[Bug c++/63262] New: internal compiler error including math.h

[Bug c++/63262] New: internal compiler error including math.h

[karo13de at googlemail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262

            Bug ID: 63262
           Summary: internal compiler error including math.h
           Product: gcc
           Version: 5.0
            Status: UNCONFIRMED
          Severity: blocker
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: karo13de at googlemail dot com

Compilation of  source crashes with
/usr/local/bin/g++ --version
g++ (GCC) 5.0.0 20140824 (experimental) ...

In file included from /usr/local/include/c++/5.0.0/cmath:44:0,
                 from ./CImg.h:77,
                 from gmic.h:61,
                 from gmic.cpp:1876:
/usr/local/lib/gcc/x86_64-apple-darwin13.3.0/5.0.0/include-fixed/math.h:273:5:
internal compiler error: Illegal instruction: 4
     return __inline_isfinitef(__x) && __builtin_fabsf(__x) >= __FLT_MIN__;
     ^

/usr/local/lib/gcc/x86_64-apple-darwin13.3.0/5.0.0/include-fixed/math.h:273:5:
internal compiler error: Abort trap: 6
g++: internal compiler error: Abort trap: 6 (program cc1plus)

gcc build from "hpc.sourceforge.net"

gmic software from "git clone http://git.code.sf.net/p/gmic/source gmic"

Offending command in gmic/src

/usr/local/bin/g++ -o gmic gmic.cpp -std=c++11 -Dgmic_build
-I/opt/local/include -Dgmic_is_beta=1 -Dcimg_display=1 -Dcimg_appname=\"gmic\"
-I/usr/X11R6/include  -Dcimg_use_zlib -Dcimg_use_png -Dcimg_use_jpeg
-Dcimg_use_tiff -Dcimg_use_openexr -I/opt/local/include/OpenEXR
-Dcimg_use_fftw3  -Wno-unused-parameter -Dgmic_is_parallel -Dcimg_use_opencv
-I/opt/local/include/opencv -I/opt/local/include   -Dgmic_is_beta=1  
-Dgmic_main -L/opt/local/lib -L/usr/X11R6/lib -lX11 -lpthread  -lz -lpng -lz
-ljpeg -ltiff -lIlmImf -lHalf -lfftw3 -lfftw3_threads  -Dcimg_use_opencv
-L/opt/local/lib -lopencv_calib3d -lopencv_contrib -lopencv_core
-lopencv_features2d -lopencv_flann -lopencv_gpu -lopencv_highgui
-lopencv_imgproc -lopencv_legacy -lopencv_ml -lopencv_nonfree
-lopencv_objdetect -lopencv_photo -lopencv_stitching -lopencv_superres
-lopencv_ts -lopencv_video -lopencv_videostab

OS Mac OS 10.9.4 (Mavericks) and 10.10 beta2 (Yosemite)

Compilation works with native g++ 
 g++ --version (Mavericks)
Configured with: --prefix=/Applications/Xcode.app/Contents/Developer/usr
--with-gxx-include-dir=/usr/include/c++/4.2.1
Apple LLVM version 5.1 (clang-503.0.40) (based on LLVM 3.4svn)
Target: x86_64-apple-darwin13.3.0
Thread model: posix
and Yosemite LLVM version 6.0  (based on LLVM 3.5svn)

[Bug c++/63262] internal compiler error including math.h

[fxcoudert at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262

Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |fxcoudert at gcc dot gnu.org

--- Comment #1 from Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> ---
Could you come up with a preprocessed source file showing the bug?

Rerun your command ("/usr/local/bin/g++ -o gmic gmic.cpp ...") adding the
-save-temps option. This should generate a gmic.ii file, which you can attach
here. It will allow us to reproduce the issue without having to download and
set up the full gmic source code.

[Bug c++/63262] internal compiler error including math.h

[karo13de at googlemail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262

--- Comment #2 from karsten <karo13de at googlemail dot com> ---
Created attachment 33646
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=33646&action=edit
compilation with "-save-temps" using the actual gmic git version


This ii file was generated with the gmic git version. Messages are attached:

/usr/local/bin/g++ -o gmic gmic.cpp -std=c++11 -Dgmic_build
-I/opt/local/include -Dgmic_is_beta=1 -Dcimg_display=1 -Dcimg_appname=\"gmic\"
-I/usr/X11R6/include  -Dcimg_use_zlib -Dcimg_use_png -Dcimg_use_jpeg
-Dcimg_use_tiff -Dcimg_use_openexr -I/opt/local/include/OpenEXR
-Dcimg_use_fftw3  -save-temps -mmacosx-version-min=10.7 -Wno-unused-parameter
-Wno-deprecated-register -Dcimg_use_opencv -I/opt/local/include/opencv
-I/opt/local/include   -Dgmic_is_beta=1  -Dgmic_main -L/opt/local/lib
-L/usr/X11R6/lib -lX11 -lpthread  -lz -lpng -lz -ljpeg -ltiff -lIlmImf -lHalf
-lfftw3 -lfftw3_threads  -L/opt/local/lib -lopencv_calib3d -lopencv_contrib
-lopencv_core -lopencv_features2d -lopencv_flann -lopencv_gpu -lopencv_highgui
-lopencv_imgproc -lopencv_legacy -lopencv_ml -lopencv_nonfree
-lopencv_objdetect -lopencv_photo -lopencv_stitching -lopencv_superres
-lopencv_ts -lopencv_video -lopencv_videostab 
In file included from /usr/local/include/c++/5.0.0/cmath:44:0,
                 from ./CImg.h:77,
                 from gmic.h:61,
                 from gmic.cpp:1876:
/usr/local/lib/gcc/x86_64-apple-darwin13.3.0/5.0.0/include-fixed/math.h:1:0:
internal compiler error: Illegal instruction: 4
 /*  DO NOT EDIT THIS FILE.
 ^

