Figuring out how to use gfortran

Arnaud Desitter arnaud.desitter@ouce.ox.ac.uk
Fri Aug 6 09:48:00 GMT 2004


----- Original Message ----- 
From: "Steve Kargl" <sgk@troutmask.apl.washington.edu>
Newsgroups: gmane.comp.gcc.fortran
Cc: <fortran@gcc.gnu.org>
Sent: Friday, August 06, 2004 6:31 AM
Subject: Re: Figuring out how to use gfortran


> On Thu, Aug 05, 2004 at 09:16:44PM -0700, Jerry DeLisle wrote:
> > On Thu, 2004-08-05 at 21:10, Steve Kargl wrote:
> > > On Thu, Aug 05, 2004 at 08:56:43PM -0700, Jerry DeLisle wrote:
> > > > Well,
> > > >
> > > > I have gfortran running.  Where can I find gfortran compiler options
for
> > > > different dialects listed.  I have some sample f90 programs that
seem to
> > > > give errors.
> > > >
> > >
> > > what are the errors?
> > >
> > [jerry@quantum fortran]$ foogfortran 1000d.f -o x1000d
> > /tmp/ccqPypK7.o(.text+0x6f): In function `MAIN__':
> > : undefined reference to `dfloat_'
> > /tmp/ccqPypK7.o(.text+0x83): In function `MAIN__':
> > : undefined reference to `dfloat_'
> > /tmp/ccqPypK7.o(.text+0x9f): In function `MAIN__':
> > : undefined reference to `dfloat_'
> > collect2: ld returned 1 exit status
> >
>
> DFLOAT is a nonstandard intrinsic, but many compilers
> support it.  Can you file a bug report to support DFLOAT
> as an extension?  To move forward with your code, you can
> make a global substitution of DFLOAT to DREAL.  Note, DREAL
> is also an extension, but gfortran already supports it.
> The correct Fortran 90/95 fix is to change DFLOAT(X) to
> REAL(X,KIND(1.d0)).

Using DBLE(x), which is standard, is much easier.

Regards,



More information about the Fortran mailing list