Improve feature detection in py-netif
Currently feature detection is made using freebsd numerical version (uname -K), that is fragile and hard to maintain. We would like to improve this possibly using autoconf so it can detect the features by trying to compile small tests softwares.
We risk breaking current feature detection in py-netif
Feature detection should be working in py-netif and we should have a more neat way of doing it instead of using uname -K.
( Different sections will be updated as the ticket progresses )