[ENet-discuss] Data loss on disconnect?

[Lee Salzman]

At current, the ENet user must ensure all data on the connection has 
been sent before he attempts a disconnect, if that is his wish.

Adam D. Moss wrote:
> Lee Salzman wrote:
> 
>> I had debated this and ultimately decided that disconnect should 
>> override any pending requests, and other behaviors can easily be 
>> emulated over it. You need to send a packet to the peer that marks the 
>> end of any communication, which the peer should acknowledge after it's 
>> processed all pending packets before it. After you receive the 
>> acknowledgement, then you do your disconnect.
> 
> 
> So if I understand you correctly, non-reliable-clobbering
> disconnection has to be implemented at the application protocol
> level?
> 
> Regards,
> --Adam
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> ENet-discuss mailing list
> ENet-discuss at cubik.org
> http://lists.cubik.org/mailman/listinfo/enet-discuss