Ensure we have both _initialize() and a corresponding _cleanup() routines for
authorJeff Morriss <jeff.morriss.ws@gmail.com>
Wed, 7 Aug 2013 16:46:08 +0000 (16:46 -0000)
committerJeff Morriss <jeff.morriss.ws@gmail.com>
Wed, 7 Aug 2013 16:46:08 +0000 (16:46 -0000)
commit2c81a97d55f4aee096933e6dd555efff300c2eec
treeba7f2fbbbe0e0b3966b9ccb90996dbfb00f688ce
parent2dab57f425b34026d772d2ba7df66ab4a21cbdc3
Ensure we have both _initialize() and a corresponding _cleanup() routines for
the various name resolvers; put those two routines next to each other.

Add generic addr_resolv_init() and addr_resolv_cleanup() routines which call
all of those internal routines.

Call the generic init/cleanup routine from epan_init() and epan_cleanup().

Create the hash tables for each name resolver in those initialization routines
in order to avoid having to repeatedly check if the table is already created
or not (and to avoid glib warnings if we neglected to perform that check):

http://www.wireshark.org/lists/wireshark-dev/201308/msg00012.html

Don't clean up   hostnames in init_dissection():    it's done already in cleanup_dissection().
Don't initialize hostnames in cleanup_dissection(): it's done already in init_dissection().

svn path=/trunk/; revision=51191
epan/addr_resolv.c
epan/addr_resolv.h
epan/epan.c
epan/packet.c