/usr/local/lib/gcc/x86_64-apple-darwin13.3.0/5.0.0/include-fixed/math.h:1:0:
internal compiler error: Abort trap: 6
g++: internal compiler error: Abort trap: 6 (program cc1plus)
-bash: line 64: 67127 Abort trap: 6           /usr/local/bin/g++ -o gmic
gmic.cpp -std=c++11 -Dgmic_build -I/opt/local/include -Dgmic_is_beta=1
-Dcimg_display=1 -Dcimg_appname=\"gmic\" -I/usr/X11R6/include -Dcimg_use_zlib
-Dcimg_use_png -Dcimg_use_jpeg -Dcimg_use_tiff -Dcimg_use_openexr
-I/opt/local/include/OpenEXR -Dcimg_use_fftw3 -save-temps
-mmacosx-version-min=10.7 -Wno-unused-parameter -Wno-deprecated-register
-Dcimg_use_opencv -I/opt/local/include/opencv -I/opt/local/include
-Dgmic_is_beta=1 -Dgmic_main -L/opt/local/lib -L/usr/X11R6/lib -lX11 -lpthread
-lz -lpng -lz -ljpeg -ltiff -lIlmImf -lHalf -lfftw3 -lfftw3_threads
-L/opt/local/lib -lopencv_calib3d -lopencv_contrib -lopencv_core
-lopencv_features2d -lopencv_flann -lopencv_gpu -lopencv_highgui
-lopencv_imgproc -lopencv_legacy -lopencv_ml -lopencv_nonfree
-lopencv_objdetect -lopencv_photo -lopencv_stitching -lopencv_superres
-lopencv_ts -lopencv_video -lopencv_videostab

[Bug c++/63262] internal compiler error including math.h

[fxcoudert at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262

Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|---                         |WORKSFORME

--- Comment #3 from Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> ---
OK, so the preprocessed source is truncated (due to the error, which happens
very early). However, I downloaded gmic and dependencies, and have been able to
compile it without issue (trunk revision 215887). I suggest the problem you are
having is either fixed, or was due to a building issue. You may want to contact
the hpc.sourceforge.net team for help.

In the mean time, I am closing this PR here. If you can reproduce the bug with
a recent compiler, please reopen and give full details.

[Bug c++/63262] internal compiler error including math.h

[karo13de at googlemail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262

karsten <karo13de at googlemail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |UNCONFIRMED
         Resolution|WORKSFORME                  |---

--- Comment #4 from karsten <karo13de at googlemail dot com> ---

After using the last built from latest development “snapshot” source for 4.9.2
from Oct 29th, according the maintainer of http://hpc.sourceforge.net/ the
error appears still!

System: Darwin ... 13.4.0 Darwin Kernel Version 13.4.0: Sun Aug 17 19:50:11 PDT
2014; root:xnu-2422.115.4~1/RELEASE_X86_64 x86_64

Source: gmic.souceforge.net git

Compiler command: 
/usr/local/bin/g++ -o gmic_gimp.o -c gmic.cpp -std=c++11 -mtune=native -g
-Dgmic_build -I/opt/local/include -Wall -W ${BETA} -Dcimg_use_fftw3
-Dcimg_use_png -Dcimg_use_zlib -Dcimg_display=1 -Dcimg_appname=\"gmic\"
-I/usr/X11R6/include -Dcimg_use_rng -Dgmic_gimp

Error: 
In file included from /usr/local/include/c++/4.9.2/cmath:44:0,
                 from ./CImg.h:77,
                 from gmic.h:61,
                 from gmic.cpp:1876:
/usr/local/lib/gcc/x86_64-apple-darwin14.0.0/4.9.2/include-fixed/math.h:273:5:
interner Compiler-Fehler: Illegal instruction: 4
     return __inline_isfinitef(__x) && __builtin_fabsf(__x) >= __FLT_MIN__;
     ^

/usr/local/lib/gcc/x86_64-apple-darwin14.0.0/4.9.2/include-fixed/math.h:273:5:
interner Compiler-Fehler: Abort trap: 6
g++: interner Compiler-Fehler: Abort trap: 6 (Programm cc1plus)
Abort trap: 6

gmic.ii is stored as zipped file under
https://drive.google.com/file/d/0B-rhVuF5w5sgTFllWEdfdzE3NnM/view?usp=sharing
since larger than 1MB.
>From gcc-bugs-return-465580-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Sun Nov 02 21:41:02 2014
Return-Path: <gcc-bugs-return-465580-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 2990 invoked by alias); 2 Nov 2014 21:41:02 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 2974 invoked by uid 48); 2 Nov 2014 21:40:59 -0000
From: "cordlandwehr at kde dot org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/63716] GCC Internal Error
Date: Sun, 02 Nov 2014 21:41:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: c++
X-Bugzilla-Version: 4.9.1
X-Bugzilla-Keywords: memory-hog
X-Bugzilla-Severity: normal
X-Bugzilla-Who: cordlandwehr at kde dot org
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields:
Message-ID: <bug-63716-4-9j1J2IvTqw@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-63716-4@http.gcc.gnu.org/bugzilla/>
References: <bug-63716-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2014-11/txt/msg00052.txt.bz2
Content-length: 137

https://gcc.gnu.org/bugzilla/show_bug.cgi?idc716

--- Comment #2 from cordlandwehr at kde dot org ---
* RAM: 12 GB
* Swap space: 4 GB

[Bug c++/63262] internal compiler error including math.h

[fxcoudert at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262

--- Comment #5 from Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> ---
(In reply to karsten from comment #4)
> After using the last built from latest development “snapshot” source for
> 4.9.2 from Oct 29th, according the maintainer of http://hpc.sourceforge.net/
> the error appears still!

