20010626: problem with 8-byte reals and doubleprecisions: "pgf90 -r8"

Dear Zion Wang;

I regret to inform you that the Fortran-90 interface of the netCDF
package doesn't support environments where real and doubleprecision
values are both 8-bytes wide -- and we have no immediate plans to
address this situation.

Workarounds include the following:

    1.  Not using the Fortran-90 interface (as you saw, the Fortran-77
        interface works OK in this situation);

    2.  Modifying your code -- replacing the use of real values with
        doubleprecision ones where appropriate.

Good luck.

Regards,
Steve Emmerson   <http://www.unidata.ucar.edu>

> To: <address@hidden>
> From: "Zion S. Wang" <address@hidden>
> Subject: problem trying to compile in double precision
> Organization: UCAR/Unidata
> Keywords: 200106262124.f5QLOT107006
...
> I'm trying to compile netcdf under Linux Redhat 7.1 w/ pgf90.  
> I can compile (make) and run the test program (make test) 
> successfully.  The results are shown in the attached 
> file - no.problem.report.
> 
> However, when I try to compile a double precision (-r8) option 
> of netcdf, compilation (make) ran fine, but make test did not.  
> The results are shown in the attached file - problem.report.
> 
> Please feel free to contact me at address@hidden or 
> (909) 781-5655 if you need any additional information.
> 
> Thanks.
> 
> Zion Wang
> 
> * * * * * * * * * * * * * * * * * *
> Zion Wang, Ph.D. (address@hidden)
> CE-CERT, UC Riverside
> (909) 781-5655
> (909) 781-5790 (fax)
...