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

Re: some thing wrong about FPP

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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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: Re: some thing wrong about FPP
  • Date: Fri, 18 Jul 1997 08:54:58 -0600
>To: address@hidden
>From: "Qiao Jinsong" <address@hidden>
>Subject: Problem in installing netcdf
>Organization: UCAR
>Keywords: 199707171310.HAA09585

Hi Qiao,

> Thank you for your information. You are wright, the problem about cxx is =
> because of the wrong C++ compiler. I have tried again to configure the =
> environment, but the there are some in-compatible of F90. I bypassed the =
> configure, just added the FC CC CXX manually with "-n32" flag in the =
> file of macros.make. Here's the result of "make all test" command. Could =
> you help me to determine is it OK, or what I still need to do. Is the =
> FORTRAN interface inside the directory of "fortran"? I think the make =
> has compiled it and added into the library. Is there any other version =
> of fortran That I need to do.? Is there any problem if I run make under =
> cshell?

If you choose to bypass the configure step, it will be difficult to help
much, because the configure script creates the "macros.make" file and
making inconsistent changes to this generated file "voids your warranty"
:-).  Nevertheless, I'll try to help.  The output from "make all test"
shows the C interface works fine.  

Yes, the netCDF-2 and netCDF-3 FORTRAN interfaces are in the "fortran/"
directory, and the output shows both of these were compiled into the
library.  Also, the test of the netCDF-2 FORTRAN interface succeeded.

However the test of the netCDF-3 FORTRAN interface failed because it's
trying to use a C++ compiler on Fortran code!

  ...
> Making `test' in directory /usr/people/joe/netcdf-3.3.1/src/fortran
> 
> + CC -n32 -E ftest.F
> + grep -v ^#
> CC ERROR:  cannot exec /usr/lib32/cmplrs/mfef77
> CC INTERNAL ERROR:  /usr/lib32/cmplrs/mfef77 returned non-zero status 32
> + rm ftest.f
> + exit 1
> *** Error code 1 (bu21)

The "ftest.F" file is Fortran code that has preprocessor directives in
it.  The Fortran compilers on IRIX handle ".F" files and their
preprocessor directives, but "CC" can't be used for this purpose.  The
use of CC where a Fortran compiler is needed is probably the result of
some change you have made to macros.make.

Several of the settings for IRIX systems under the "TESTED SYSTEMS"
section in the INSTALL document of the netCDF 3.3.1 distribution were
missing quotes and contained some unnecessary flags, as described in our
"Known Problems" web page, at

  http://www.unidata.ucar.edu/packages/netcdf/known_problems.html

A corrected version of the INSTALL document is available from

  http://www.unidata.ucar.edu/packages/netcdf/INSTALL

If you follow the procedures in this INSTALL document from a new
unmodified copy of the current netCDF release, things may work better.
If you run into problems, just send us all the information requested
under the "IF PROBLEMS OCCUR" heading of the INSTALL document, and we'll
try to help.

--Russ

_____________________________________________________________________

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