Can you tell us:

 - what GCC source file was used exactly (link to tarball, or revision number)
 - whether any modification (patch) was applied to the GCC source
 - what build procedure they used (GMP, MPFR, MPC versions; what "configure"
command-line; etc.)
 - what OS X version the compiler was configured on (not your machine, the one
the hpc.sourceforge.net people used)

Only my machine, with the closest build I have right now (somewhat older 4.9
branch), I don't see the bug. If I get the information above, I can try to
reproduce it, and figure it out.

Also, it might be good for the hpc.sourceforge.net team to come here and give
their input directly.
>From gcc-bugs-return-465582-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Sun Nov 02 21:59:55 2014
Return-Path: <gcc-bugs-return-465582-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 17155 invoked by alias); 2 Nov 2014 21:59:55 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 17130 invoked by uid 55); 2 Nov 2014 21:59:52 -0000
From: "karo13de at googlemail dot com" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/63262] internal compiler error including math.h
Date: Sun, 02 Nov 2014 21:59:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: c++
X-Bugzilla-Version: 5.0
X-Bugzilla-Keywords:
X-Bugzilla-Severity: blocker
X-Bugzilla-Who: karo13de at googlemail dot com
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields:
Message-ID: <bug-63262-4-QSy4BBDxcp@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-63262-4@http.gcc.gnu.org/bugzilla/>
References: <bug-63262-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2014-11/txt/msg00054.txt.bz2
Content-length: 1552

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262

--- Comment #6 from karsten <karo13de at googlemail dot com> ---
I don't have the requested information. I have forwarded your comment to the
maintainer of hpc Gaurav Khanna gkhanna@umassd.edu

Karsten

Am 02.11.2014 um 22:46 schrieb fxcoudert at gcc dot gnu.org
<gcc-bugzilla@gcc.gnu.org>:

> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262
> 
> --- Comment #5 from Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> ---
> (In reply to karsten from comment #4)
>> After using the last built from latest development “snapshot” source for
>> 4.9.2 from Oct 29th, according the maintainer of http://hpc.sourceforge.net/
>> the error appears still!
> 
> Can you tell us:
> 
> - what GCC source file was used exactly (link to tarball, or revision number)
> - whether any modification (patch) was applied to the GCC source
> - what build procedure they used (GMP, MPFR, MPC versions; what "configure"
> command-line; etc.)
> - what OS X version the compiler was configured on (not your machine, the one
> the hpc.sourceforge.net people used)
> 
> Only my machine, with the closest build I have right now (somewhat older 4.9
> branch), I don't see the bug. If I get the information above, I can try to
> reproduce it, and figure it out.
> 
> Also, it might be good for the hpc.sourceforge.net team to come here and give
> their input directly.
> 
> -- 
> You are receiving this mail because:
> You reported the bug.

Karsten
karo13de@googlemail.com
>From gcc-bugs-return-465583-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Sun Nov 02 22:09:59 2014
Return-Path: <gcc-bugs-return-465583-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 20050 invoked by alias); 2 Nov 2014 22:09:59 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 20013 invoked by uid 55); 2 Nov 2014 22:09:55 -0000
From: "karo13de at googlemail dot com" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/63262] [4.9.2] ICE including math.h on darwin
Date: Sun, 02 Nov 2014 22:09:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: c++
X-Bugzilla-Version: 4.9.2
X-Bugzilla-Keywords: ice-on-valid-code
X-Bugzilla-Severity: blocker
X-Bugzilla-Who: karo13de at googlemail dot com
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields:
Message-ID: <bug-63262-4-26Z7gm9uc5@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-63262-4@http.gcc.gnu.org/bugzilla/>
References: <bug-63262-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2014-11/txt/msg00055.txt.bz2
Content-length: 1060

https://gcc.gnu.org/bugzilla/show_bug.cgi?idc262

--- Comment #7 from karsten <karo13de at googlemail dot com> ---
I have tried the same with the hpc build of gcc 5.0 on Yosemite MacOSX 10.10
with the same result!

Should I produce a similar gmic.ii file ?

Karsten


Am 02.11.2014 um 23:03 schrieb fxcoudert at gcc dot gnu.org
<gcc-bugzilla@gcc.gnu.org>:

> https://gcc.gnu.org/bugzilla/show_bug.cgi?idc262
>
> Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> changed:
>
>           What    |Removed                     |Added
> ----------------------------------------------------------------------------
>           Keywords|                            |ice-on-valid-code
>             Target|                            |x86_64-apple-darwin13
>            Version|5.0                         |4.9.2
>            Summary|internal compiler error     |[4.9.2] ICE including
>                   |including math.h            |math.h on darwin
>
> --
> You are receiving this mail because:
> You reported the bug.

Karsten
karo13de@googlemail.com

[Bug c++/63262] [4.9.2] ICE including math.h on darwin

[fxcoudert at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262

--- Comment #8 from Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> ---
(In reply to karsten from comment #7)
> I have tried the same with the hpc build of gcc 5.0 on Yosemite MacOSX 10.10
> with the same result!
> Should I produce a similar gmic.ii file ?

You can, and I'll test it. But I strongly suspect this is an issue with the
hpc.sf.net build, because my own builds of trunk do not seem affected by this
issue (see my comment #3).

