serial.receive() returns true repeatedly with no traffic

[simm42]

After a successful packet is received and HammingSerial.receive() is called, it will continue to return true for as long as no further traffic is on the underlying Serial device.

I beleive this is undesirable behaviour as the HammingSerial.receive() check is called to determine is there is a packet arrived to process. I think the bug is cuased by the following checks in HammingSerial.receive both being conditional on _serial.available() being true.

  if ((readState == PACKET_READ_STATE_END) && _serial->available()) {
    b = _serial->read();
    if (b == PACKET_START_SYMBOL) {
      readState = PACKET_READ_STATE_START;
    } else {
      readState = PACKET_READ_STATE_UNKNOWN;
      ++errorCount;
      rx_buffer.clear();
    }
  }
  
  while ((readState == PACKET_READ_STATE_UNKNOWN) && _serial->available()) {
    b = _serial->read();
    if (b == PACKET_START_SYMBOL) {
      readState = PACKET_READ_STATE_START;
    } 
  }

I beleive the fix would be to replace the above with:

  if (readState == PACKET_READ_STATE_END) {
    readState = PACKET_READ_STATE_UNKNOWN;
  }
  
  while ((readState == PACKET_READ_STATE_UNKNOWN) && _serial->available()) {
    b = _serial->read();
    if (b == PACKET_START_SYMBOL) {
      readState = PACKET_READ_STATE_START;
    }  else {      
      ++errorCount;
    }
  }

This saves the user from needing to check that both HammingSerial.receive() is true and that rx_buffer.count is non zero before being able to assume a new packet has been received.

[bkeevil]

Did you happen to check if the tests in the demo app run with this change?

[simm42]

I’m afraid I don’t have an environment set up to run them currently.

If I get chance to do so in next few days I’ll look at setting it up and adding the test case that this fixes.

[bkeevil]

I'll do it tonight.