fix(anvil): fix incorrect op-stack deposit tx hashes

[jakim929]

Motivation

tx hash calculation for deposit tx for anvil --optimism was different from specs. this made it difficult to deterministically generate the L2 tx hash from the L1 deposit transaction.

• the EIP-2718 tx type was not included when calculating hash
• encoding was slightly different from spec
  • nonce was included in the rlp encoding
  • order of fields was different from spec

Solution

• calculate tx hash after encoding in eip-2718 envelope format
• fix order of fields / remove nonce when RLP encoding / decoding
• added tests - checks against a op-sepolia tx hash / fields to make sure that the hash / encoding is consistent

Test to repro issue

the following test would fail previously

    fn test_decode_encode_deposit_tx() {
        // https://sepolia-optimism.etherscan.io/tx/0xbf8b5f08c43e4b860715cd64fc0849bbce0d0ea20a76b269e7bc8886d112fca7
        let tx_hash: TxHash = "0xbf8b5f08c43e4b860715cd64fc0849bbce0d0ea20a76b269e7bc8886d112fca7"
            .parse::<TxHash>()
            .unwrap();

        // https://sepolia-optimism.etherscan.io/getRawTx?tx=0xbf8b5f08c43e4b860715cd64fc0849bbce0d0ea20a76b269e7bc8886d112fca7
        let raw_tx = alloy_primitives::hex::decode(
           "7ef861a0dfd7ae78bf3c414cfaa77f13c0205c82eb9365e217b2daa3448c3156b69b27ac94778f2146f48179643473b82931c4cd7b8f153efd94778f2146f48179643473b82931c4cd7b8f153efd872386f26fc10000872386f26fc10000830186a08080",
        )
        .unwrap();
        let dep_tx = TypedTransaction::decode(&mut raw_tx.as_slice()).unwrap();
        let mut encoded = Vec::<u8>::new();
        dep_tx.encode_2718(&mut encoded);
        assert_eq!(raw_tx, encoded);
        assert_eq!(tx_hash, dep_tx.hash());
    }

[mattsse]

this is great, ty

few pedantic nits

we now have this type also here https://github.com/alloy-rs/op-alloy so we could replace the entire anvil deposit type, will track this separately.

q, are you using eth_sendraw deposit txs at OP for testing?

[jakim929]

thanks for the quick review! 🙏 addressed the requested changes

q, are you using eth_sendraw deposit txs at OP for testing?

Yep, it's very useful for testing bridging! and we're using it pretty extensively for https://github.com/ethereum-optimism/supersim , which is a local environment for testing L1 + multiple L2s (anvils) for bridging, message passing etc.

Also we're happy to help maintain the optimism related features on anvil including tasks like

we now have this type also here https://github.com/alloy-rs/op-alloy so we could replace the entire anvil deposit type, will track this separately.

and improving support for deposit tx validation in the tx pool.