`canonical_execution_contracts` incorrect `code_hash` and `init_code_hash`

[weiihann]

I think this is not Xatu's problem but rather Cryo, but would like to flag it here in case somebody else has the same confusion.

Essentially, the code_hash and init_code_hash columns in the canonical_execution_contracts table are swapped.

Evidence 1: Compare with Geth

SELECT
    contract_address,
    block_number,
    init_code_hash,
    code_hash
FROM canonical_execution_contracts
FINAL
WHERE contract_address = '0x3cf9e625387a35863280e064db9b57c0cf8013d2'
LIMIT 1

Query id: 45bc7144-4382-4ef8-8360-95d89cbd2e47

   ┌─contract_address───────────────────────────┬─block_number─┬─init_code_hash─────────────────────────────────────────────────────┬─code_hash──────────────────────────────────────────────────────────┐
1. │ 0x3cf9e625387a35863280e064db9b57c0cf8013d2 │     19000000 │ 0xb89c1b3bdf2cf8827818646bce9a8f6e372885f8c55e5c07acbd307cb133b000 │ 0x56e3081a3d1bb38ed4eed1a39f7729c3cc77c7825794c15bbf326f3047fd779c │
   └────────────────────────────────────────────┴──────────────┴────────────────────────────────────────────────────────────────────┴────────────────────────────────────────────────────────────────────┘

Comparing it with a geth node:

Command: geth snapshot inspect-account 0x3cf9e625387a35863280e064db9b57c0cf8013d2

Disklayer: Root: 0673b60ffe3c1e8f18859908b1b032427a53263632d95b59f0495d27e7bcebdd
	account.nonce: 1
	account.balance: 0
	account.root: 000f3f8d0b4b80f6d9d384e239166ebb889dd788d08a168328d8127344033958
	account.codehash: b89c1b3bdf2cf8827818646bce9a8f6e372885f8c55e5c07acbd307cb133b000
	Storage:
		...

As shown, the address has a code hash of 0xb8...00, but in the table it appears in the init_code_hash column rather than code_hash.

Evidence 2: Mismatch number of unique code hashes

Distinct init_code_hash

SELECT countDistinct(init_code_hash)
FROM default.canonical_execution_contracts AS c
FINAL

Query id: 6e108300-25aa-41b8-ab40-67ea94221ef5

   ┌─countDistinc⋯_code_hash)─┐
1. │                  1762040 │ -- 1.76 million
   └──────────────────────────┘

Distinct code_hash

SELECT countDistinct(code_hash)
FROM default.canonical_execution_contracts AS c
FINAL

Query id: 643a6b22-81aa-484c-8f82-4063bde161b8

   ┌─countDistinct(code_hash)─┐
1. │                  4543590 │ -- 4.54 million
   └──────────────────────────┘

I know that from Geth and Reth node, the number of unique code hashes should only be around 1.7M. So that's another evidence that the columns are swapped.

n_code_bytes and n_init_code_bytes remain correct.

[AuburyEssentian]

Confirmed this is a Cryo bug. In crates/freeze/src/datasets/contracts.rs, the two keccak256 calls are reversed:

// Current (wrong):
store!(schema, columns, init_code_hash, keccak256(result.code.clone()).to_vec());
store!(schema, columns, code_hash,      keccak256(create.init.clone()).to_vec());

init_code_hash hashes the deployed runtime bytecode (result.code), and code_hash hashes the initialization bytecode (create.init) — exactly backwards.

Submitted a fix upstream: paradigmxyz/cryo#249