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

20040730: update: building netcdf 3.5.1 with NAG f95 v5.0 on OS X

Eric,

>Date: Fri, 30 Jul 2004 14:48:37 -0700
>From: Eric Salathe <address@hidden>
>Organization: University of Washington
>To: Steve Emmerson <address@hidden>
>Subject: Re: 20040730: update: building netcdf 3.5.1 with NAG f95 v5.0 on OS X 
> Keywords: 200407302101.i6UL1haW016337 netCDF NAG F95

The above message contained the following:

> Oops, you are right. I had made other hacks, which allowed it to work,  
> and forgot to remove them before blindly trying the cpp flag.

That make more sense.

> It is very odd, and I am sure and issue w/ NAG f95 v5.0. If I change  
> f90/netcdf_test.f90 calls to nf90_dev_var like this (add (/.../) around  
> latDimID):
> 
> <   call check(nf90_def_var(ncFileID, "lat", nf90_float, dimids =  
> latDimID, varID = latVarID) )
> ---
>  >   call check(nf90_def_var(ncFileID, "lat", nf90_float, dimids =  
> (/latDimID/), varID = latVarID) )
> 
> the f90 test completes w/o error. So the NAGware issue has something to  
> do with how the rank of dimids is detected and the correct nf_* (f77)  
> routine is selected, but I get over my head after that.
> 
> So, let me retract that NAG f95 v5.0 works. Version 4.2 successfully  
> builds netcdf 3.5.1, however.

Interesting.  The problem could lie with the compiler (we've seen another
compiler have a problem).

> -Eric
> --
> Eric Salathe
> Climate Impacts Group             <address@hidden>
> University of Washington           
> <http://www.atmos.washington.edu/~salathe>

Regards,
Steve Emmerson