Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!samsung!munnari.oz.au!metro!news
From: glenn@suphys.physics.su.OZ.AU (Glenn Geers)
Newsgroups: comp.lang.fortran
Subject: Re: f2c on a 386 machine?
Message-ID: <1990Oct12.035354.12689@metro.ucc.su.OZ.AU>
Date: 12 Oct 90 03:53:54 GMT
References: <9010110815.AA09349@euler.Berkeley.EDU>
Sender: news@metro.ucc.su.OZ.AU (news)
Organization: Uni Computing Service, Uni of Sydney, Australia
Lines: 32

From article <9010110815.AA09349@euler.Berkeley.EDU>, by ndeng@EULER.BERKELEY.EDU:
> 
> I just wonder if anybody has adopted the f2c program (avail. from netlib) to
> a PC environment (Presumably 386 boxes), either under dos-extender, or os/2,
> or SCO Unix/Xenix? Can this be done? If anybody has done this, I believe many
> netters (especially in this group) will be interested in hearing your experience.


f2c works fine under SCO Xenix 2.3.2 and so presumably under SCO Unix (just 
use the Xenix binary :-) ). The C compiler I am using is gcc (with Steve
Bleazard's patches). The largest all fortran code I have compiled is spice2g.6
and that went through without a hitch. You do have to make a couple of minor
changes to the makefiles in libI77 and libF77 as the Xenix supplied /lib/ldr
does not support the -x option.

I'm currently working on some mixed language stuff for my PhD (C + fortran)
and I'm having no problem using f2c. The code passes 2-d arrays of structs
(complex numbers, ok) back and forth between C and fortran without trouble.

f2c is a *great* piece of work, but treat it as a *compiler* some of the C
code is really ugly.
						Cheers,
								Glenn

glenn@qed.physics.su.oz.au


--
Glenn Geers                       | "So when it's over, we're back to people.
Department of Theoretical Physics |  Just to prove that human touch can have
The University of Sydney          |  no equal."
Sydney NSW 2006 Australia         |  - Basia Trzetrzelewska, 'Prime Time TV'