[ENet-discuss] (Plain-text) Rationale of ENET_PACKET_FLAG_UNRELIABLE_FRAGMENT?

[Lee Salzman]

The incidence of packets being dropped or malordered skyrockets once
packets start to fragment. It becomes the other way that the moment you
fragment your packets just stop getting through in a useful fashion. So for
better or worse I made the decision that packets by default force
reliability once they fragment to combat this. This was much later allowed
to be overridden by the unreliable fragment flag. Love it or hate it, this
is how the library behaves and the knobs now exist for you to change it
with flags. I have no plans to change it, sorry

On Fri, Jun 28, 2024, 16:29 Чёрный Думер <bdmailb at gmail.com> wrote:

> Good day everyone,
>
> Could somebody please explain to me the reasoning behind the
> ENET_PACKET_FLAG_UNRELIABLE_FRAGMENT flag?
>
> Why isn't this the default behavior for large unreliable packets
> exceeding the MTU?
>
> Do I understand correctly that without this flag, large unreliable
> packets become de facto reliable? In my opinion, this violates the
> principle of least astonishment, but I also think I'm probably missing
> something important.
>
> Thanks in advance.
>
> P.S. I apologize for sending this letter in MIME HTML format first time:
> http://lists.cubik.org/pipermail/enet-discuss/2024-June/002501.html
> Yandex.Mail sucks.
> _______________________________________________
> ENet-discuss mailing list
> ENet-discuss at puremagic.com
> https://lists.puremagic.com/mailman/listinfo/enet-discuss
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puremagic.com/pipermail/enet-discuss/attachments/20240628/9aaa704b/attachment.htm>