chore(pevm): Post-processing lazy evaluation: assert that the evaluated nonce matches the transaction's

[tsnewnami]

This PR addresses #310

[tsnewnami]

Hey @hai-rise, I've tried to have a crack at this one. And I'm not sure if I've identified a bug or not (thinking possibly yes as the other MemoryValue::Basic checks pass).

The mainnet_blocks_from_disk is failing. I'd love to try and debug this more to better understand the stack, but I may need some pointers if that's okay 🙏 .

Just for my understanding of what's going on here: this is essentially validating the incarnation (tx) that resulted in the particular state mutation of an account, which is encoded in the write_history. Where the write_history holds all particular state updates for a given memory location from the resulting incarnations?

Here is the particular tx_env where this check fails:

TX TxEnv {
    caller: 0xa9cefd4474086c52f15abc0e604284440768da57,
    gas_limit: 23100,
    gas_price: 54641531805,
    transact_to: Call(
        0xcff0b829fd9bd7a58ca8d62eef8105506f859592,
    ),
    value: 500000000000000,
    data: 0x,
    nonce: Some(
        18,
    ),
    chain_id: Some(
        1,
    ),
    access_list: [],
    gas_priority_fee: Some(
        3358468195,
    ),
    blob_hashes: [],
    max_fee_per_blob_gas: None,
    authorization_list: None,
}

Tyvm.

[hai-rise]

Ah, tx.nonce only applies to the sender, while we can have the recipient (or more addresses in the future) in MemoryValue::Basic and other arms.

I think we can add this line below the match memory_entry:

debug_assert!(tx.caller != address || tx.nonce.map_or(true, |n| n + 1 == nonce));

[tsnewnami]

Thank you for the feedback @hai-rise. That makes sense. Was way off the mark with that bug... Thank you for the learning here.

[hai-rise]

Thank you!