Per our discussion in the trimalyzer thread, can you speculate why eehack isn't discarding trashed frames?

Am I correct to assume that's what's supposed to happen here in aldl_read_reply()?

Code:
  // look for bad checksum
  if(checksum_test() == false) {
    serial_error_occurred();
    pkt_fail_counter++;
    return 0;
  }
Here are three adjacent frames from a log taken this morning. Note how knock count "jumps" in the middle. This one is actually pretty subtle but I've found frames like this where VSS will be 255mph and there are three or four trouble codes, MAF reading 400g/s, etc.

frame-37168.png
frame-37169.png
frame-37170.png

This is by no means a huge problem, but I've been chasing my tail on false knock and the huge knock events these generate make it really hard to see the "real" false knock.

BTW, I have some patches I want to send you relevant to the knock scope that adds the timestamp to the tool tip bubble. Will try to push to git today.