[Bug c++/63262] [4.9.2] ICE including math.h on darwin

[karo13de at googlemail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262

--- Comment #9 from karsten <karo13de at googlemail dot com> ---
Anfang der weitergeleiteten Nachricht:

> Von: Gaurav Khanna <gkhanna@umassd.edu>
> Betreff: Aw: [Bug c++/63262] internal compiler error including math.h
> Datum: 2. November 2014 23:15:50 MEZ
> An: Karsten <karo13de@googlemail.com>
> 
> 
> http://mirrors.concertpass.com/gcc/snapshots/LATEST-4.9/gcc-4.9-20141029.tar.bz2
> 
> No modifications
> 
> ./configure —enable-languages=c++,fortran
> GMP 6.0.0a, MPFR 3.1.2 (w/ cumulative patch), MPC 1.0.2
> OS X Yosemite (current) on (late 2013) MacPro
> 
> 
> ---------------------------------------------------------------------------------------
> GAURAV KHANNA, (508) 910 6605, http://gravity.phy.umassd.edu
> Associate Professor, Physics Department, College of Engineering
> Assoc. Director, Center for Scientific Computing & Visualization Research
> Graduate Program Director, Engg & Appl. Sci. Ph.D. Program 
> University of Massachusetts Dartmouth
> 
> "Black holes are where God divided by zero." - Steven Wright
> 
> On Nov 2, 2014, at 4:53 PM, Karsten <karo13de@googlemail.com> wrote:
> 
>> 
>> Hi Gaurav,
>> 
>> thank you for your response. I have tried several options without improvement.
>> 
>> I have reopend the ticket and one of the maintainers asks for some more information concerning the build of gcc. It would be very helpful if you could communicate the questions. 
>> 
>> Thanks in advance
>> 
>> Karsten
>> 
>> Anfang der weitergeleiteten Nachricht:
>> 
>>> Von: "fxcoudert at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
>>> Betreff: [Bug c++/63262] internal compiler error including math.h
>>> Datum: 2. November 2014 22:46:28 MEZ
>>> An: karo13de@googlemail.com
>>> 
>>> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262
>>> 
>>> --- Comment #5 from Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> ---
>>> (In reply to karsten from comment #4)
>>>> After using the last built from latest development “snapshot” source for
>>>> 4.9.2 from Oct 29th, according the maintainer of http://hpc.sourceforge.net/
>>>> the error appears still!
>>> 
>>> Can you tell us:
>>> 
>>> - what GCC source file was used exactly (link to tarball, or revision number)
>>> - whether any modification (patch) was applied to the GCC source
>>> - what build procedure they used (GMP, MPFR, MPC versions; what "configure"
>>> command-line; etc.)
>>> - what OS X version the compiler was configured on (not your machine, the one
>>> the hpc.sourceforge.net people used)
>>> 
>>> Only my machine, with the closest build I have right now (somewhat older 4.9
>>> branch), I don't see the bug. If I get the information above, I can try to
>>> reproduce it, and figure it out.
>>> 
>>> Also, it might be good for the hpc.sourceforge.net team to come here and give
>>> their input directly.
>>> 
>>> -- 
>>> You are receiving this mail because:
>>> You reported the bug.
>> 
>> Karsten
>> karo13de@googlemail.com
>> 
>> 
>> 
> 

Karsten
karo13de@googlemail.com
>From gcc-bugs-return-465586-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Sun Nov 02 22:20:46 2014
Return-Path: <gcc-bugs-return-465586-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 25494 invoked by alias); 2 Nov 2014 22:20:45 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 25467 invoked by uid 48); 2 Nov 2014 22:20:42 -0000
From: "fxcoudert at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/63262] [4.9.2] ICE including math.h on darwin
Date: Sun, 02 Nov 2014 22:20:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: c++
X-Bugzilla-Version: 4.9.2
X-Bugzilla-Keywords: ice-on-valid-code
X-Bugzilla-Severity: blocker
X-Bugzilla-Who: fxcoudert at gcc dot gnu.org
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields:
Message-ID: <bug-63262-4-1tTNbZUPfl@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-63262-4@http.gcc.gnu.org/bugzilla/>
References: <bug-63262-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2014-11/txt/msg00058.txt.bz2
Content-length: 449

https://gcc.gnu.org/bugzilla/show_bug.cgi?idc262

--- Comment #10 from Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> ---
Could you backtrace the segfault?

1. Run g++ with added option "-v", and find out the command-line with "cc1plus"
in the output.
2. Then run this "cc1plus" command-line under a debugger, prefixing your
command line with "lldb -- " or "gdb -args "
3. When it segfaults, get a backtrace through debugger command "bt"

[Bug c++/63262] [4.9.2] ICE including math.h on darwin

[karo13de at googlemail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262

--- Comment #11 from karsten <karo13de at googlemail dot com> ---
I could not forward the mail from Gaurav:

Von: Gaurav Khanna <gkhanna@umassd.edu>
Betreff: Aw: [Bug c++/63262] internal compiler error including math.h
Datum: 2. November 2014 23:15:50 MEZ
An: Karsten <karo13de@googlemail.com>


http://mirrors.concertpass.com/gcc/snapshots/LATEST-4.9/gcc-4.9-20141029.tar.bz2

No modifications

./configure �enable-languages=c++,fortran
GMP 6.0.0a, MPFR 3.1.2 (w/ cumulative patch), MPC 1.0.2
OS X Yosemite (current) on (late 2013) MacPro


---------------------------------------------------------------------------------------
GAURAV KHANNA, (508) 910 6605, http://gravity.phy.umassd.edu
Associate Professor, Physics Department, College of Engineering
Assoc. Director, Center for Scientific Computing & Visualization Research
Graduate Program Director, Engg & Appl. Sci. Ph.D. Program 
University of Massachusetts Dartmouth

