From bug-octave-request at bevo dot che dot wisc dot edu  Wed Jul 14 12:31:18 1999
Subject: [jwe at bevo dot che dot wisc dot edu: Crashing after complex division by zero]
From: craig at jcb-sc dot com
To: bug-octave at bevo dot che dot wisc dot edu
Date: Wed, 14 Jul 1999 12:31:10 -0500 (CDT)


Enclosed is a bug report against libF77.  I'm not sure how valid the
complaint is, or what all the implications of making the recommended
change are.  I'm pretty sure we won't be making any g77-specific patches
to libf2c (in what we call libg2c) anymore, except in specific instances,
such as when we're pretty sure you're going to make the same bug-fix
shortly anyway -- definitely not the case here!

Also, I'll note that, as of gcc 2.95 aka g77 0.5.25, libg2c's c_div and
z_div routines aren't used by default anymore.  But they're still used
if the user compiles with the `-Os' (optimize for space) option, so we're
still interested in keeping them current vis-a-vis netlib libf2c.

        tq vm, (burley)


------- Start of forwarded message -------
Mailing-List: contact egcs-bugs-help at egcs dot cygnus dot com; run by ezmlm
Precedence: bulk
Sender: owner-egcs-bugs at egcs dot cygnus dot com
From: "John W. Eaton" <jwe at bevo dot che dot wisc dot edu>
Content-Type: text/plain; charset=us-ascii
Date: Wed, 14 Jul 1999 11:21:34 -0500 (CDT)
To: Thomas Hoffmann <hoffmann at ehmgs2 dot et dot tu-dresden dot de>
Cc: bug-octave at bevo dot che dot wisc dot edu, egcs-bugs@egcs.cygnus.com
Subject: Crashing after complex division by zero
In-Reply-To: <199907141202 dot OAA07624 at ehmgs2 dot et dot tu-dresden dot de>
X-UIDL: 54c7c006e30bc79c1a63b6b14c7950e7

[The following bug was reported for Octave 2.0.14, but it is really a
problem with libg2c, so I'm also sending this report to the egcs-bug
list.  --jwe]

On 14-Jul-1999, Thomas Hoffmann <hoffmann at ehmgs2 dot et dot tu-dresden dot de> wrote:

| Bug report for Octave 2.0.14 configured for hppa1.1-hp-hpux10.20
| 
| Description:
| -----------
| octave:2> b=[0 i 
| > 0 0]
| b =
| 
|   0 + 0i  0 + 1i
|   0 + 0i  0 + 0i
| 
| octave:3> inv(b)
| complex division by zero
| Abort(coredump)
| 
| Repeat-By:
| ---------
| 
| b=[0 i 
| > 0 0]
| 
| inv(b)

This problem is due to a bug in the complex division routine used by
g77.  The following patch will make it behave the same as the complex
division code used by g++ (which was adapted from the libf2c code).

The code in c_div.c is not relevant to Octave, but it has the same
problem.

To take advantage of this patch with Octave, you will have to
recompile the libg2c library from egcs then relink Octave.

Thanks,

jwe


Wed Jul 14 09:58:23 1999  John W. Eaton  <jwe at bevo dot che dot wisc dot edu>

	* libF77/z_div.c (z_div): Don't call sig_die for division by zero.
	* libF77/c_div.c (c_div): Likewise.


*** egcs-1.1.2/libf2c/libF77/z_div.c	1999/07/14 14:54:15	1.1
- --- egcs-1.1.2/libf2c/libF77/z_div.c	1999/07/14 14:55:00
***************
*** 18,25 ****
  	abi = - abi;
  if( abr <= abi )
  	{
- - 	if(abi == 0)
- - 		sig_die("complex division by zero", 1);
  	ratio = b->r / b->i ;
  	den = b->i * (1 + ratio*ratio);
  	res.r = (a->r*ratio + a->i) / den;
- --- 18,23 ----

*** egcs-1.1.2/libf2c/libF77/c_div.c	1999/07/14 14:54:19	1.1
- --- egcs-1.1.2/libf2c/libF77/c_div.c	1999/07/14 14:54:52
***************
*** 19,26 ****
  	abi = - abi;
  if( abr <= abi )
  	{
- - 	if(abi == 0)
- - 		sig_die("complex division by zero", 1);
  	ratio = (double)b->r / b->i ;
  	den = b->i * (1 + ratio*ratio);
  	res.r = (a->r*ratio + a->i) / den;
- --- 19,24 ----
------- End of forwarded message -------



---------------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.  To ensure
that development continues, see www.che.wisc.edu/octave/giftform.html
Instructions for unsubscribing: www.che.wisc.edu/octave/archive.html
---------------------------------------------------------------------