From octave-maintainers-request at bevo dot che dot wisc dot edu  Tue Nov 19 14:35:48 1996
Subject: endlos loop with segmentation fault
From: "John W. Eaton" <jwe at bevo dot che dot wisc dot edu>
To: Friedrich Leisch <leisch at ci dot tuwien dot ac dot at>
Cc: Andreas dot Weingessel at ci dot tuwien dot ac dot at, Kurt.Hornik@ci.tuwien.ac.at,         octave-maintainers@bevo.che.wisc.edu
Date: Tue, 19 Nov 1996 14:34:06 -0600

On 19-Nov-1996, Friedrich Leisch <leisch at ci dot tuwien dot ac dot at> wrote:

: it happenend already two times at our department that somebody wanted
: to quit octave and got caught in an infinite loop with octave messages
: "Segmentation fault ... trying to stop myself" rushing over the screen. 

Here's a patch that might be able to prevent the endless loop:

Index: sighandlers.cc
===================================================================
RCS file: /home/jwe/src/master/octave/src/sighandlers.cc,v
retrieving revision 1.44
diff -c -r1.44 sighandlers.cc
*** sighandlers.cc	1996/11/14 08:39:47	1.44
--- sighandlers.cc	1996/11/19 20:30:59
***************
*** 24,30 ****
--- 24,32 ----
  #include <config.h>
  #endif
  
+ #include <cstdlib>
  #include <csignal>
+ 
  #include <new>
  
  #include <iostream.h>
***************
*** 88,98 ****
  static void
  my_friendly_exit (const char *sig_name, int sig_number)
  {
!   error ("%s -- stopping myself...", sig_name);
  
!   save_user_variables ();
  
!   clean_up_and_exit (sig_number);
  }
  
  // I know, not really a signal handler.
--- 90,117 ----
  static void
  my_friendly_exit (const char *sig_name, int sig_number)
  {
!   static bool been_there_done_that = false;
! 
!   if (been_there_done_that)
!     {
! #ifdef SIGABRT
!       octave_set_signal_handler (SIGABRT, SIG_DFL);
! #endif
! 
!       error ("attempted clean up seems to have failed -- aborting...");
! 
!       abort ();
!     }
!   else
!     {
!       been_there_done_that = true;
! 
!       error ("%s -- stopping myself...", sig_name);
  
!       save_user_variables ();
  
!       clean_up_and_exit (sig_number);
!     }
  }
  
  // I know, not really a signal handler.


If you happen to find a way to repeat this, I'd definitely like to be
able to fix the problem that is actually causing the segfault.

Thanks,

jwe