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

[netCDF #GPS-461433]: netcdf-4: libsrc/v2i.c needs nc.h

This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.

  • Subject: [netCDF #GPS-461433]: netcdf-4: libsrc/v2i.c needs nc.h
  • Date: Thu, 29 Oct 2009 06:33:18 -0600
> Hi Ed,
> 
> > Well I looked into this issue again, and I see it is a complex one which 
> > has haunted us in the past.
> >
> > So I would like to fist see if your build works with --disable-v2.
> 
> well, I did a successful native build (but not a cross-compile) with
> --disable-v2 on the NEC SX with the latest snapshot.  I had to disable
> Fortran90 support for a different reason with the NetCDF sources.
> No HDF5 yet, so netcdf4 not enabled.  No patches applied.
> 
> > This will test whether the problem is confined to the v2i.c file, or exists 
> > in other part of the code.
> >
> > Please do send me *all* output, including all make check output and the 
> > config.log. It is hard to know what I will need to see, so please send it 
> > all.
> 
> See the attached log files (gzip-compressed).
> 
> log.config.sx8r contains the output of a job running the configure
> script.  Notice the following output:
> 
> checking for Fortran-equivalent to netCDF "byte"... integer
> checking for Fortran-equivalent to netCDF "byte"...
> integer(selected_int_kind(2))
> checking for Fortran-equivalent to netCDF "short"... integer*2
> checking for Fortran-equivalent to netCDF "short"... integer*2
> 
> (see my other mail) and
> 
> checking size of short... 2
> checking size of int... 4
> checking size of long... 8
> checking size of float... 4
> checking size of double... 8
> checking size of off_t... 8
> checking size of size_t... 4
> 
> (these values are correct).
> 
> 
> log.make.sx8r is the output of "make -k check".  There are two
> failures.  The ncgen/ncdump test hits a known bug in printf()
> on SuperUX (poor rounding).  The C++ failure is likely a compiler
> bug which I did not investigate because I do not care about the
> C++ interface.
> 
> So it looks like the problem in the current report is indeed limited to
> V2 support.
> 
> Cheers,
> Harald


Howdy Harald!

Well, I have kept this support issue open all this time (nearly a year!) hoping 
to get back to this problem and try and get your patch working everywhere. But, 
alas, the time passes and the difficult support request remains unsolved. 
Meanwhile, the --disable-v2 option hopefully provided a useful workaround for 
you.

Now, of course, the code has moved along a great deal. Not only has netCDF-4 
changed internally, but many new features have been added, including HDF4 and 
HDF5 interoperability, parallel I/O for classic files, and, most of all, the 
new opendap client, which allows remote access to netCDF data files, and has 
had a huge impact on our codebase.

I am going to close this support issue. If this problem comes up with the 
updated codebase, maybe we can take another stab at it.

Thanks,

Ed

Ticket Details
===================
Ticket ID: GPS-461433
Department: Support netCDF
Priority: Critical
Status: Closed