Re: 20030403:netcdf, cfortran.h and lf95

>To: address@hidden
>cc: address@hidden
>From: Gareth Williams <address@hidden>
>Subject: netcdf, cfortran.h and lf95
>Organization: UCAR/Unidata
>Keywords: 200304030651.h336pN7U022849

Hi Gareth,

> I have just compiled netcdf-3.5.1-beta10 using the Lahey fortran compiler
> lf95 (version L6.00a).
> 
> lf95 by default decorates names by converting to lower case and adding a
> single underscore, so the settings for -Df2cFortran (which I tried
> first) are not appropriate.
> 
> To get it to compile I used -DNAGf90Fortran to get the right behavior from
> cfortran.h and avoid the f2cFortran kludge in nf_test/fortlib.c
> 
> Could you please add this information to the documentation for cfortran.h
> and the install docs for netcdf.
> 
> Perhaps a new define, (-DlaheyFortran?) would be appropriate, but that
> would require co-ordinateion between cfortran.h and netcdf which is why I
> am writing to you both at once.  Note also that the version of cfortran.h
> in netcdf-3.5.1-beta10 is 4.1 whereas version 4.3 is available from
> http://www-zeus.desy.de/~burow/cfortran/ -- but still with no lf95
> support.

We currently have a pointers to information from Dirk Slawinski about
using Lahey F95 version 5.5 with netCDF on Linux and Windows NT:

  http://www.unidata.ucar.edu/packages/netcdf/other-builds.html#lahey-linux
  http://www.unidata.ucar.edu/packages/netcdf/other-builds.html#lahey-nt

where he apparently had success using -DpgiFortran instead of
-DNAGf90Fortran.  The information is in our "Other Builds of the
netCDF Package" rather than the standard install docs, because we
don't maintain a version of the Lahey Fortran compile environment here
to use in testing.  I've added your note there as well.

--Russ

_____________________________________________________________________

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