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

building using NAG f90



Steve,

>Date: 13 Jun 1997 13:02:06 -0500 
>From: "Steve Mauget" <address@hidden>
>Organization: USDA
>To: "Steve E" <address@hidden>
>Subject: Latest tests... 
>Keywords: 199706021417.IAA27579

In the above message, you wrote:

>     Here are the results from your suggested test...
> 
>                                                   Steve M.
> 
> I. make -n ftest.o >& ftest.log
> 
> in ftest.log....
> 
>       case "" in      \
>           '') \
>               set -x; \
>               /bin/c89 -E  ftest.F | grep -v '^#' >ftest.f ||         \
>                   (rm ftest.f ; exit 1);      \
>               /usr/local/bin/f90 -c -g ftest.f || (rm ftest.f; exit 1);       
> \
>               rm ftest.f;     \
>               ;;      \
>           *)  \
>               set -x; \
>                ftest.F;       \
>               ;;      \
>       esac
> 
> II. make ftest.o FFLAGS='-mismatch' >>& ftest.log
> ( not clear about the reference to flags from previous step, only
>   included -mismatch ...)

You could have set FFLAGS to "-mismatch -g" because the "-g" option
appeared in the previous steve.  No matter, though.

> 
> in ftest.log....
> 
> + /bin/c89 -E ftest.F
> + grep -v ^#
> + 1> ftest.f
> + /usr/local/bin/f90 -c -mismatch ftest.f
> Warning: ftest.f, line 1446: Unused external reference NF_INQ_NDIMS
>          detected at END@<end-of-statement>
> Warning: ftest.f, line 1446: Unused external reference NF_GET_VARA_INT1
>          detected at END@<end-of-statement>
>                                      "
>                       Many warning messages edited out,

I wonder if NAG f90 has an option to turn off warning messages about
unused external references.  Is so, then you should add it to FFLAGS
when you run the configure script.

>                                      "
> Warning: ftest.f, line 13572: Inconsistent datatype for arg 5 in call to NCVGT
> Warning: ftest.f, line 13583: Inconsistent datatype for arg 5 in call to NCVGT
> Warning: ftest.f, line 13603: Inconsistent datatype for arg 5 in call to NCVGT
> Warning: ftest.f, line 14933: Inconsistent datatype for arg 4 in call to
> NCVGT1
> Warning: ftest.f, line 14939: Inconsistent datatype for arg 4 in call to
> NCVGT1
> Warning: ftest.f, line 14945: Inconsistent datatype for arg 4 in call to
> NCVGT1
> Warning: ftest.f, line 14952: Inconsistent datatype for arg 4 in call to
> NCVGT1
> Warning: ftest.f, line 16276: Inconsistent datatype for arg 5 in call to NCVPT
> Warning: ftest.f, line 16284: Inconsistent datatype for arg 5 in call to NCVPT
> Warning: ftest.f, line 16300: Inconsistent datatype for arg 5 in call to NCVPT
> Warning: ftest.f, line 16316: Inconsistent datatype for arg 5 in call to NCVPT
> Warning: ftest.f, line 17623: Inconsistent datatype for arg 4 in call to
> NCVPT1
> Warning: ftest.f, line 17627: Inconsistent datatype for arg 4 in call to
> NCVPT1
> Warning: ftest.f, line 17631: Inconsistent datatype for arg 4 in call to
> NCVPT1
> Warning: ftest.f, line 17635: Inconsistent datatype for arg 4 in call to
> NCVPT1
> Warning: ftest.f, line 18923: Inconsistent datatype for arg 4 in call to
> NCVGT1
> Warning: ftest.f, line 18924: Inconsistent datatype for arg 4 in call to
> NCVGT1
> Warning: ftest.f, line 18925: Inconsistent datatype for arg 4 in call to
> NCVGT1
> Warning: ftest.f, line 18926: Inconsistent datatype for arg 4 in call to
> NCVGT1
> + rm ftest.f

Well, it looks like the "-mismatch" option transforms the errors into
warnings.  This is good.

To see if the netCDF-2 (not the new netCDF-3) Fortran API works, try the
following:

    1.  Go to the top-level source directory.

    2.  Execute the command "make libsrc/all fortran/all".
        Redirect standard output and standard error to a file.
        If this step fails, then send me the file and stop.

    3.  Execute the command "make fortran/test".
        Redirect standard output and standard error to a file.
        If this step fails, then send me the file and stop.

--------
Steve Emmerson   <address@hidden>