[Support #UPB-190795]: Compiling netcdf-4-daily & netcdf-4.1 on Altix 64-bit ia64 machines

> Hi Ed,
> 
> Its Paida once more, I was trying to build netcdf for use on the ia64
> cluster. I am getting some errors. Built and checked zzlib-1.2.3 and
> hdf5-1.8.4 just fine. Here is my environment
> CC=gcc
> CXX=g++
> FC=gfortran
> FXX=gfortran

Howdy Paida!

OK, I have recently completed a sweep through the code, fixing some memory 
problems, in particular one related to variable length types (and the string, 
where you are having problems, is a variable length type under the covers.)

So if you could try again with the latest snapshot, we will see if I caught 
your problem too.

For the purposes of testing this, you can disable the fortran and C++ APIs, and 
things will go faster. To do this, use --disable-fortran --disable-cxx when 
running configure. Also if HDF5 and zlib are installed in the same directory, 
you only need to give the location once, so just do --with-hdf5=/someplace and 
it will find zlib there too.

Find the daily snapshot here:
ftp://ftp.unidata.ucar.edu/pub/netcdf/snapshot/netcdf-4-daily.tar.gz

If this doesn't work, send me the complete output of configure and make, and 
the config.log file, in your response. I need these to further debug your 
problem (if you still have one after getting this snapshot).

Thanks!

Ed

> 
> configure command: (same command structure works on the i686's and the
> x86_64's)
> ./configure
> --with-hdf5=/home/pem/ia64/netcdf-4.1-snapshot2009121010_libs
> --with-zlib=/home/pem/ia64/netcdf-4.1-snapshot2009121010_libs
> --prefix=/home/pem/ia64/netcdf-4.1-snapshot2009121010_libs
> 
> I do a gmake check install
> 
> These are the errors:
> 
> ================================================
> 2 of 52 tests failed
> Please report to address@hidden
> ================================================
> 
> *** Tests successful!
> PASS: tst_opaques
> 
> *** Testing netcdf-4 string type.
> *** testing string attribute...ok.
> *** testing string var functions.../bin/sh: line 1: 236908
> Aborted                 ${dir}$tst
> FAIL: tst_strings
> 
> *** Testing HDF5/NetCDF-4 interoperability...
> *** Creating a HDF5 file with one var with two dimension scales...ok.
> *** Checking that HDF5 file can be read by netCDF-4, and adding an att...ok.
> *** Checking that one var, two dimscales, one att file can still be read
> by HDF5...ok.
> *** Creating a HDF5 file with one var and no dimension scales...ok.
> *** Creating a HDF5 file with one var and no dimension scales, without
> creation ordering...ok.
> *** Creating a HDF5 file with two vars and no dimension scales, without
> creation ordering...ok.
> *** Creating a HDF5 file with fixed length string...ok.
> *** Checking a HDF5 file with scalar, fixed-length string dataset...ok.
> *** Tests successful!
> PASS: tst_interops
> 
> 
> 
> and-------------------------
> 
> *** Testing UTF-8 names.
> *** creating UTF-8 names in classic model netcdf files...ok.
> *** renaming to UTF-8 names in classic model netcdf files...ok.
> *** creating UTF-8 names in netcdf-4 file...ok.
> *** Tests successful!
> PASS: tst_utf8
> 
> *** Testing fill values.
> *** testing empty fill values of a string var.../bin/sh: line 1: 236929
> Aborted                 ${dir}$tst
> FAIL: tst_fills
> 
> *** Testing for a netCDF-4 fill-value bug.
> *** Testing fill-values...ok.
> *** Tests successful!
> PASS: tst_fillbug
> 
> 
> Oh when I tried building the stable release, I came across the same
> problem yana came across, see
> http://www.unidata.ucar.edu/support/help/MailArchives/netcdf/msg05206.html
> 
> I get the same error with the above environment settings.
> 
> Thanks,
> 
> Paida
> 
> 


Ticket Details
===================
Ticket ID: UPB-190795
Department: Support netCDF
Priority: Critical
Status: Closed