ports/174753: change to boost-lib causes adversary effects

[JINMEI Tatuya]

>Number:         174753
>Category:       ports
>Synopsis:       change to boost-lib causes adversary effects
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    freebsd-ports-bugs
>State:          open
>Quarter:        
>Keywords:       
>Date-Required:
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Thu Dec 27 22:20:00 UTC 2012
>Closed-Date:
>Last-Modified:
>Originator:     JINMEI Tatuya
>Release:        8.3-RELEASE
>Organization:
ISC
>Environment:
FreeBSD mon.jinmei.org 8.3-RELEASE FreeBSD 8.3-RELEASE #0: Sun Dec 16 19:53:01 PST 2012     root at newmon.jinmei.org:/usr/src/sys/i386/compile/PFHOST  i386
>Description:
The change to the boost-lib port introduced in
http://www.freebsd.org/cgi/query-pr.cgi?pr=144336
caused other regression.  For example, see this:
http://bind10.isc.org/ticket/1991

I confirmed this behavior with FreeBSD 8.3-RELEASE (i386) and
boost-libs-1.48.0_1 and g++ (GCC) 4.2.2 20070831 prerelease [FreeBSD].

It can more easily be confirmed by compiling this code:
#include <boost/numeric/conversion/cast.hpp>

#include <stdint.h>

int
main() {
    return (boost::numeric_cast<uint16_t>(0));
}

as follows: g++ -Werror -I/usr/local/include -c testcode.cc

It was also discussed with the boost developers:
https://svn.boost.org/trac/boost/ticket/4381
(see also https://svn.boost.org/trac/boost/ticket/4669)
and, as discussed there, it seems the change to the port is not the
right solution.

For that matter, I just confirmed I can successfully build
libtorrent-rasterbar-15-python even suppressing the corresponding
change in /usr/local/include/boost/config/compiler/gcc.hpp by hand.

So, I now propose reverting the change.  It seems it suffices if we
can simply remove
ports/devel/boost-libs/files/patch-boost__config__compiler__gcc.hpp

>How-To-Repeat:
See the description.
>Fix:
See the description.

>Release-Note:
>Audit-Trail:
>Unformatted: