feat(cli): add FRC-102 compatible EIP-191 signing via --fevm flag (#13256)

[aceppaluni]

Related Issues

This PR aims to fix Issue #13256

Proposed Changes

This PR updates the lotus wallet sign command to implement FRC-102, allowing users to sign messages with an EIP-191 (Ethereum-style) prefix for FEVM compatibility.

Additional Info

• Added --fevm flag to wallet sign command.
• Prepends the appropriate prefix to messages before signing (EIP-191 for FEVM, Filecoin prefix for native messages).
• Supports interoperability with other Filecoin wallets using FRC-102.
• Returns signatures in the same format as the existing command.

Checklist

Before you mark the PR ready for review, please make sure that:

• Commits have a clear commit message.
• PR title conforms with contribution conventions
• Update CHANGELOG.md or signal that this change does not need it per contribution conventions
• New features have usage guidelines and / or documentation updates in
  • Lotus Documentation
  • Discussion Tutorials
• Tests exist for new functionality or change in behavior
• CI is green

[aceppaluni]

Hi all 👋 Just wanted to follow up on this PR.
No rush.
Please let me know if there’s anything I can update to help move it forward!

[rvagg]

@hugomrdias @eshon @bumblefudge can you help us out here -- this PR does two things, it changes the default signature to be the version 0x45 (E) style signature but adds a --fevm argument to do the version 0x46 (F) style signature. I don't use this API for signing and I don't know what current users would expect but this would be a breaking change.
There are 4 variants in FRC-0102, should we support all of them and a bare form for backward compatibility? We could do --type=[0|1|E|F|bare] (defaulting to F perhaps?). What do ledger and filsnap do here?

[hugomrdias]

@hugomrdias @eshon @bumblefudge can you help us out here -- this PR does two things, it changes the default signature to be the version 0x45 (E) style signature but adds a --fevm argument to do the version 0x46 (F) style signature. I don't use this API for signing and I don't know what current users would expect but this would be a breaking change. There are 4 variants in FRC-0102, should we support all of them and a bare form for backward compatibility? We could do --type=[0|1|E|F|bare] (defaulting to F perhaps?). What do ledger and filsnap do here?

in this command we only care about 0x46 "personal_sign", filsnap and ledger only implement 0x46 for personal messages.

yes, its a breaking change there no more unprefixed signatures exactly frc-0102 purpose

https://github.com/hugomrdias/filecoin/blob/main/packages/iso-filecoin/src/wallet.js#L274-L285

[rvagg]

in this command we only care about 0x46 "personal_sign", filsnap and ledger only implement 0x46 for personal messages.

OK, @aceppaluni would you mind please just removing the --fevm and the branch that allows that and just going with the 0x46 prefixed version for all signing.

But then also, we need to fix the "verify" command to be the reverse of "sign", so it also needs to know about the prefix. When you're done, you should be able to pass the output of "sign" to "verify" and it pass.

[BigLep]

@aceppaluni : please fil any CI checks that are failing (e.g., gofmt) and then go ahead and mark "ready for review" (assuming other feedback has been incoporated).

[rvagg]

I've just noticed we have 3 issues that this will close:

• Should we use have a EIP191 like arbitrary bytes signatures #12765
• Implement FRC-102 in Lotus CLI for consistency across Filecoin wallets #13256
• Add a Signing Envelope (FRC-102) when signing arbitrary messages in Lotus CLI (and API?) #13373

So to get them off the board I've been a bit heavy handed in here and did a bunch of clean up, expanded tests, got the prefix correct (I think it got messed up somewhere in a merge/rebase) and a --raw flag working.

Thanks for the work on this @aceppaluni.