At least in my compile environment, we don't need "inet_v6defs.h" to
authorguy <guy@f5534014-38df-0310-8fa8-9805f1628bb7>
Sat, 16 Nov 2002 21:36:39 +0000 (21:36 +0000)
committerguy <guy@f5534014-38df-0310-8fa8-9805f1628bb7>
Sat, 16 Nov 2002 21:36:39 +0000 (21:36 +0000)
commit367ff66c7099f910fba611764941b95028bbee4b
tree111df52574d40b35e719061398ec52c7e384a2a1
parentca072f82f7eb6bee22b4f0d4d1b4dba5ea6a30b7
At least in my compile environment, we don't need "inet_v6defs.h" to
compile "epan/packet.c", and including it without including <winsock2.h>
before it means that, as <winsock2.h> appears to get included after we
include "inet_v6defs.h", we get complaints about a redefinition of
AF_INET6; removing the include of "inet_v6defs.h" sqelches that warning.

If we *do* need "inet_v6defs.h" on some platforms, put it back, but put
in an include of <winsock2.h> before it, to keep the warning away - if
we ever use AF_INET6 in "epan/packet.c", we want to use the one from
<winsock2.h> if it defines it, as that's what the rest of Ethereal uses.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@6649 f5534014-38df-0310-8fa8-9805f1628bb7
epan/packet.c