GNOME Bugzilla – Bug 394024
crash in giop_send_buffer_append_string
Last modified: 2007-01-13 03:13:32 UTC
What were you doing when the application crashed? I was testing out the Gnucash program in order to see if I could get a crash on exit as described by anouther development user. I was able to create an account and then exit. This failure happened on failure. I will get the debuginfo packages installed for more details later for this report. Distribution: Fedora Core release 6 (Rawhide) Gnome Release: 2.17.2 2006-11-07 (Red Hat, Inc) BugBuddy Version: 2.17.3 System: Linux 2.6.19-1.2904.fc7 #0 SMP Tue Jan 2 00:25:46 EST 2007 i686 X Vendor: The X.Org Foundation X Vendor Release: 70101000 Selinux: No Accessibility: Enabled Memory status: size: 102215680 vsize: 0 resident: 102215680 share: 0 rss: 43499520 rss_rlim: 0 CPU usage: start_time: 1168203026 rtime: 0 utime: 1811 stime: 0 cutime:1623 cstime: 0 timeout: 188 it_real_value: 0 frequency: 11 Backtrace was generated from '/usr/bin/gnucash' (no debugging symbols found) Using host libthread_db library "/lib/libthread_db.so.1". (no debugging symbols found) [Thread debugging using libthread_db enabled] [New Thread -1208588592 (LWP 3368)] 0x00390402 in __kernel_vsyscall ()
+ Trace 100267
Thread 1 (Thread -1208588592 (LWP 3368))
----------- .xsession-errors (12 sec old) --------------------- (gnucash:3368): GLib-CRITICAL **: Invalid key name: Show Hidden (gnucash:3368): GLib-CRITICAL **: Invalid key name: Show Zero Total (gnucash:3368): GLib-CRITICAL **: Invalid key name: Selected Account (gnucash:3368): GLib-CRITICAL **: Invalid key name: Open Account 1 (gnucash:3368): GLib-CRITICAL **: Invalid key name: Number of Open Accounts GTK Accessibility Module initialized Bonobo accessibility support initialized GTK Accessibility Module initialized Bonobo accessibility support initialized "/usr/bin/gnucash": not in executable format: File format not recognized --------------------------------------------------
Installed debuginfo package and ran again. Debuginfo in-line. Distribution: Fedora Core release 6 (Rawhide) Gnome Release: 2.17.2 2006-11-07 (Red Hat, Inc) BugBuddy Version: 2.17.3 System: Linux 2.6.19-1.2904.fc7 #0 SMP Tue Jan 2 00:25:46 EST 2007 i686 X Vendor: The X.Org Foundation X Vendor Release: 70101000 Selinux: No Accessibility: Enabled Memory status: size: 80408576 vsize: 0 resident: 80408576 share: 0 rss: 43659264 rss_rlim: 0 CPU usage: start_time: 1168203915 rtime: 0 utime: 1194 stime: 0 cutime:1110 cstime: 0 timeout: 84 it_real_value: 0 frequency: 11 Backtrace was generated from '/usr/bin/gnucash' Using host libthread_db library "/lib/libthread_db.so.1". [Thread debugging using libthread_db enabled] [New Thread -1209092400 (LWP 22838)] 0x00f0b402 in __kernel_vsyscall ()
+ Trace 100273
Thread 1 (Thread -1209092400 (LWP 22838))
----------- .xsession-errors (16 sec old) --------------------- (gnucash:22838): GLib-CRITICAL **: Invalid key name: Page Name (gnucash:22838): GLib-CRITICAL **: Invalid key name: Register Type (gnucash:22838): GLib-CRITICAL **: Invalid key name: Account Name (gnucash:22838): GLib-CRITICAL **: Invalid key name: Register Style (gnucash:22838): GLib-CRITICAL **: Invalid key name: Double Line Mode GTK Accessibility Module initialized Bonobo accessibility support initialized GTK Accessibility Module initialized Bonobo accessibility support initialized "/usr/bin/gnucash": not in executable format: File format not recognized --------------------------------------------------
NOTE: someone mentioned ORBit2 possibly being the problem on a test list for Fedora.
Thanks for this detailed bugreport. Which version of ORBit2 do you have installed? The crash seems to be caused in ORBit's giop-send-buffer.c:515 where the strlen() argument isn't checked for NULL. Changing product accordingly.
I have ORBit2-2.14.4-1.fc7 version installed. There is a bug report where downgrading the ORBit2 version stops the launching of bug buddy. I have not downgraded the package. Tom tested upgrading the package again and the error returned. Related report below. http://bugzilla.gnome.org/show_bug.cgi?id=390362 Downstream report below: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=221778
Sending NULL strings via CORBA is illegal / not-possible. So - this is in fact expected - although, prolly we should throw a nice exception rather than segv'ing ;-) The app passing this data needs fixing. In this case - it seems the a11y bridge.
The several apps where I have seen the problem now do not segfault after at-spi was fixed. I agree that the applications need fixing since it happened due to one package across the whole system. bug 364839 influenced the crash and its repair hid this problem once again.
*** This bug has been marked as a duplicate of 364839 ***