be more tolerant of laggy networks;

client has a 5sec timeout before it retries the RRQ when
initiating a download

previously, if client eventually received a response for
both of the issued RRQ's, it would panic and disconnect

instead, ignore the duplicate ack/oack and keep going

partftpy/TftpStates.py

@@ -530,6 +530,10 @@ def handle(self, pkt, raddress, rport):
             self.sendError(TftpErrors.IllegalTftpOp)
             raise TftpException("Received ERR from peer: " + str(pkt))
 
+        elif isinstance(pkt, TftpPacketACK) or isinstance(pkt, TftpPacketOACK):
+            log.warning("Discarding unexpected packet type (retransmission?): %s", pkt)
+            return self
+
         else:
             self.sendError(TftpErrors.IllegalTftpOp)
             raise TftpException("Received unknown packet type from peer: " + str(pkt))