Yoav Weiss

I already tested an identical fix, it stops the false positive and the signal gets recognized correctly. But I don't have any gate_tx device so I can't verify that it...

This question has come up a couple of times before. There are a few different ways to address it. Extending 1271 to interact with a factory contract though a new...

What @drortirosh described is a generalization of (4) and seems like the cleanest approach. Makes it possible to simulate account creation combined with an arbitrary call (such as isValidSignature). It...

How will the magic byte be distinguished from a signature? isValidSignature doesn't require a specific type of signature, so any byte we choose could also be a valid first byte...

If it's a new standard anyway, wouldn't it be safer to make it explicit and not embed it in a single signature field?

We could do that with a prefix but it would have to be quite long to make collisions unlikely. How many transporters would have to be updated to support the...

@Ivshti The proposal looks good, and I also think the solution should be an extension of 1271 rather than 4337. It was easier to solve it in EntryPoint as part...

The singleton approach keeps the implementation simple for library maintainers, and doesn't add any cost (in terms of `eth_call`s). They already have to perform at least one call to support...

This issue has become more critical with the growing popularity of account abstraction, so I would like to resume this discussion. Account abstraction adds a new challenge that, to the...