Great, infra seems to have unborked the mailinglists (My stuff wasn't going through)
I think maybe the best option would be to create an abstraction of the real Socket implementation and to start with pcap4j and use that to implement the raw-socket Netty support. As soon as that's running, I think I can spare some extra cycles to try to implement something that's fully Apache 2.0 and doesn't have too big drawbacks. I think pcap4j does offer a huge load of functionality which we don't need and that might even decrease performance. Right now I'll stick to this option in order to continue implementing (Did waste quite a lot of time for trying to get RockSaw working) Chris Am 22.01.18, 01:13 schrieb "Justin Mclean" <[email protected]>: Hi, > Pcap4j is licensed with a MIT license: > https://github.com/kaitoy/pcap4j/#license Pcap4j is MIT with bit of BSD and Apache 2.0 all good. (And their logo is CC-BY-4.0 also ok) > So, I guess we can use that. As the library requires libpcap, the license of that is the 3-clause BSD: > http://www.tcpdump.org/license.html libpcap contains a number of licenses including MIT, BSD, dual licensed MIT/GPL (which is OK), some autogenerated GPL make scripts (also OK) and some BSD 4 clause licenses (not OK as that’s Category X). IMO there a good change we could get those changed to 3 clause BSD given they are from NetBSD which have retroactively changed the license. This has come up recently in another project (Mynewt) and there’s an outstanding question/JIRA on legal discuss about it. Thanks, Justin
