convert create2 test

[winsvega]

🗒️ Description

Port

• src/GeneralStateTestsFiller/stCreate2/call_outsize_then_create2_successful_then_returndatasizeFiller.json
• src/GeneralStateTestsFiller/stCreate2/call_then_create2_successful_then_returndatasizeFiller.json

🔗 Related Issues

none

✅ Checklist

NOT READY FOR MERGE (NEED coverage)

• All: Set appropriate labels for the changes.
• All: Considered squashing commits to improve commit history.
• All: Added an entry to CHANGELOG.md.
• All: Considered updating the online docs in the ./docs/ directory.
• Tests: All converted JSON/YML tests from ethereum/tests have been added to converted-ethereum-tests.txt.
• Tests: Included the type and version of evm t8n tool used to locally execute test cases: e.g., ref with commit hash or geth 1.13.1-stable-3f40e65.
• Tests: Ran mkdocs serve locally and verified the auto-generated docs for new tests in the Test Case Reference are correctly formatted.

[winsvega]

@marioevz do you want the test coverage to be mandatory checkpoint as well? or perhaps we can setup a cli job to do it using evmone script that I was developing. hm..

[marioevz]

@marioevz do you want the test coverage to be mandatory checkpoint as well? or perhaps we can setup a cli job to do it using evmone script that I was developing. hm..

I think we could start with a CLI entry point to check coverage, if you have an idea of implementation we could open another PR and follow it up there.

After that PR is merged we can do a follow up PR to try to embed it into the CI to automatically check it.

[marioevz]

Looks great overall! Just some suggestions.

[marioevz]

Suggested change

	+ Op.SSTORE(slot_code_worked, 1)
	+ Op.SSTORE(slot_code_worked, 1)
	+ Op.RETURNDATACOPY(0, 0, Op.RETURNDATASIZE())
	+ Op.SSTORE(slot_return_data_sha3, Op.SHA3(0, Op.RETURNDATASIZE()))

What do you think about also SHA3-comparing the return data just to test the integrity of the return data after the create opcode has been called?

We can get the expected hash as follows:

from ethereum.crypto.hash import keccak256

...

call_return_data = int.to_bytes(0x0000111122223333444455556666777788889999AAAABBBBCCCCDDDDEEEEFFFF, 32, byteorder='big').ljust(call_return_size, b'\0')[0:call_return_size]

...
        storage={
            slot_code_worked: 1,
            slot_returndatasize_before_create: 32,
            slot_returndatasize_after_create: 0,
            slot_return_data_sha3: keccak256(call_return_data),
        }

[winsvega]

hm, the values does not match when returnsize is not 32. need to debug
ah I see. so we can ask the call to return into 0..x but the return datasize will always sys what was the actual return (0, y) inside call. this is another test need to see that return(0,y) inside the call does only write to allowed (0,x) when we called the contract

[marioevz]

I have not reviewed this file yet, but I think we need to split this file into its own PR to have it properly reviewed, since it's completely orthogonal to test implementation.

[winsvega]

Yes, but on this pr, I can test directly that it works with the given test changes.

It's not ready yet.