[Fwd: 02-ac-lang-program-fortran-77.patch]

[Paul Martinolich]

Sorry!

Paul Martinolich wrote:
> 
> "Lars J. Aas" wrote:
> >
> > On Wed, Nov 08, 2000 at 06:12:18PM -0600, Paul Martinolich wrote:
> > : Akim Demaille wrote:
> > : > But indeed I overlooked the problem of the extension :(.
> > : >
> > : > Still, how do we want to neutralize this issue.  Do we want to keep
> > : > the GNUness test inside AC_PROG_F77 or not?
> > :
> > : With the extension change, then I think that the GNUness test should
> > : stay since this is GNU autoconf.
> > :
> > : My only reservation with the extension is what do we do when some
> > : unknown
> > : Fortran compiler won't compile code with a '.F' (assuming that is the
> > : one used).  I guess we can cross that bridge when we find one.
> >
> > Autoconf must probably soon be extended to use compiler wrapper scripts.
> > Dealing with unsupported file-extensions is really the domain of those
> > wrappers, so is it worth it to complicate Autoconf with that right now
> > when it will likely be removed again later?
> >
> >   Lars J
> 
> If the compiler wrapper scripts are to be in place before 2.50, then you
> are right.  But for this particular case, determining the GNUness of the
> Fortran compiler, Autoconf is currently and has in the past used an
> "unsupported" file-extension, like '.F'.  Autoconf 2.14 used the '.fpp'
> extension and G77 documents it.
> 
> Of course, I am biased to '.f' as being the standard way of indicating
> Fortran source because I've been doing UNIX for so long.  But, I seem to
> remember DOS using '.FOR' and Vax and Gould using '.F' (but that is
> really
> cloudy).
> 
> Will the compiler wrapper scripts be a major change?
> 
> Paul

-- 
I don't know Marge.  Trying is the first step to failure.
                                          -- Homer