[ENet-discuss] A few questions about unreliable delivery

[Lee Salzman]

Unreliable is sequenced, unless you specify the unsequenced flag.

On 07/28/2014 10:19 AM, Tom Spencer-Smith wrote:
>
> Thanks. So 0 flags means unreliable but sequenced.
> If I send U traffic on its own channel without any R on that channel, using 0 flag, is it still sequenced?
>
> Cheers
>
>
> On Mon, Jul 28, 2014 at 3:10 PM, Lee Salzman <lsalzman at gmail.com <mailto:lsalzman at gmail.com>> wrote:
>
>     1. 0 flags means unreliable.
>
>     2. Unsequenced are both unreliable and unsequenced, so they can be discarded and come out of order.
>
>     3. Neither has duplicates, so you will only ever get a packet once or never.
>
>     4. Unreliable packets sent after a reliable packet stall on the reliable packet, whereas unsequenced won't stall.
>
>
>     On 07/28/2014 06:27 AM, Tom Spencer-Smith wrote:
>
>
>         I'm a little confused on how to get certain delivery characteristics for unreliable packets.
>
>         1. By default (using 0 for the flags), U packets are delivered in order (but with possible gaps in sequence)?
>
>         2. ENET_PACKET_FLAG_UNSEQUENCED delivers them potentially out-of-order, i.e. discards nothing?
>
>         3. What about duplicates, for both cases above?
>
>         4. If R and U packets are sent on the same channel, do U packets stall on earlier undelivered R packets? So most people would want to use different channels for R and U?
>
>         Thanks
>
>         Tom
>
>
>         _______________________________________________
>         ENet-discuss mailing list
>         ENet-discuss at cubik.org <mailto:ENet-discuss at cubik.org>
>         http://lists.cubik.org/mailman/listinfo/enet-discuss
>
>
>     _______________________________________________
>     ENet-discuss mailing list
>     ENet-discuss at cubik.org <mailto:ENet-discuss at cubik.org>
>     http://lists.cubik.org/mailman/listinfo/enet-discuss
>
>
>
>
> _______________________________________________
> ENet-discuss mailing list
> ENet-discuss at cubik.org
> http://lists.cubik.org/mailman/listinfo/enet-discuss