day14.md

Day 14

Today I upgraded my EVM setup by pulling the newest version of the evm-from-scratch repo. Some things have been refactored and some have been added aswell. So I spent a few minutes adjusting the new tests to my old code. During the process, I had to implement a couple Opcodes that were added in order to get back at my last result, namely GAS and LOG.

GAS is simple, it pushes the remaining available gas to the stack, but I haven't implemented a full gas-tracking system yet, so this is still a work in progress I would say.

LOG required some extra work: I had to add the logs field to the MachineState and return it to the test result. Here's how I implemented the actual LOG runners (in a single function for all LOG Opcodes):

function LOG(ms: MachineState) {
  // take the number of arguments from the opcode
  const n = ms.code[ms.pc] - 0xa0;

  // get the memory offset and size to read
  const [memOffset, size] = ms.stack.popN(2);

  // get the actual log topics from stack
  const topics = ms.stack.popN(n);

  // read the memory and parse the topics
  const data = ms.memory.read(Number(memOffset), Number(size));
  const topicsHex = topics.map(parsers.BigintIntoHexString);

  // push the log with the appropriate data and topics
  ms.logs.push({
    address: ms.txData.to,
    data: data.toString("hex"),
    topics: topicsHex,
  });
}

Also since today I had more time to dedicate, I implemented a big refactoring of the EVM class introducing a Logger, which is a way to visualize exactly what is happening on every execution step: stack, memory and storage dump.

I also added a way to start the EVM execution from a specified bytecode or an input file instead of only relying on provided tests, so I can potentially use it for more than just testing in the future... :D