Skip Menu |

This queue is for tickets about the Math-GSL CPAN distribution.

Report information
The Basics
Id: 88922
Status: new
Priority: 0/
Queue: Math-GSL

People
Owner: Nobody in particular
Requestors: ghaverla [...] materialisations.com
Cc:
AdminCc:

Bug Information
Severity: (no value)
Broken in: (no value)
Fixed in: (no value)



Subject: New attempt from Debian/unstable
Date: Mon, 23 Sep 2013 22:18:44 -0600
To: bug-Math-Gsl [...] rt.cpan.org
From: ghaverla <ghaverla [...] materialisations.com>
Hello. I ran into problems a year or so ago, which seemed to be mostly Swig/XS stack related. I had no idea how to proceed, and so moved to other things. I got a reason to look into Math::GSL again recently, so today I tried the build process again. I got a bunch of strange messages, so I just blew away the original build directory and tried installing again. I recently went through the Debian 5.14 to 5.18 transition. Anything that looks like Swig or XS seems to be up to date. All the problems up until the build fails, deal with FFT. If FFT for GSL means using FFTW, I have updates on FFTW scheduled but stalled. I don't know that GSL is using Boost libraries, but I also have a bunch of Boost libraries scheduled for update, but stalled. Googling, I am remembering what I was seeing a year ago: Bizarre copy of UNKNOWN. Partway through (Sun, 15 Jan 2012 12:13:21 +0100) http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=655748 XS perl some variable returned following something which looks like sv_setsv() isn't doing what it is supposed to. A later email in that thread, talks about LFS (Large FileSystem?) flags, and their affect on Stat_t variables (which includes PL_sv_undef) (Mon, 20 Feb 2012 19:05:43 +0200). In any event, the errors I am seeing are: Building Math-GSL Process XS files version 1.15 In file included from /usr/lib/perl/5.18/CORE/perl.h:5156:0, from xs/FFT_wrap.1.15.c:731: xs/FFT_wrap.1.15.c: In function ‘boot_Math__GSL__FFT’: xs/FFT_wrap.1.15.c:8304:61: error: ‘forward’ undeclared (first use in this function) sv_setsv(sv, SWIG_From_int SWIG_PERL_CALL_ARGS_1((int)(forward))); ^ /usr/lib/perl/5.18/CORE/embed.h:644:59: note: in definition of macro ‘sv_setsv_flags’ #define sv_setsv_flags(a,b,c) Perl_sv_setsv_flags(aTHX_ a,b,c) ^ xs/FFT_wrap.1.15.c:8304:5: note: in expansion of macro ‘sv_setsv’ sv_setsv(sv, SWIG_From_int SWIG_PERL_CALL_ARGS_1((int)(forward))); ^ xs/FFT_wrap.1.15.c:8304:33: note: in expansion of macro ‘SWIG_PERL_CALL_ARGS_1’ sv_setsv(sv, SWIG_From_int SWIG_PERL_CALL_ARGS_1((int)(forward))); ^ xs/FFT_wrap.1.15.c:8304:61: note: each undeclared identifier is reported only once for each function it appears in sv_setsv(sv, SWIG_From_int SWIG_PERL_CALL_ARGS_1((int)(forward))); ^ /usr/lib/perl/5.18/CORE/embed.h:644:59: note: in definition of macro ‘sv_setsv_flags’ #define sv_setsv_flags(a,b,c) Perl_sv_setsv_flags(aTHX_ a,b,c) ^ xs/FFT_wrap.1.15.c:8304:5: note: in expansion of macro ‘sv_setsv’ sv_setsv(sv, SWIG_From_int SWIG_PERL_CALL_ARGS_1((int)(forward))); ^ xs/FFT_wrap.1.15.c:8304:33: note: in expansion of macro ‘SWIG_PERL_CALL_ARGS_1’ sv_setsv(sv, SWIG_From_int SWIG_PERL_CALL_ARGS_1((int)(forward))); ^ xs/FFT_wrap.1.15.c:8309:61: error: ‘backward’ undeclared (first use in this function) sv_setsv(sv, SWIG_From_int SWIG_PERL_CALL_ARGS_1((int)(backward))); ^ /usr/lib/perl/5.18/CORE/embed.h:644:59: note: in definition of macro ‘sv_setsv_flags’ #define sv_setsv_flags(a,b,c) Perl_sv_setsv_flags(aTHX_ a,b,c) ^ xs/FFT_wrap.1.15.c:8309:5: note: in expansion of macro ‘sv_setsv’ sv_setsv(sv, SWIG_From_int SWIG_PERL_CALL_ARGS_1((int)(backward))); ^ xs/FFT_wrap.1.15.c:8309:33: note: in expansion of macro ‘SWIG_PERL_CALL_ARGS_1’ sv_setsv(sv, SWIG_From_int SWIG_PERL_CALL_ARGS_1((int)(backward))); ^ error building .o file from 'xs/FFT_wrap.1.15.c' at inc/GSLBuilder.pm line 286. Hopefully something in there rings a bell. Have a great day! Gord
Subject: Re: [rt.cpan.org #88922] New attempt from Debian/unstable
Date: Tue, 24 Sep 2013 17:10:35 -0600
To: bug-Math-Gsl [...] rt.cpan.org
From: ghaverla <ghaverla [...] materialisations.com>
I brought this up with the Debian maintainer for GSL, mentioning the LFS connection to another perl problem, and he doesn't think the problem is with the various packages present from Debian involving GSL. But, until a person solves the problem, one never knows. How would a person install a second version of GSL that is to be used by Math::GSL only? Have a great day! Gord