[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: octave 3.4.0 on Scientific Linux 5
From: |
John W. Eaton |
Subject: |
Re: octave 3.4.0 on Scientific Linux 5 |
Date: |
Fri, 25 Feb 2011 10:58:45 -0500 |
On 25-Feb-2011, Michael Martins wrote:
| Thank you for the fast response. Unfortunately I am a little bit helpless
with
| this type of error. The machine is a Quadcore Dual Xeon systems running a
| 64bit SL5.
|
| All the best
| Michael
| -----
|
| Here is the backtrace
|
| GNU gdb Fedora (6.8-27.el5)
| Copyright (C) 2008 Free Software Foundation, Inc.
| License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
|
| [Thread debugging using libthread_db enabled]
| [New Thread 0x2b10bfb0dc40 (LWP 29863)]
|
| Program received signal SIGSEGV, Segmentation fault.
| 0x00002b10bd7cabe8 in sigemptyset () from
| /scratch/martins/octave-3.4.0/src/.libs/liboctinterp-3.4.0.so
| (gdb) bt
| #0 0x00002b10bd7cabe8 in sigemptyset () from
| /scratch/martins/octave-3.4.0/src/.libs/liboctinterp-3.4.0.so
| #1 0x00002b10bcd87ae1 in octave_set_signal_handler (sig=2,
| handler=0x2b10bcd88050 <sigint_handler>, restart_syscalls=true) at
| sighandlers.cc:225
| #2 0x00002b10bcd87baa in octave_catch_interrupts () at sighandlers.cc:460
| #3 0x00002b10bcd87c13 in install_signal_handlers () at sighandlers.cc:529
| #4 0x00002b10bcd5a547 in octave_main (argc=6, argv=0x7fffaddee7c8,
| embedded=0) at octave.cc:665
| #5 0x0000003897c1d994 in __libc_start_main () from /lib64/libc.so.6
| #6 0x0000000000400659 in _start ()
This appears to be the same as this problem:
https://savannah.gnu.org/bugs/?30685
As I have said in the comments to the bug report, I still have no clue
precisely what is causing this problem, and since I can't reproduce it
on any system I have, someone who can reproduce the problem is going
to have to do some debugging and analysis to find the cause.
Looking at this problem a little more now, I see that the crash is
apparently happening inside the gnulib sigemptyset function, which is
just
int
sigemptyset (sigset_t *set)
{
*set = 0;
return 0;
}
The call to this function is happening with
gnulib::sigemptyset (&act.sa_mask);
and my guess is that your definition of the sigaction struct has
sa_mask is not compatible with what the gnulib sigemptyset function is
expecting. So I think this is a gnulib bug.
At the point of the crash, what do you see if you execute the
following gdb commands?
(gdb) up
(gdb) ptype act.sa_mask
?
Please follow up by posting this info to the bug report linked above.
Is the gnulib sigprocmask file being compiled on your system? If so,
please add -save-temps to CFLAGS and compile this file and attach the
resulting sigprocmask.ii file to the bug report.
Can you also determine why the gnulib sigprocmask.c file is being
compiled? I don't see any information in the config.log file that you
posted that explains why it would be compiled on your system.
Thanks,
jwe