"Black holes are where God divided by zero." - Steven Wright

Karsten
karo13de@googlemail.com
>From gcc-bugs-return-465588-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Sun Nov 02 22:25:38 2014
Return-Path: <gcc-bugs-return-465588-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 29983 invoked by alias); 2 Nov 2014 22:25:38 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 29960 invoked by uid 55); 2 Nov 2014 22:25:35 -0000
From: "karo13de at googlemail dot com" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/63262] [4.9.2] ICE including math.h on darwin
Date: Sun, 02 Nov 2014 22:25:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: c++
X-Bugzilla-Version: 4.9.2
X-Bugzilla-Keywords: ice-on-valid-code
X-Bugzilla-Severity: blocker
X-Bugzilla-Who: karo13de at googlemail dot com
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields:
Message-ID: <bug-63262-4-dpgnhaxXcX@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-63262-4@http.gcc.gnu.org/bugzilla/>
References: <bug-63262-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2014-11/txt/msg00060.txt.bz2
Content-length: 1158

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262

--- Comment #12 from karsten <karo13de at googlemail dot com> ---
> I could not forward the mail from Gaurav:
> 
> Von: Gaurav Khanna <gkhanna@umassd.edu>
> Betreff: Aw: [Bug c++/63262] internal compiler error including math.h
> Datum: 2. November 2014 23:15:50 MEZ
> An: Karsten <karo13de@googlemail.com>
> 
> 
> http://mirrors.concertpass.com/gcc/snapshots/LATEST-4.9/gcc-4.9-20141029.tar.bz2
> 
> No modifications
> 
> ./configure �enable-languages=c++,fortran
> GMP 6.0.0a, MPFR 3.1.2 (w/ cumulative patch), MPC 1.0.2
> OS X Yosemite (current) on (late 2013) MacPro
> 
> 
> ---------------------------------------------------------------------------------------
> GAURAV KHANNA, (508) 910 6605, http://gravity.phy.umassd.edu
> Associate Professor, Physics Department, College of Engineering
> Assoc. Director, Center for Scientific Computing & Visualization Research
> Graduate Program Director, Engg & Appl. Sci. Ph.D. Program 
> University of Massachusetts Dartmouth
> 
> "Black holes are where God divided by zero." - Steven Wright

Karsten
karo13de@googlemail.com
>From gcc-bugs-return-465589-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Sun Nov 02 22:34:30 2014
Return-Path: <gcc-bugs-return-465589-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 569 invoked by alias); 2 Nov 2014 22:34:29 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 540 invoked by uid 55); 2 Nov 2014 22:34:25 -0000
From: "karo13de at googlemail dot com" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/63262] [4.9.2] ICE including math.h on darwin
Date: Sun, 02 Nov 2014 22:34:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: c++
X-Bugzilla-Version: 4.9.2
X-Bugzilla-Keywords: ice-on-valid-code
X-Bugzilla-Severity: blocker
X-Bugzilla-Who: karo13de at googlemail dot com
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields:
Message-ID: <bug-63262-4-oSzT5lR9yX@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-63262-4@http.gcc.gnu.org/bugzilla/>
References: <bug-63262-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2014-11/txt/msg00061.txt.bz2
Content-length: 5706

https://gcc.gnu.org/bugzilla/show_bug.cgi?idc262

--- Comment #13 from karsten <karo13de at googlemail dot com> ---
It gets late, here is the output of lldb. I will be busy tomorrow so I will not
respond to quick!

lldb -- /usr/local/libexec/gcc/x86_64-apple-darwin14.0.0/4.9.2/cc1plus -quiet
-v -I /opt/local/include -I /usr/X11R6/include -D__DYNAMIC__ -D gmic_build -D
gmic_is_beta=1 -D cimg_use_fftw3 -D cimg_use_png -D cimg_use_zlib -D
cimg_display=1 -D cimg_appname="gmic" -D cimg_use_rng -D gmic_gimp gmic.cpp
--param l1-cache-size2 --param l1-cache-line-sized --param
l2-cache-sizea44 -mtune=core2 -fPIC -feliminate-unused-debug-symbols -quiet
-dumpbase gmic.cpp -mmacosx-version-min\x10.9.4 -auxbase-strip gmic_gimp.o -g
-Wall -Wextra -std=c++11 -version -o
/var/folders/gs/rk3h54tr8xl48f008k8wf9n80000gn/T//ccymxQht.s
(lldb) target create
"/usr/local/libexec/gcc/x86_64-apple-darwin14.0.0/4.9.2/cc1plus"
Current executable set to
'/usr/local/libexec/gcc/x86_64-apple-darwin14.0.0/4.9.2/cc1plus' (x86_64).
(lldb) settings set -- target.run-args  "-quiet" "-v" "-I" "/opt/local/include"
"-I" "/usr/X11R6/include" "-D__DYNAMIC__" "-D" "gmic_build" "-D"
"gmic_is_beta=1" "-D" "cimg_use_fftw3" "-D" "cimg_use_png" "-D" "cimg_use_zlib"
"-D" "cimg_display=1" "-D" "cimg_appname=gmic" "-D" "cimg_use_rng" "-D"
"gmic_gimp" "gmic.cpp" "--param" "l1-cache-size2" "--param"
"l1-cache-line-sized" "--param" "l2-cache-sizea44" "-mtune=core2" "-fPIC"
"-feliminate-unused-debug-symbols" "-quiet" "-dumpbase" "gmic.cpp"
"-mmacosx-version-min\x10.9.4" "-auxbase-strip" "gmic_gimp.o" "-g" "-Wall"
"-Wextra" "-std=c++11" "-version" "-o"
"/var/folders/gs/rk3h54tr8xl48f008k8wf9n80000gn/T//ccymxQht.s"
run
(lldb) run
Process 1864 launched:
'/usr/local/libexec/gcc/x86_64-apple-darwin14.0.0/4.9.2/cc1plus' (x86_64)
GNU C++ (GCC) version 4.9.2 20141029 (prerelease) (x86_64-apple-darwin14.0.0)

    compiled by GNU C version 4.9.2 20141029 (prerelease), GMP version 6.0.0,
MPFR version 3.1.2-p10, MPC version 1.0.2

