NetBSD Problem Report #48272

From prlw1@inf.phy.cam.ac.uk  Fri Oct  4 13:02:36 2013
Return-Path: <prlw1@inf.phy.cam.ac.uk>
Received: from mail.netbsd.org (mail.netbsd.org [149.20.53.66])
	(using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits))
	(Client CN "mail.NetBSD.org", Issuer "Postmaster NetBSD.org" (verified OK))
	by mollari.NetBSD.org (Postfix) with ESMTPS id 160927109D
	for <gnats-bugs@gnats.NetBSD.org>; Fri,  4 Oct 2013 13:02:36 +0000 (UTC)
Message-Id: <E1VS51X-0003Us-Oh@quartz.inf.phy.cam.ac.uk>
Date: Fri, 04 Oct 2013 14:02:27 +0100
From: prlw1@cam.ac.uk
Reply-To: prlw1@cam.ac.uk
To: gnats-bugs@NetBSD.org
Subject: octave and blas build failure
X-Send-Pr-Version: 3.95

>Number:         48272
>Category:       pkg
>Synopsis:       octave and blas fail to build
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    pkg-manager
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Fri Oct 04 13:05:00 +0000 2013
>Last-Modified:  Fri Sep 04 13:50:00 +0000 2015
>Originator:     Patrick Welche
>Release:        NetBSD 6.99.24
>Organization:
>Environment:
NetBSD-current/amd64 with base GCC
>Description:
Building octave gets stuck with:

  /usr/lib/libstdc++.so.7: undefined reference to `_Unwind_GetIPInfo@GCC_4.2.0'

Thinking this may have something to do with fortran being built with a
different compiler to the one which build libstdc++, I tried rebuilding
g95 and libtool-fortran. Now octave fails on its dependency blas with:

--- libblas.la ---
libtool-fortran  --mode=link --tag=F77 g95 -o libblas.la caxpy.lo ...
   ztrsv.lo   -rpath /usr/pkg/lib -version-info 1:1
libtool-fortran: link: g77 -shared   .libs/caxpy.o ...  .libs/ztrsv.o
   -L/usr/obj/pkgsrc/math/blas/work.x86_64/.buildlink/lib
   -soname libblas.so.1 -o .libs/libblas.so.1.0.1
g95: libblas.so.1: No such file or directory
g95: unrecognized option '-soname'
*** [libblas.la] Error code 1 
>How-To-Repeat:
Try to build octave.
>Fix:

>Audit-Trail:
From: Joerg Sonnenberger <joerg@britannica.bec.de>
To: gnats-bugs@NetBSD.org
Cc: pkg-manager@netbsd.org, gnats-admin@netbsd.org, pkgsrc-bugs@netbsd.org
Subject: Re: pkg/48272: octave and blas build failure
Date: Sun, 6 Oct 2013 05:31:42 +0200

 On Fri, Oct 04, 2013 at 01:05:00PM +0000, prlw1@cam.ac.uk wrote:
 > >Number:         48272
 > >Category:       pkg
 > >Synopsis:       octave and blas fail to build
 > >Confidential:   no
 > >Severity:       serious
 > >Priority:       medium
 > >Responsible:    pkg-manager
 > >State:          open
 > >Class:          sw-bug
 > >Submitter-Id:   net
 > >Arrival-Date:   Fri Oct 04 13:05:00 +0000 2013
 > >Originator:     Patrick Welche
 > >Release:        NetBSD 6.99.24
 > >Organization:
 > >Environment:
 > NetBSD-current/amd64 with base GCC
 > >Description:
 > Building octave gets stuck with:
 > 
 >   /usr/lib/libstdc++.so.7: undefined reference to `_Unwind_GetIPInfo@GCC_4.2.0'

 Well, the problem here is that the libgcc from g95 is picked up, which
 doesn't support the (newer) interface system libstdc++ expects. Not sure
 how to fix this, not providing a shared Fortran RTL also breaks things.

 Joerg

From: Thomas Klausner <wiz@NetBSD.org>
To: NetBSD bugtracking <gnats-bugs@NetBSD.org>
Cc: 
Subject: Re: pkg/48272: octave and blas build failure
Date: Tue, 5 May 2015 11:33:59 +0200

 I've just successfully built math/octave on 7.99.15/amd64.
  Thomas

From: Patrick Welche <prlw1@cam.ac.uk>
To: gnats-bugs@NetBSD.org
Cc: 
Subject: Re: pkg/48272: octave and blas build failure
Date: Tue, 5 May 2015 23:52:24 +0100

 >  I've just successfully built math/octave on 7.99.15/amd64.

 Did you set PKGSRC_FORTRAN=gfortran or somesuch?

