max_datagram_size() seems to be off by 2

[jfro]

Not sure if this should just be documented or if there should be a fix to the calculation.
If I try to send a packet via send_datagram of data at the size returned by max_datagram_size() I'll get a Toolarge error. I've had to reduce it by 2 to get it to pass the check & send the datagram.
I'm calling max_datagram_size() at time of data generation & immediate send_datagram() call so it should be current.

[Ralith]

Thanks for the report! We should really have a unit test for this.

[Ralith]

I'm not able to reproduce this on 0.8 or main. Can you provide a repro?

[jfro]

I'll try, but my first attempt at a repro for the assertion bug didn't yield the same problem the original code producing this issue is, haven't found what the difference between the 2 is so far.

[terrarier2111]

I had a similar issue where when i sent a packet with a size of max_datagram_size it wouldn't reach its destination, but when subtracting 1 or more from it, it worked

[Ralith]

That was #1413, which was fixed in #1415.

[Ralith]

Closing as not reproducible.