Dan:

f2c:  I have f2c, but I believe it is a bit slower than g77, though I could
be corrected.  I have not tried f2c under NetBSD 1.5 yet.

I have compiled this model with g77 under mklinux on the mac, and found no
difficulty.  The problem is confusing.  I have previous model that compiled
and ran with g77 under macppc and i386 NetBSD without difficulty, which
suggests that this problem will not be easy to find.

Jim


-----Original Message-----
From: mcmahill@mtl.mit.edu <mcmahill@mtl.mit.edu>
To: jnelson@wiai.com <jnelson@wiai.com>
Cc: port-macppc@netbsd.org <port-macppc@netbsd.org>
Date: Wednesday, February 28, 2001 4:56 AM
Subject: Re: Fortran


>
>
>
>
>On Tue, 27 Feb 2001 jimtnelson@earthlink.net wrote:
>
>> I have a seismic forward modeling program running fine on an i386
>> machine, but has a big problem with NaN's on my mac 7500.  I think i
>> have an alignment problem, but can't tell.  I would like to detect
>> floating point exceptions.  I tried looked for the fpu_control.h include
>> file described in "info g77", but can not find it.  Is there a provision
>> for floating point exception handling for g77 on macppc?  If so, where
>> do I find it?
>>
>
>> More generally, g77 crashed on my during compiling with optimization set
>> to "O".  Are there big problems with g77 for macppc netbsd 1.5?
>
>I've not used g77 on macppc, but have found it somewhat buggy on some
>other non-i386 ports.  You might try the lang/f2c package.  That package
>includes a script 'f2c-f77' which does a pretty good job of acting just
>like a fortran compiler.  It should accept most of the same flags as g77.
>
>-Dan
>
>