new(tests): Identity precompile test #1047
Conversation
|
@shemnon I was reviewing your memory expansion tests and remembered this case, so decided to add a test that can be filled for the |
| post = { | ||
| contract_address: Account( | ||
| storage={ | ||
| 1: keccak256(bytes([1, 2, 3, 4, 4]).ljust(32, b"\0")), |
There was a problem hiding this comment.
can this benefit from leftpadding?
| code = ( | ||
| sum(Op.MSTORE8(offset=i, value=(i + 1)) for i in range(4)) # memory = [1, 2, 3, 4] | ||
| + call_opcode( | ||
| address=4, |
There was a problem hiding this comment.
shall we have global space constants for the precompiles?
| """Test the return data of the identity precompile overwriting its input.""" | ||
| env = Environment() | ||
| code = ( | ||
| sum(Op.MSTORE8(offset=i, value=(i + 1)) for i in range(4)) # memory = [1, 2, 3, 4] |
There was a problem hiding this comment.
if possible I prefer to have concrete looking memory input like
0x000102030405 because when reading the code we will have to stop and calculate or sometimes we don't see what exactly the result of a calculation and when its not obvious we can have a bug and difficulty to find it.
|
Tests look fine WRT EXTCALL vs CALL. The only twist may be to test for the "corrupted" version in the memory write version when the opcodes are the EOF variants to verify no memory writes occured. |
🗒️ Description
Adds a test for the identity precompile based on this https://github.com/ethereum/go-ethereum/blob/master/docs/postmortems/2021-08-22-split-postmortem.md
Also adds a new variant that can be filled for EOF
EXT*CALLopcodes .🔗 Related Issues
✅ Checklist
mkdocs servelocally and verified the auto-generated docs for new tests in the Test Case Reference are correctly formatted.