On Fri, Jan 17, 2014 at 12:34 PM, Warren Young <war...@etr-usa.com> wrote: > On 1/17/2014 11:58, Aaron Humphrey wrote: >> >> I had thought that cygwin 1.7 had IPv6 support, > > > It does, to the extent that the underlying Winsock APIs do. Basically, you > want to be on Vista or newer if you're going to depend on IPv6 under Cygwin. > IPv6 support for earlier versions of Windows was available only in > "experimental" form. (Microsoft's term.)
I am using Windows 7, so I was pretty sure that I had it. But I suppose that means that building it will still mean it only works on machines with Vista or later. That's probably okay. >> Is there an inherent limitation in IPv6 support under Cygwin > > Here's the relevant section of the configure output: > > checking whether to include IPv6 support... yes > checking for netinet/ip6.h... no > configure: WARNING: include file netinet/ip6.h not found, disabling IP6 > checking netinet6/in6.h usability... no > checking netinet6/in6.h presence... no > checking for netinet6/in6.h... no > So, apparently Cygwin needs several header files, or socat needs to be > tweaked to use the definitions in the files that already exist. I ran across a reference that implied that the regular netinet/ip.h files should work, so I gave that a try. I just disabled the check for netinet/ip6.h and forced the WITH_IP6 flag, and after that it built just fine. Probably I could accomplish the same thing by putting in an empty or stub ip6.h file. So why does it think it requires ip6.h if it compiles fine without it? Is Cygwin unusual in what it has in its header files, compared to other *n*x variants? It seems to work fine, too: starting up a simple IP4 to IP6 relay, my IP4 app can now connect to the IP6 server just fine. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
