Build Information:
This build information is after the problem has been fixed using the included patch and other patches....
wireshark 1.7.1-SVN-40442 (SVN Rev Unknown from unknown)
Copyright 1998-2012 Gerald Combs <gerald@wireshark.org> and contributors.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
Compiled (64-bit) with GTK+ 3.2.3, with Cairo 1.10.2, with Pango 1.29.4, with
GLib 2.30.2, with libpcap (version unknown), with libz 1.2.3, without POSIX
capabilities, with SMI 0.4.8, with c-ares 1.7.5, with Lua 5.1, with Python
2.7.1, with GnuTLS 2.12.16, with Gcrypt 1.4.6, with Heimdal Kerberos, with
GeoIP, with PortAudio V19-devel (built Dec 9 2011 02:30:18), without AirPcap.
Running on OpenBSD 5.0, without locale, with OpenBSD libpcap, with libz 1.2.3,
GnuTLS 2.12.16, Gcrypt 1.4.6.
Built using gcc 4.2.1 20070719 .
OpenbSD 5.0 current amd64, when building both version 1.6.5 and SVN version 40442 with gtk3 configure option, fail during configure.
The check on the gtk+-3.0 version uses this in configure
Configure scripts are not written by hand, they're generated from configure.in by autoconf, so the patch can't be applied - there's nothing to which to apply it (the configure script isn't in SVN, and won't ever be in SVN, only the Wireshark source files from which it's generated, such as configure.in, acinclude.m4, etc., are).
This might be a consequence of the configure script being built on a system with GTK+ 2.x installed, or with an older version of it installed; the 2.x version of the configure macros might not understand that there's no GLib 3.x and that GTK+ 3.x works with GLib 2.
Configure scripts are not written by hand, they're generated from configure.in by autoconf, so the patch can't be applied - there's nothing to which to apply it (the configure script isn't in SVN, and won't ever be in SVN, only the Wireshark source files from which it's generated, such as configure.in, acinclude.m4, etc., are).
This might be a consequence of the configure script being built on a system with GTK+ 2.x installed, or with an older version of it installed; the 2.x version of the configure macros might not understand that there's no GLib 3.x and that GTK+ 3.x works with GLib 2.
Both wirehark 1.6.5, and the nightly SVN-40442 used the configure file that is in the source tarball, the configure script is used unaltered when building the port on OpenBSD, apart from applying a set of patches only where necessary. It is the configure.in generating configure for the source tarballs that is wrong.
configure.in has ----
if test "x$with_gtk3" = "xyes"; then AM_PATH_GTK_3_0(3.0.0, [ CFLAGS="$CFLAGS $GTK_CFLAGS" CXXFLAGS="$CXXFLAGS $GTK_CFLAGS" AC_DEFINE(HAVE_GTK, 1, [Define to 1 if compiling with GTK]) ], GTK_OK=no, gthread)
The modules, clearly are optional, so removing the gthread from AM_PATH_GTK_3_0 in configure.in should have the same effect as the patch to configure when used to generate configure file for the source tarballs.
aclocal-fallback/gtk-3.0.m4 has this
pkg_config_args=gtk+-3.0 for module in . $4 do case "$module" in gthread) pkg_config_args="$pkg_config_args gthread-2.0" ;; esac done
gthread is already checked later in configure.in as part of Glib-2.0, so doesn't need checking with gtk+-3.0.
[I'm not sure why my--and apparently everybody else's--pkg-config seems to accept --atleast-version 3.0.0 with gthread-2.0 version 2.something.]
OpenBSD now has its own non-GPLed pkg-config, written as a Perl script; I'm not sure when they did that, but if that was the case 11 years ago, that might be the issue.