NetBurner Community Forum

There are 2 examples of receiving UDP data in the programmers manual.

One method is:
RegisterUDPFifo(port, &FIFO);

And the other is:
int UdpFd = CreateRxUdpSocket( port );

I am confused as to why I would use one over the other.
More importantly is the upkt.Validate() call.

if I use the later UDP listener, does it automatically take care of validation ?

For example, if I use:
int len = recvfrom( UdpFd, (BYTE *)buffer, 80, &SrcIpAddr, &LocalPort,
&SrcPort );

when it returns len > 0 has it already confirmed that the UDP packet checksum is valid ?

or if I want to make sure the checksum is valid, I would have to use the first example and call upkt.Validate() myself as in the example.

One of the cool things is that all the source is provided...
So the UDPSocket code lcan be checked if you look in nburn\system\udpsocket.cpp
you will see that at line 273 the receive routine is there.
If the packet fails validation it returns a zero for the size.

Paul

That clears the main part of my question, thank you.
But now, why would one chose one method over the other ?

The NetBurner system originally had the class based UDP system.
Yet the Unix/bsd/nromal sockets standard was the socket based connection.

People porting existing UDP apps to the netburner had problems because the interfaces were so different.

So we wrote the udpsocket code to provide an easier to use adapter.
The udpsocket stuff usues the udp classes underneath.

So the plain udp classes are slightly more efficient, but if you want to write code that is more portable to other
environments use the socket.

Paul