Charles Lepple wrote: >>> I forget, is IPv6 enabled in the NUT ./configure script? >> It is, otherwise it wouldn't show 'setuptcp: try to bind to 0.0.0.0 port >> 3493' in the debug output (this line is only in the new IPv4/IPv6 code). > Sorry, my brain shuts down when I see functions longer than about one > page. I'm not a C preprocessor by trade :-)
Indeed setuptcp() is long than I'd prefer to see. As long as we're not sure that getaddrinfo() function is supported for all platforms we intend to support, we'll have to keep the old code in there as well. > Zoltan: in any case, the more details you can provide about the > system, the better. Arjen had some good questions about IPv4 vs IPv6 > and system hardening, and the combined truss+-DDD output will help us > figure out exactly which function is causing the problem. Even without truss, it would be worthwhile to see what is going on with the latest version from SVN. I expect that Zoltan's system is doing something really unusual here. That's why I added an additional check for an empty addrinfo list, although that should be covered by successful completion of the getaddrinfo() call (which is 'guaranteed' to have at least one element then). I suspect, in this case it hasn't, or none with a useful sockettype and protocol combination, which shouldn't happen either. Running the latest version with '-DDD' will unveil what is going on here. Best regards, Arjen _______________________________________________ Nut-upsuser mailing list Nut-upsuser@lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/nut-upsuser