GGC heuristics: --param ggc-min-expand\x100 --param ggc-min-heapsize\x131072

ignoring nonexistent directory
"/usr/local/lib/gcc/x86_64-apple-darwin14.0.0/4.9.2/../../../../x86_64-apple-darwin14.0.0/include"

#include "..." search starts here:

#include <...> search starts here:

 /opt/local/include

 /usr/X11R6/include


/usr/local/lib/gcc/x86_64-apple-darwin14.0.0/4.9.2/../../../../include/c++/4.9.2


/usr/local/lib/gcc/x86_64-apple-darwin14.0.0/4.9.2/../../../../include/c++/4.9.2/x86_64-apple-darwin14.0.0


/usr/local/lib/gcc/x86_64-apple-darwin14.0.0/4.9.2/../../../../include/c++/4.9.2/backward

 /usr/local/lib/gcc/x86_64-apple-darwin14.0.0/4.9.2/include

 /usr/local/include

 /usr/local/lib/gcc/x86_64-apple-darwin14.0.0/4.9.2/include-fixed

 /usr/include

 /System/Library/Frameworks

 /Library/Frameworks

End of search list.

GNU C++ (GCC) version 4.9.2 20141029 (prerelease) (x86_64-apple-darwin14.0.0)

    compiled by GNU C version 4.9.2 20141029 (prerelease), GMP version 6.0.0,
MPFR version 3.1.2-p10, MPC version 1.0.2

GGC heuristics: --param ggc-min-expand\x100 --param ggc-min-heapsize\x131072

Compiler executable checksum: 10c0f5a43ac831149a38356dc3f34f45

Process 1864 stopped
* thread #1: tid = 0x1329d, 0x00000001415c747a libgmp.10.dylib`__gmpn_popcount
+ 26, queue = 'com.apple.main-thread', stop reason = EXC_BAD_INSTRUCTION
(code=EXC_I386_INVOP, subcode=0x0)
    frame #0: 0x00000001415c747a libgmp.10.dylib`__gmpn_popcount + 26
libgmp.10.dylib`__gmpn_popcount + 26:
-> 0x1415c747a:  popcntq (%rdi,%rcx,8), %r8
   0x1415c7480:  popcntq 0x8(%rdi,%rcx,8), %r9
   0x1415c7487:  jmp    0x1415c7516               ; __gmpn_popcount + 182
   0x1415c748c:  movl   $0x2, %ecx
bt
(lldb) bt
* thread #1: tid = 0x1329d, 0x00000001415c747a libgmp.10.dylib`__gmpn_popcount
+ 26, queue = 'com.apple.main-thread', stop reason = EXC_BAD_INSTRUCTION
(code=EXC_I386_INVOP, subcode=0x0)
  * frame #0: 0x00000001415c747a libgmp.10.dylib`__gmpn_popcount + 26
    frame #1: 0x0000000141578f67 libmpfr.4.dylib`parsed_string_to_mpfr + 2567
    frame #2: 0x000000014157847b libmpfr.4.dylib`mpfr_strtofr + 2011
    frame #3: 0x00000001006f4d11 cc1plus`real_from_string(real_value*, char
const*) + 817
    frame #4: 0x00000001006f5741 cc1plus`real_from_string3(real_value*, char
const*, machine_mode) + 33
    frame #5: 0x000000010015f7b2 cc1plus`interpret_float(cpp_token const*,
unsigned int, char const*, overflow_type*) + 354
    frame #6: 0x000000010016059c cc1plus`c_lex_with_flags(tree_node**, unsigned
int*, unsigned char*, int) + 716
    frame #7: 0x000000010007dba0
cc1plus`cp_lexer_get_preprocessor_token(cp_lexer*, cp_token*) + 32
    frame #8: 0x00000001000b0de9 cc1plus`c_parse_file() + 265
    frame #9: 0x0000000100165c3b cc1plus`c_common_parse_file() + 123
    frame #10: 0x000000010078741e cc1plus`compile_file() + 46
    frame #11: 0x000000010078936a cc1plus`toplev_main(int, char**) + 3002
    frame #12: 0x00007fff8d8755fd libdyld.dylib`start + 1
    frame #13: 0x00007fff8d8755fd libdyld.dylib`start + 1


Am 02.11.2014 um 23:20 schrieb fxcoudert at gcc dot gnu.org
<gcc-bugzilla@gcc.gnu.org>:

> https://gcc.gnu.org/bugzilla/show_bug.cgi?idc262
>
> --- Comment #10 from Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> ---
> Could you backtrace the segfault?
>
> 1. Run g++ with added option "-v", and find out the command-line with "cc1plus"
> in the output.
> 2. Then run this "cc1plus" command-line under a debugger, prefixing your
> command line with "lldb -- " or "gdb -args "
> 3. When it segfaults, get a backtrace through debugger command "bt"
>
> --
> You are receiving this mail because:
> You reported the bug.

