[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[netCDF #PVO-840684]: Question About Make Check In netcdf4.1.1



Hi Tony,

Sorry to have taken so long to respond to your question.  We're short-handed 
due to vacations, and have a growing backlog of support questions, with only
two developers available and an oversubscribed list of commitments.  I was 
actually looking at your question when your "SECOND ATTEMPT" email arrived.

We haven't been able to duplicate the problem here, but that may be because
our only AIX platform is an old powerpc-ibm-aix5.1.0.0, on which we still
haven't installed HDF5-1.8.5.  We haven't seen the problem you reported on
any other of our testing platforms.

We run different tests than the HDF5 tests, so it's not surprising that we
could get different results.  The problem you reported is a symptom of getting
an error return from the HDF5 call in libsrc4/tst_h_files.c

  if (H5Dwrite(datasetid, native_typeid, mem_spaceid, file_spaceid, 
xfer_plistid, bufr) < 0) ERR;

so my next step would be to rebuild with HDF5 detailed logging turned on by 
configuring with --enable-logging and making an initial call to set the HDF5
logging level to get more detail by calling

  nc_set_log_level(5);

sometime before the error occurs.  That will tell what HDF5 error is actually
occurring.  Or you could see it by using a debugger to descend into the HDF5
level calls.  The person who would usually do this won't be back until Aug 13, 
but I'll try to investigate further before then ...

--Russ


> I installed hdf5-1.8.5.  The resulting "make check" revealed no problems
> (none that I can find).  Next I installed netcdf4.1.1.  The resulting
> "make check" is saying there are errors in hdf5-1.8.5.  So I've got
> netcdf4.1.1's "make check" saying there are errors in hdf5-1.8.5.  (See
> attached files for hdf5-1.8.5 and netcdf4.1.1.)  I wasn't sure how to
> proceed so I came to you first.  Is netcdf4.1.1's "make check" telling
> the truth or is hdf5-1.8.5's "make check" telling the truth?
 ...
> Making check in libsrc4
 ...
> *** Checking HDF5 file functions.
> *** Checking HDF5 file creates and opens...ok.
> *** Checking HDF5 file creates and opens some more...ok.
> *** Creating file...ok.
> *** Creating HDF5 file in the canonical netCDF-4 way...ok.
> *** large file test for HDF5...Sorry! Unexpected result, tst_h_files.c, line: 
> 400
> 1 failures
> 1 errors detected! Sorry!
> FAIL: tst_h_files

  
Russ Rew                                         UCAR Unidata Program
address@hidden                      http://www.unidata.ucar.edu



Ticket Details
===================
Ticket ID: PVO-840684
Department: Support netCDF
Priority: Normal
Status: Closed