Many thanks for detailed instructions David. See below.
On Fri, Jan 15, 2010 at 8:55 AM, David Mertens <[email protected]> wrote:
> A few thoughts on what troubles you:
>
>>
>> I downloaded PDL 2.4.6, and following my previous install, config-ed
>> the perldl.conf to build without PLPLOT and Karma but with HDF5.
>
> I am 99% certain that the HDF library that comes with PDL is not HDF5 but
> HDF4. For HDF5, I found this on CPAN. I'm not sure how the two compare,
> but I was led to believe that HDF4 is quite different from HDF5.
Yes, HDF5 is very different from HDF4, and HDF5 itself has two
branches. I installed my own HDF5, the latest branch numbered 1.8.x.
Downloaded, untarred/ungzipped, configure, make, make install. The src
is in
~/Projects/hdf5-1.8.2
So, 'make install' installed everything in
punkish ~/Projects/hdf5-1.8.2/hdf5$ll
total 0
drwxr-xr-x 19 root punkish 646 Jan 14 16:15 bin/
drwxr-xr-x 41 root punkish 1394 Jan 14 16:15 include/
drwxr-xr-x 11 root punkish 374 Jan 14 16:15 lib/
punkish ~/Projects/hdf5-1.8.2/hdf5$
I updated my perldl.conf like so
WITH_HDF => 1,
HDF_LIBS => '/Users/punkish/Projects/hdf5-1.8.2/hdf5/lib',
HDF_INC => '/Users/punkish/Projects/hdf5-1.8.2/hdf5/include',
>
>>
>> ===== during the Makefile process =======
>> punkish ~/Projects/PDL-2.4.6$perl Makefile.PL
>>
>> ..
>>
>> Writing Makefile for PDL::IO::FITS
>> Will forcibly try and build PDL::IO::HDF on this system
>> WARNING: Unknown cpu type i386! Not setting $hdf_defs. (This may not
>> be a bad thing)
>> Final $hdf_defs flags: ''
>> Building PDL::IO::HDF. Turn off WITH_HDF if there are any problems
>
>
> That business about 'This may not be a bad thing' is correct in your case.
> This warning message gets made in the process of attempting to set
> Linux-specific compiler settings, which is not a very portable idea to begin
> with. Furthermore, this stage in the process obtains the cpu type with the
> command `uname -m`, which seems to me to have portability problems written
> all over it. (In fact, it is preceeded with a comment that states "NOTE TO
> SELF: The main PDL developers may not like this...") The whole check should
> probably be replaced or removed.
>
> To work around this, edit the file IO/HDF/Makefile.PL and comment out lines
> 106, 107, and 109-118. Then run perl Makefile.PL and see what you get.
>
>> Found FFTW2 double version in /usr/local/include/fftw.h !
>> Libs: 'libfftw.a', 'librfftw.a' Not Found
>> Looked in: '/lib', '/usr/lib', '/usr/local/lib'
>> Skipping build of PDL::FFTW.
>> writing dummy Makefile
>
> Sorry, I don't know what's going on here. I don't know enough about FFTW2.
> However, PDL has its own FFT routines, so unless you really need the
> capabilities of FFTW2, I wouldn't worry much about it. You ignore this, or
> modify your configuration file appropriately.
>
>>
>> Skip ../../blib/lib/PDL/IO/FITS.pm (unchanged)
>> Manifying ../../blib/man3/PDL::IO::FITS.3
>>
>> make[2]: Makefile: No such file or directory
>
>
> Well, this is strange, now isn't it? I'm not sure quite what's going on.
> Here's my recommendation to try and remedy the problem:
>
> # Navigate over to PDL/IO/FITS in the source tarball
> cd IO/FITS
> perl Makefile.PL
> make clean
> make
> cd ../..
> make
>
> Hopefully having make clean up whatever is in IO/FITS will fix the problem
> for you. Write back if it still gives trouble.
>
See below.
punkish ~/Projects/PDL-2.4.6/IO/FITS$make clean
rm -f \
*.a core \
core.[0-9] blib/arch/auto/PDL/IO/FITS/extralibs.all \
core.[0-9][0-9] FITS.bso \
pm_to_blib.ts core.[0-9][0-9][0-9][0-9] \
FITS.x \
perl tmon.out \
*.o pm_to_blib \
blib/arch/auto/PDL/IO/FITS/extralibs.ld blibdirs.ts \
core.[0-9][0-9][0-9][0-9][0-9] *perl.core \
core.*perl.*.? Makefile.aperl \
perl FITS.def \
core.[0-9][0-9][0-9] mon.out \
libFITS.def perlmain.c \
perl.exe so_locations \
FITS.exp
rm -rf \
blib
mv Makefile Makefile.old > /dev/null 2>&1
punk...@dyn-72-33-135-24 ~/Projects/PDL-2.4.6/IO/FITS$make
make: *** No targets specified and no makefile found. Stop.
punkish ~/Projects/PDL-2.4.6/IO/FITS$
> Hope that helps!
>
> David
>
--
Puneet Kishor http://www.punkish.org
Carbon Model http://carbonmodel.org
Charter Member, Open Source Geospatial Foundation http://www.osgeo.org
Science Commons Fellow, http://sciencecommons.org/about/whoweare/kishor
Nelson Institute, UW-Madison http://www.nelson.wisc.edu
-----------------------------------------------------------------------
Assertions are politics; backing up assertions with evidence is science
=======================================================================
Sent from Madison, Wisconsin, United States
_______________________________________________
Perldl mailing list
[email protected]
http://mailman.jach.hawaii.edu/mailman/listinfo/perldl