Karsten
karo13de@googlemail.com

[Bug c++/63262] [4.9.2] ICE including math.h on darwin

[fxcoudert at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262

--- Comment #14 from Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> ---
(In reply to karsten from comment #13)
> libgmp.10.dylib`__gmpn_popcount + 26, queue = 'com.apple.main-thread', stop
> reason = EXC_BAD_INSTRUCTION (code=EXC_I386_INVOP, subcode=0x0)
>   * frame #0: 0x00000001415c747a libgmp.10.dylib`__gmpn_popcount + 26
>     frame #1: 0x0000000141578f67 libmpfr.4.dylib`parsed_string_to_mpfr + 2567
>     frame #2: 0x000000014157847b libmpfr.4.dylib`mpfr_strtofr + 2011
>     frame #3: 0x00000001006f4d11 cc1plus`real_from_string(real_value*, char
> const*) + 817
>     frame #4: 0x00000001006f5741 cc1plus`real_from_string3(real_value*, char
> const*, machine_mode) + 33

So the segfault happens in GMP. I also confirm that I don't see it after doing
a clean build of the 4.9.2 sources you linked, which makes sense because I do
my builds with an older GMP version.

What I would advise is building GMP without too much optimization, or with no
optimization at all, and see if it fixes your issue.

PS: the attached short testcase might be useful. Could you confirm that it does
trigger the ICE (internal compiler error)?

[Bug c++/63262] [4.9.2] ICE including math.h on darwin

[fxcoudert at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262

--- Comment #15 from Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> ---
Created attachment 33869
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=33869&action=edit
Shorted testcase should reproduce the issue

[Bug c++/63262] [4.9.2] ICE including math.h on darwin

[fxcoudert at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262

Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |WAITING
   Last reconfirmed|                            |2014-11-02
     Ever confirmed|0                           |1
           Severity|blocker                     |normal

[Bug c++/63262] [4.9.2] ICE including math.h on darwin

[karo13de at googlemail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262

--- Comment #16 from karsten <karo13de at googlemail dot com> ---
Hmm, hmm I am a packager in this case and don't know how to build that library.
It is delivered with the compiler from hpc.
What I could try is to replace it with install_name_tool with my MacPorts
library libgmp.10.dylib. But that will last some time.

By the way, I would appreciate a nice complete test case.

Good night,

Karsten

otool -L /usr/local/libexec/gcc/x86_64-apple-darwin14.0.0/4.9.2/cc1plus
/usr/local/libexec/gcc/x86_64-apple-darwin14.0.0/4.9.2/cc1plus:
    /usr/lib/libiconv.2.dylib (compatibility version 7.0.0, current version
7.0.0)
    /usr/local/lib/libmpc.3.dylib (compatibility version 4.0.0, current version
4.0.0)
    /usr/local/lib/libmpfr.4.dylib (compatibility version 6.0.0, current
version 6.2.0)
    /usr/local/lib/libgmp.10.dylib (compatibility version 13.0.0, current
version 13.0.0)
    /usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version
1213.0.0)

Am 02.11.2014 um 23:51 schrieb fxcoudert at gcc dot gnu.org
<gcc-bugzilla@gcc.gnu.org>:

> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262
> 
> --- Comment #14 from Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> ---
> (In reply to karsten from comment #13)
>> libgmp.10.dylib`__gmpn_popcount + 26, queue = 'com.apple.main-thread', stop
>> reason = EXC_BAD_INSTRUCTION (code=EXC_I386_INVOP, subcode=0x0)
>>  * frame #0: 0x00000001415c747a libgmp.10.dylib`__gmpn_popcount + 26
>>    frame #1: 0x0000000141578f67 libmpfr.4.dylib`parsed_string_to_mpfr + 2567
>>    frame #2: 0x000000014157847b libmpfr.4.dylib`mpfr_strtofr + 2011
>>    frame #3: 0x00000001006f4d11 cc1plus`real_from_string(real_value*, char
>> const*) + 817
>>    frame #4: 0x00000001006f5741 cc1plus`real_from_string3(real_value*, char
>> const*, machine_mode) + 33
> 
> So the segfault happens in GMP. I also confirm that I don't see it after doing
> a clean build of the 4.9.2 sources you linked, which makes sense because I do
> my builds with an older GMP version.
> 
> What I would advise is building GMP without too much optimization, or with no
> optimization at all, and see if it fixes your issue.
> 
> PS: the attached short testcase might be useful. Could you confirm that it does
> trigger the ICE (internal compiler error)?
> 
> -- 
> You are receiving this mail because:
> You reported the bug.

Karsten
karo13de@googlemail.com

[Bug c++/63262] [4.9.2] ICE including math.h on darwin

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org

--- Comment #17 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
The existence of the popcntq instruction suggests your libgmp has been
configured for Nehalem or later CPU, which you apparently don't have.
So, either convince whoever built it to configure it for earlier CPUs, or build
it on your own.  In any case, this doesn't sound like a GCC bug in any way.

[Bug c++/63262] [4.9.2] ICE including math.h on darwin

[fxcoudert at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262

Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |RESOLVED
         Resolution|---                         |INVALID

--- Comment #18 from Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> ---
(In reply to Jakub Jelinek from comment #17)
> The existence of the popcntq instruction suggests your libgmp has been
> configured for Nehalem or later CPU, which you apparently don't have.
> So, either convince whoever built it to configure it for earlier CPUs, or
> build it on your own.  In any case, this doesn't sound like a GCC bug in any
> way.

Thanks Jakub, I had missed that. It's actually much simpler. Karsten, you need
to ask Gaurav to build his GMP/MPFR/MPC libraries with an explicit build
parameter:

./configure --build=x86_64-apple-darwin14.0.0

(for Yosemite). Otherwise, the GMP library is not portable. And if I recall
correctly, MPFR and MPC use the GMP settings for their own compilation.

[Bug c++/63262] [4.9.2] ICE including math.h on darwin

[karo13de at googlemail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262

--- Comment #19 from karsten <karo13de at googlemail dot com> ---
Merci François for your efforts.

