much much investigation, I found out that I can export my trezor private wallet key by following this youtube tutorial:

https://www.youtube.com/watch?v=aBy3EUqQLrg

I installed Tails onto a virtual box,
Downloaded the BIP39 Tool inside of the Tails VM.
Turned off my internet connection of my computer.
Inputted my mnemonic phrase and passphrase
Chose my crypto coin
Scrolled down to the table
and finally, scanned the private key QR code of the wallet I needed with Metamask mobile.

Note: you have to hover over the private key for the QR code to appear

By doing these steps, I was able to take out my NFT's from my trezor and put them in my ledger right away.

bondo1079 · 2022-01-16T13:22:05Z

i tred to enter it withmeta mask wallet, but cant, then most of my etheirium is gone less than1% was lest in my account. what do i do

fun4cash · 2022-01-24T10:03:44Z

why cant i sign in to dcl and pls join my account i cant open more im getting lost
fun4 cash and necitizen2

Maxim Kupriianov added 3 commits April 13, 2021 15:46

Implemented EIP-712 typed data signatures in Ethereum app.

d0d4c82

make protobuf

0135ed9

Add eth_abi into pyproject deps

52ed7bf

xlab requested review from matejcik and tsusanka as code owners April 13, 2021 21:59

This was referenced Apr 13, 2021

eth_signTypedData on Trezor account fails without callback MetaMask/metamask-extension#5052

Open

Support EIP-712 signing for Ledger accounts MetaMask/metamask-extension#10240

Closed

trezor deleted a comment from thedarkring May 14, 2021

tsusanka added this to the 21.07 milestone May 20, 2021

trezor deleted a comment from PRMomma May 21, 2021

trezor deleted a comment from blueblack2 May 21, 2021

trezor deleted a comment from PRMomma May 23, 2021

mmilata requested changes May 28, 2021

View reviewed changes

trezor deleted a comment from lamart1 May 31, 2021

szymonlesisz mentioned this pull request Jun 21, 2021

Ethereum EIP-712 structured signatures isn't supported trezor/connect#825

Closed

trezor deleted a comment Sep 15, 2021

fubuloubu mentioned this pull request Sep 20, 2021

EIP-712 support ApeWorX/ape-trezor#8

Open

grdddj mentioned this pull request Oct 6, 2021

EIP-712 support for Trezor T #1835

Merged

wyc mentioned this pull request Oct 14, 2021

Create EIP-4361: Sign-In with Ethereum ethereum/EIPs#4361

Merged

4 tasks

matejcik closed this Oct 26, 2021

evilrobot-01 mentioned this pull request Nov 29, 2021

Make Opensea transactions work BitBoxSwiss/bitbox02-firmware#843

Closed

xlab deleted the ethereum-eip712 branch December 7, 2021 18:07

bangjelkoski mentioned this pull request Mar 10, 2022

EIP712 signing fails when struct is only used as an array element #2167

Closed

	primary_type = m.string[m.start():m.end()]
	primary_type = m.group(0)


		return result, result_indexed

		allowed_typed_data_properties = ["types", "primaryType", "domain", "message"]

	return type_name[len(type_name) - 1] == ']'
	return type_name[-1] == ']'

Adding support of EIP-712 into Ethereum app #1568

Adding support of EIP-712 into Ethereum app #1568

Conversation

xlab commented Apr 13, 2021 • edited Loading

Intro

Data exchange flow

Ethereum ABI

UI/UX Flow

1. Domain brief 👀

2. Domain inspection 👀

3. Message brief 👀

4. Message inspection 👀

Notes

bangjelkoski commented Apr 22, 2021 • edited Loading

xlab commented Apr 25, 2021 • edited Loading

prusnak commented Apr 25, 2021 • edited Loading

Howard3 commented May 5, 2021

PSA10GEMMT commented May 9, 2021

PRMomma commented May 16, 2021

tsusanka commented May 20, 2021 • edited Loading

PRMomma commented May 20, 2021 • edited Loading

bangjelkoski commented May 20, 2021

prusnak commented May 21, 2021

mmilata left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

xlab May 28, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

xlab commented May 28, 2021

mmilata commented May 31, 2021

tsusanka commented Jun 11, 2021

xlab commented Jun 11, 2021 • edited Loading

blueblack2 commented Jun 12, 2021

0xGitGuy commented Sep 16, 2021

ernest22 commented Sep 24, 2021

AarbozT commented Sep 28, 2021

gewure commented Oct 22, 2021

matejcik commented Oct 26, 2021

remixie commented Dec 1, 2021 • edited Loading

roshhhyy commented Dec 1, 2021

mohamedmansour commented Dec 1, 2021

remixie commented Dec 1, 2021

prusnak commented Dec 6, 2021

AarbozT commented Dec 7, 2021

roshhhyy commented Dec 7, 2021

AarbozT commented Dec 7, 2021 • edited Loading

PRMomma commented Dec 7, 2021

remixie commented Dec 7, 2021 • edited Loading

bondo1079 commented Jan 16, 2022

fun4cash commented Jan 24, 2022

xlab commented Apr 13, 2021 •

edited

Loading

bangjelkoski commented Apr 22, 2021 •

edited

Loading

xlab commented Apr 25, 2021 •

edited

Loading

prusnak commented Apr 25, 2021 •

edited

Loading

tsusanka commented May 20, 2021 •

edited

Loading

PRMomma commented May 20, 2021 •

edited

Loading

xlab May 28, 2021 •

edited

Loading

xlab commented Jun 11, 2021 •

edited

Loading

remixie commented Dec 1, 2021 •

edited

Loading

AarbozT commented Dec 7, 2021 •

edited

Loading

remixie commented Dec 7, 2021 •

edited

Loading