From: Thomas Klausner <wiz@NetBSD.org>
To: NetBSD bugtracking <gnats-bugs@NetBSD.org>
Cc: 
Subject: Re: pkg/48272: octave and blas build failure
Date: Wed, 6 May 2015 09:15:05 +0200

 On Tue, May 05, 2015 at 10:55:00PM +0000, Patrick Welche wrote:
 >  Did you set PKGSRC_FORTRAN=gfortran or somesuch?

 No.
  Thomas

From: Aleksej Saushev <asau@inbox.ru>
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: pkg/48272: octave and blas build failure
Date: Wed, 06 May 2015 23:55:07 +0300

 Thomas Klausner <wiz@NetBSD.org> writes:

 > The following reply was made to PR pkg/48272; it has been noted by GNATS.
 >
 > From: Thomas Klausner <wiz@NetBSD.org>
 > To: NetBSD bugtracking <gnats-bugs@NetBSD.org>
 > Cc: 
 > Subject: Re: pkg/48272: octave and blas build failure
 > Date: Wed, 6 May 2015 09:15:05 +0200
 >
 >  On Tue, May 05, 2015 at 10:55:00PM +0000, Patrick Welche wrote:
 >  >  Did you set PKGSRC_FORTRAN=gfortran or somesuch?
 >  
 >  No.

 You're using clang, that's why.
 (My guess.)

From: Thomas Klausner <wiz@NetBSD.org>
To: NetBSD bugtracking <gnats-bugs@NetBSD.org>
Cc: 
Subject: Re: pkg/48272: octave and blas build failure
Date: Wed, 6 May 2015 23:30:13 +0200

 On Wed, May 06, 2015 at 09:20:01PM +0000, Aleksej Saushev wrote:
 >  You're using clang, that's why.
 >  (My guess.)

 I'm pretty sure that I use gcc. I do have clang installed, but
 PKGSRC_COMPILER, CLANGBASE, and HAVE_LLVM are not set. I use modular
 xorg in case it makes a difference.
  Thomas

From: David Holland <dholland-pbugs@netbsd.org>
To: gnats-bugs@NetBSD.org
Cc: 
Subject: Re: pkg/48272: octave and blas build failure
Date: Thu, 9 Jul 2015 17:26:47 +0000

 On Tue, May 05, 2015 at 09:35:01AM +0000, Thomas Klausner wrote:
  >  I've just successfully built math/octave on 7.99.15/amd64.

 fwiw, I've been able to build blas just fine lately, although I
 haven't tried octave.

 -- 
 David A. Holland
 dholland@netbsd.org

From: Patrick Welche <prlw1@cam.ac.uk>
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: pkg/48272: octave and blas build failure
Date: Fri, 4 Sep 2015 14:47:28 +0100

 octave and friends now build successfully with g95. The problem is that
 the binary it generates doesn't behave e.g.:

 make test
 ...
   src/DLD-FUNCTIONS/besselj.cc ........................... PASS  180/180 
   src/DLD-FUNCTIONS/betainc.cc ........................... PASS    6/6   
   src/DLD-FUNCTIONS/bsxfun.cc ............................terminate called after throwing an instance of 'octave_execution_exception'
 panic: Abort trap -- stopping myself...
 Makefile:2127: recipe for target 'check' failed
 gmake[1]: *** [check] Abort trap (core dumped)
 gmake[1]: Leaving directory '/tmp/obj/math/octave/work/octave-3.6.4/test'
 Makefile:2549: recipe for target 'check' failed

 note the .cc => c++ and Joerg's comment.

 Simply quitting octave:

 octave:1> quit
 terminate called after throwing an instance of 'octave_interrupt_exception'
 panic: Abort trap -- stopping myself...
 attempting to save variables to 'octave-core'...
 save to 'octave-core' complete
 [1]   Abort trap (core dumped) octave


 A blas, lapack and octave built with PKGSRC_FORTRAN=gfortran is OK:

   src/DLD-FUNCTIONS/besselj.cc ........................... PASS  180/180 
   src/DLD-FUNCTIONS/betainc.cc ........................... PASS    6/6   
   src/DLD-FUNCTIONS/bsxfun.cc ............................ PASS   73/73  
   src/DLD-FUNCTIONS/cellfun.cc ........................... PASS  122/122 
   src/DLD-FUNCTIONS/chol.cc .............................. PASS   27/27  
   src/DLD-FUNCTIONS/conv2.cc ............................. PASS   23/23  
   src/DLD-FUNCTIONS/convhulln.cc ......................... PASS    4/4   
   src/DLD-FUNCTIONS/dassl.cc ............................. PASS    4/4   
 ...
 Summary:

   PASS   9863
   FAIL      6

NetBSD Home
NetBSD PR Database Search

(Contact us) $NetBSD: query-full-pr,v 1.39 2013/11/01 18:47:49 spz Exp $
$NetBSD: gnats_config.sh,v 1.8 2006/05/07 09:23:38 tsutsui Exp $
Copyright © 1994-2014 The NetBSD Foundation, Inc. ALL RIGHTS RESERVED.