20001026: NetCDF-Perl-1.2.1 "make test" failure

Michael,

>Date: Thu, 26 Oct 2000 14:17:39 -0700
>From: "Michael P. Urban" <address@hidden>
>Organization: NASA/JPL
>To: Steve Emmerson <address@hidden>
>Subject: Re: 20001026: NetCDF-Perl-1.2.1 "make test" failure 
>Keywords: 200010261832.e9QIWj405557

The above message contained the following:

...
> ./configure: ERROR: The following variables need values:
> 
> VARIABLE     DESCRIPTION                EXAMPLE
> --------     -------                    -------
> CPP_NETCDF   netCDF header              -I/usr/local/unidata/include
> LD_NETCDF    netCDF library             -L/usr/local/unidata/lib -lnetcdf
> 
> For each variable above, this script will now request that you input
> a value appropriate for your system (note that this value will not
> be interpolated by a shell -- so don't use shell substitutions).
> Alternatively, you can interrupt this script, set the above variables
> in the environment or in the file CUSTOMIZE, and then re-execute this
> script.  (Variables referring to executable programs needn't be set if
> the relevant directory is added to PATH.  See file INSTALL for details.)
> 
> Enter a value for the netCDF header (e.g. "-I/usr/local/unidata/include"):
> -I/tmp

Does the file "/tmp/netcdf.h" exist?

> 
> Enter a value for the netCDF library (e.g. "-L/usr/local/unidata/lib 
> -lnetcdf"):
> -L/tmp -lnetcdf

Does the file "/tmp/libnetcdf.a" exist?

> updating cache ./config.cache
> creating ./config.status
> creating Makefile
> creating perl/Makefile.PL
> creating port/master.mk
> creating port/Makefile
> expanding `include's in file `Makefile'
> expanding `include's in file `perl/Makefile.PL'
> expanding `include's in file `port/master.mk'
> expanding `include's in file `port/Makefile'
> % make perl/Makefile
> cd perl && perl Makefile.PL POLLUTE=1
> Checking if your kit is complete...
> Looks good
> Writing Makefile for NetCDF
> % make
> 
> making `dynamic' in directory /tmp/netcdf-perl-1.2.1/src/perl
> 
> /usr/local/bin/perl -I/usr/local/lib/perl5/5.6.0/sun4-solaris 
> -I/usr/local/lib/perl5/5.6.0 /usr/local/lib/perl5/5.6.0/ExtUtils/xsubpp  
> -typemap /usr/local/lib/perl5/5.6.0/ExtUtils/typemap NetCDF.xs > NetCDF.xsc 
> && mv NetCDF.xsc NetCDF.c
> Please specify prototyping behavior for NetCDF.xs (see perlxs manual)
> gcc -c -I/tmp -fno-strict-aliasing -I/usr/local/include -D_LARGEFILE_SOURCE 
> -D_FILE_OFFSET_BITS=64 -DPERL_POLLUTE -O      -DVERSION=\"

\"  
> -DXS_VERSION=\"

\" -fPIC -I/usr/local/lib/perl5/5.6.0/sun4-solaris/CORE  
> NetCDF.c
> mkdir blib
> mkdir blib/arch
> mkdir blib/arch/auto
> mkdir blib/arch/auto/NetCDF
> Running Mkbootstrap for NetCDF ()
> chmod 644 NetCDF.bs
> LD_RUN_PATH="/tmp" gcc -o blib/arch/auto/NetCDF/NetCDF.so -R/tmp -G 
> -L/usr/local/lib NetCDF.o    -L/tmp -lnetcdf 
> chmod 755 blib/arch/auto/NetCDF/NetCDF.so
> cp NetCDF.bs blib/arch/auto/NetCDF/NetCDF.bs
> chmod 644 blib/arch/auto/NetCDF/NetCDF.bs
> 
> returning to directory /tmp/netcdf-perl-1.2.1/src
> 
> test ! -f perl/so_locations || chmod a+w perl/so_locations

The above looks identical to my build of netcdf-perl version 1.2.1 on my
SunOS 5.7 system using perl 5.6.0 except that I used Sun's C compiler 
(/opt/SUNWspro/bin/cc) rather than "gcc".  Sun's C compiler was also
used to build my perl utility.  What compiler was used to build yours
(try "perl -V")?

> % make test
> 
> making `test' in directory /tmp/netcdf-perl-1.2.1/src/perl
> 
> mkdir blib/lib
> mkdir blib/lib/auto
> mkdir blib/lib/auto/NetCDF
> cp NetCDF.pm blib/lib/NetCDF.pm
> AutoSplitting blib/lib/NetCDF.pm (blib/lib/auto/NetCDF)
> PERL_DL_NONLAZY=1 /usr/local/bin/perl -Iblib/arch -Iblib/lib 
> -I/usr/local/lib/perl5/5.6.0/sun4-solaris -I/usr/local/lib/perl5/5.6.0 test.pl
> Can't load 'blib/arch/auto/NetCDF/NetCDF.so' for module NetCDF: ld.so.1: 
> /usr/local/bin/perl: fatal: relocation error: file 
> blib/arch/auto/NetCDF/NetCDF.so: symbol Vdetach: referenced symbol not found 
> at /usr/local/lib/perl5/5.6.0/sun4-solaris/DynaLoader.pm line 200.
>  at test.pl line 3

I don't get this error and am at a loss to understand why you do.

Regards,
Steve Emmerson   <http://www.unidata.ucar.edu>