eth_signjust went through the web3 library and didn't touch the node, but I was wrong about that.
eth_signseems fine. I mostly consider it deprecated because it's not a well-defined standard so it's got pretty sloppy adherence (geth, web3.js, and trezor all have slightly different implementations of message signing, last I checked, which affect some but not all inputs)
eth_signtests in web3.py passed, because IIRC it should be possible to send a message that's not hashed before adding that preamble. I commented on that PR you linked.
eth_sign. I just verified
eth_sign. I guess he uses the tool for testing both frontend and backend.
py-evmversions was easy turns out
hello. I am working with eth-tester and would like to use the py-evm backend. I wanted to write some eth-tester source code that would allow me to remove a transaction from the transaction list in a previous block and then recalculate the entire state from that block to the canonical head as if that transaction never existed. However i do not want it to mine the blocks again with new headers. It will just revert the state to the state at the start of that block, and apply transactions until the latest block and then just continue with that state.
to my understanding, eth-tester API is essentially a passthrough to functions in the py-evm backend. To my knowledge, there is no API in py-evm that allows me to remove a transaction from the transaction list.
I was wondering if I could store a db in eth-tester of all transactions in a block, and remove it from there, and then use EthereumTester.call(transaction, block_number='latest') on all transactions to achieve this result?
eth-testerby default uses the
eth.db.backends.memory.MemoryDB) backend for storage. You can change this by setting environment variable
eth.db.backends.level.LevelDB) backend. You’ll have to play around it a bit to get it to work with a pre-filled state.
levelDbcannot be faster than
levelDbwrites to disk.
eth-testeruse whatever the latest VM class is?