Mon Aug 23 17:52:51 CEST 2010
Way back in the thread starting with
http://www.erlang.org/cgi-bin/ezmlm-cgi/4/42722, I complained about "bad
behaviour" when some gen_tcp and gen_udp functions were passed an IPv6
8-tuple without the 'inet6' (or 'inet') option, suggesting (with
patches:-) that it would be more reasonable if IPv6 and IPv4 worked the
same - i.e. an 8-tuple would imply 'inet6'. Raimo made me happy with the
reply (in part):
> Not really. But as you see for gen_tcp:connect below, the inet6
> option is currently required there even for an IPv6 address tuple.
> These are corners we have not smoothed yet.
However the "bad behaviour" was fixed in R13B, not by auto-recognizing
the 8-tuple but by exit(badarg), and this remains the case in R14A.
Should I take this to mean that the auto-recognition won't happen? FWIW,
the second version of the patches in that thread has been working fine
in production for over a year - initially in R10B-10, later in R13B03 -
and seems to apply cleanly and work fine with R14A too.
More information about the erlang-questions