Dror Tirosh

Yes, this is not a problem of **typechain** per se, but the **hardhat-typechain** plugin. The plugin automatically generates types on `hardhat compile` task, but doesn't automatically generate types on `hardhat...

This issue is a major use-case of the #12725 issue. I do think this issue requires language support and not require developers to resort to assembly.

didn't get any comment on this. I think it should be quite easy to implement, as currently it is quite annoying to read through the assembly code.

actually, there is a simpler solution: the comment can be re-constructed from the methodsig.

1. You could lazily resolve remappings, (cast command don't use remappings at all) 2. Rethink if remappings if all indirectly referenced projects is required.. I have a project with 4-5...

Yes, it is a valid issue, that a manually created structures can create different and even non-unique UserOp hash. However: 1. While it is discouraged, userop uniqueness is up to...

The `validateUserOp` is given a userOpHash to the validateUserOp, which is a hash over all fields - including all offsets to dynamic parameters, so that a bundler is prevented from...

Not checking the **userOpHash** is dangerous: you expose your account to replay attacks, or gas-attacks (where someone passes huge values for gaslimit/gasprice) to drain your balance. Even if it validates...

> Our logic of skipping signature check is based on enabled 'module' as for example social recovery module recoverUsingSocial() I don't understand: anonymous users can execute this recover module, at...

> In any case, these solutions don't in themselves solve the problem I brought up, since the verifier still needs to be aware of ERC-4337 and update their verification flow...