Actually within ENet there is undocumented callback that get triggered when a 
reliable packet was succesfully received. Logically speaking,
ENet must know the packet has been received properly, otherwise ENet will 
resend the packet (that is requirement for reliable packets).




________________________________
 From: Benoit Germain <[email protected]>
To: Discussion of the ENet library <[email protected]> 
Sent: Wednesday, February 27, 2013 4:19 PM
Subject: Re: [ENet-discuss] enet 1.3.6 dll for .net
 




2013/2/26 James Bellinger <[email protected]>

With either referenceCount or freeCallback, how are you differentiating 
disconnection from acknowledge?
>For example, enet_peer_disconnect (Peer.Disconnect) calls 
>enet_peer_reset_queues.
>

I don't. It happens that I haven't encountered a situation where I care though, 
so I didn't think about it :-). 


>As a general thought:
>
>I wonder if it'd be useful to have a 'delivered count' on the packet itself.
>
>What is the specific use both of you get from knowing if a packet has been 
>delivered though?
>
I've not needed it myself, since if it's delivered, the protocol continues, and 
if it's not delivered, well,
>reliable delivery will eventually either deliver it or the remote party will 
>disconnect, so I don't have
>to pay attention to this.
>
>

I have a toy chat application where I can send data to several recipients. I 
use it to log the fact that all recipients did receive the text I sent. Nothing 
critical.


 
Benoit:
>
>I don't see ENet setting the packet freeCallback anywhere. I believe it's 
>unrelated to the malloc/free.
>

Yes you are right, I didn't check the code. Well then it means I can change my 
binding to use this instead of maintaining a list of packets for acknowledgment 
notification. (this packet callback didn't exist when I implemented this 
feature).
 
-- 
Benoit.

_______________________________________________
ENet-discuss mailing list
[email protected]
http://lists.cubik.org/mailman/listinfo/enet-discuss
_______________________________________________
ENet-discuss mailing list
[email protected]
http://lists.cubik.org/mailman/listinfo/enet-discuss

Reply via email to