(Hold off on merging this. I realized I need to also map the -Ex version to get v4 and v6 stats separately.)

Good to go.

Looks good to me. Thank you.

@dbwiddis could you please revisit the new unittests? They fail on appveyor. Here is a trace with added verbose info:

https://ci.appveyor.com/project/matthiasblaesing/jna/builds/32772146#L1580

My gut feeling: It is EstabResets, but _Curr_Estab. I.e. the former deals with a summed total, while the latter deals with the current (at this time) number. If that is so, the assert would indeed be invalid.

@matthiasblaesing Yes, you are correct. I did the Windows implementation first and made that assumption based on reading the docs -- and it didn't fail on my AppVeyor tests. Later on when implementing cross-platform code using similar fields on different OS's (sysctl on mac, netstat command on *nix) I removed that test.

I remember thinking, "I guess Windows interprets it differently." I suppose I was wrong and just got lucky with Appveyor. Do you want me to submit the fix?

@dbwiddis a fix would be great - the appveyor unittest runs are not perfect, but should be kept runnable none the less.