Return to BSD News archive
Newsgroups: comp.os.386bsd.bugs
Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!msuinfo!agate!news.ucdavis.edu!csus.edu!netcom.com!netcom20!ceb
From: ceb@netcom20.netcom.com (Ch. Buckley)
Subject: Re: Please Help with FORTRAN
In-Reply-To: john@physiol.su.OZ.AU's message of Fri, 7 Oct 1994 22:25:26 GMT
Message-ID: <CEB.94Oct9224050@netcom20.netcom.com>
Sender: ceb@netcom.com (Ch. Buckley)
Organization: Mauto, Palo Alto
References: <1994Sep27.233949.26036@cae.daikin.co.jp> <36kqbs$5fe@pdq.coe.montana.edu>
<1994Oct7.222526.6011@physiol.su.OZ.AU>
Date: Mon, 10 Oct 1994 05:40:38 GMT
Lines: 38
In article <1994Oct7.222526.6011@physiol.su.OZ.AU> john@physiol.su.OZ.AU (John Mackin) writes:
In article <36kqbs$5fe@pdq.coe.montana.edu>,
nate@bsd.coe.montana.edu (Nate Williams) writes:
> It's possible that the version of f2c that FreeBSD is using is older and
> has bugs in it. It should be updated to a newer version.
It's not difficult to get even the fractured version of f2c that's on
netlib (an e-mail oriented archive last time I checked -- and I
sincerely hope it's no longer true), and beat it into submission for
FreeBSD. I've run some pretty wierd stuff that way, and had few
difficulties.
I believe that Jonas Olsson & Rich Murphy did a more or less official
port, which was quite thorough, though it had one or two major gaffes
which made merging their changes with the latest version (FSF have now
taken f2c under their wing and fixed some highly scary errors that I
never caught) hard. It's probably haunting the ports hierarchy at
cdrom.com. For the latest gnu version, go to prep.
I have a version that has worked adequately for my needs, but Jordan
Hubbard's package creation program hates me ;'{, and every time I use
it I have to go and do major housecleaning on my filesystem. I won't
try to put out my stuff in package form until that's fixed. (He knows
about it, and has acknowledged it must be fixed -- this is not one of
those invidious backhanded nags.)
One question: the original poster mentioned "g++". Whaaat?
We _certainly_ didn't try to link Fortran and C++... as
Mary says in _The American Way,_ `I didn't know they could.'
Well, g++ is gcc called a certain way, which is what I think the
original poster was getting at. As I recall, the trick here was to
get f2c to put out ANSI output (-A), and life got a lot easier fast.
--