[ENet-discuss] Question about RTT/PL measurement when peer is dead

[Martin Man]

Hi Lee,

thanks for your great work, since you (and others) touched the RTT measurements in preparation for 1.3.15, I wanted to ask a question.

- What I observed is that in scenario A and B are talking, RTT/PL gets updated fine.

- When B dies and stops responding to pings and reliable data, RTT and PL get stuck on their last values and never update until the peer times out. This way peer that will timeout in 15 seconds indicates that its RTT is 10ms and PL 0.1% (based on last successful communication).

Is my observation correct? Or am I missing something?

What can be done to improve this to potentially adjust RTT and PL somehow with every enet_host_service tick so that you can spot early there is something going on with the peer?

What I’m doing now is to compare lastReceiveTime of a peer with host’s serviceTime to see whether the peer starts to fail receiving ping replies but do not know whether this is a correct thing to do.

thanks for your thoughts,
Martin