I have replaced the resp. library by my MacPorts build and it compiles now!
Perhaps a hacker way, however. Perhaps I install the compiler from MacPorts
after these experiences. And all that only to try openMP on Mac!

Still linking delivers now seemingly a followup-error:
Undefined symbols for architecture x86_64:
  "___cxa_throw_bad_array_new_length", referenced from:
      create_parameters_gui(bool) in cc2Df7NY.o



Am 03.11.2014 um 09:47 schrieb fxcoudert at gcc dot gnu.org
<gcc-bugzilla@gcc.gnu.org>:

> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262
> 
> Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> changed:
> 
>           What    |Removed                     |Added
> ----------------------------------------------------------------------------
>             Status|WAITING                     |RESOLVED
>         Resolution|---                         |INVALID
> 
> --- Comment #18 from Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> ---
> (In reply to Jakub Jelinek from comment #17)
>> The existence of the popcntq instruction suggests your libgmp has been
>> configured for Nehalem or later CPU, which you apparently don't have.
>> So, either convince whoever built it to configure it for earlier CPUs, or
>> build it on your own.  In any case, this doesn't sound like a GCC bug in any
>> way.
> 
> Thanks Jakub, I had missed that. It's actually much simpler. Karsten, you need
> to ask Gaurav to build his GMP/MPFR/MPC libraries with an explicit build
> parameter:
> 
> ./configure --build=x86_64-apple-darwin14.0.0
> 
> (for Yosemite). Otherwise, the GMP library is not portable. And if I recall
> correctly, MPFR and MPC use the GMP settings for their own compilation.
> 
> -- 
> You are receiving this mail because:
> You reported the bug.

Karsten
karo13de@googlemail.com
>From gcc-bugs-return-465637-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Mon Nov 03 12:12:31 2014
Return-Path: <gcc-bugs-return-465637-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 14493 invoked by alias); 3 Nov 2014 12:12:31 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 14424 invoked by uid 48); 3 Nov 2014 12:12:25 -0000
From: "fxcoudert at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/63262] [4.9.2] ICE including math.h on darwin
Date: Mon, 03 Nov 2014 12:12:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: c++
X-Bugzilla-Version: 4.9.2
X-Bugzilla-Keywords: ice-on-valid-code
X-Bugzilla-Severity: normal
X-Bugzilla-Who: fxcoudert at gcc dot gnu.org
X-Bugzilla-Status: RESOLVED
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields:
Message-ID: <bug-63262-4-629utGbv40@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-63262-4@http.gcc.gnu.org/bugzilla/>
References: <bug-63262-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2014-11/txt/msg00109.txt.bz2
Content-length: 639

https://gcc.gnu.org/bugzilla/show_bug.cgi?idc262

--- Comment #20 from Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> ---
(In reply to karsten from comment #19)
> I have replaced the resp. library by my MacPorts build and it compiles now!
> Perhaps a hacker way, however. Perhaps I install the compiler from MacPorts
> after these experiences. And all that only to try openMP on Mac!

Maybe try one of the builds on my website: http://coudert.name
They are very user-friendly, with graphical installer (they install in
/usr/local/gfortran). It says they're gfortran packages, but they also contain
the full C and C++ compilers.

[Bug c++/63262] [4.9.2] ICE including math.h on darwin

[karo13de at googlemail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262

--- Comment #21 from karsten <karo13de at googlemail dot com> ---
Thank you, but the 4.9.0 Mavericks built throws the same error:

Undefined symbols for architecture x86_64:
  "___cxa_throw_bad_array_new_length", referenced from:
      create_parameters_gui(bool) in ccnFtqfx.o
      cimg_library::CImg<unsigned long>::CImg(unsigned int, unsigned int,
unsigned int, unsigned int) in ccnFtqfx.o
      cimg_library::CImg<int>::CImg(unsigned int, unsigned int, unsigned int,
unsigned int) in ccnFtqfx.o
      void cimg_library::CImgDisplay::_render_resize<unsigned char, unsigned
char>(unsigned char const*, unsigned int, unsigned int, unsigned char*,
unsigned int, unsigned int) in ccnFtqfx.o
      void cimg_library::CImgDisplay::_render_resize<unsigned short, unsigned
short>(unsigned short const*, unsigned int, unsigned int, unsigned short*,
unsigned int, unsigned int) in ccnFtqfx.o
      void cimg_library::CImgDisplay::_render_resize<unsigned int, unsigned
int>(unsigned int const*, unsigned int, unsigned int, unsigned int*, unsigned
int, unsigned int) in ccnFtqfx.o
      cimg_library::CImg<double>::CImg(unsigned int, unsigned int, unsigned
int, unsigned int) in ccnFtqfx.o
      ...
ld: symbol(s) not found for architecture x86_64
collect2: error: ld returned 1 exit status

Unluckily I have no expertise to overcome such cryptic things.

Am 03.11.2014 um 13:12 schrieb fxcoudert at gcc dot gnu.org
<gcc-bugzilla@gcc.gnu.org>:

> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63262
> 
> --- Comment #20 from Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> ---
> (In reply to karsten from comment #19)
>> I have replaced the resp. library by my MacPorts build and it compiles now!
>> Perhaps a hacker way, however. Perhaps I install the compiler from MacPorts
>> after these experiences. And all that only to try openMP on Mac!
> 
> Maybe try one of the builds on my website: http://coudert.name
> They are very user-friendly, with graphical installer (they install in
> /usr/local/gfortran). It says they're gfortran packages, but they also contain
> the full C and C++ compilers.
> 
> -- 
> You are receiving this mail because:
> You reported the bug.

Karsten
karo13de@googlemail.com