[vtkusers] segfault Delete()ing vtkInterpolateDataSetAttributes

[tom fogal]

I believe I am closer to tracking this down. Below is a gdb backtrace
where gdb gives an interesting message:

warning: can't find linker symbol for virtual table for `vtkDataObject'
value

Also I believe this may be (very vaguely) related to bug 189,

http://www.itk.org/Bug/bug.php?op=vote&bugid=189

due to the fact that they both crash in the same part of libc.

>From what I gather through extensive googling, main_arena is apparently
some type of flag telling glibc where a particular section of memory
was malloc()d -- certain applications implement their own malloc() and
that is not considered part of the 'main_arena', for example.

Perhaps a linking error is causing
vtkDataObject::GetNetReferenceCount() to not be found in this context?
This is a little beyond my current C/C++ debugging knowledge.

Any ideas would be GREATLY appreciated.
Also, should I perhaps submit this as a bug or should we continue to
blame glibc until more information is known?

-tom

---- gdb trace ----
tfogal at thetis ggcm_anim $ gdb ggcm_animator 
GNU gdb 6.1
Copyright 2004 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and
you are
welcome to change it and/or distribute copies of it under certain
conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB.  Type "show warranty" for
details.
This GDB was configured as "x86_64-suse-linux"...Using host
libthread_db library "/lib64/tls/libthread_db.so.1".

(gdb) run --config jxr.cfg 
Starting program: /home/tfogal/ggcm/ggcm_anim/ggcm_animator --config
jxr.cfg
[Thread debugging using libthread_db enabled]
[New Thread 182953122592 (LWP 31484)]
rendering [4] streamlines
rendering... 0% complete
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 182953122592 (LWP 31484)]
0x0000002a9686ac18 in main_arena () from /lib64/tls/libc.so.6
(gdb) bt
#0  0x0000002a9686ac18 in main_arena () from /lib64/tls/libc.so.6
#1  0x0000002a958285b2 in vtkSource::UnRegister (this=0x52c980, o=0x0)
at /home/tfogal/tarballs/VTK/Common/vtkSource.cxx:740
#2  0x0000002a957d704f in vtkObjectBase::Delete (this=0x52c980) at
/home/tfogal/tarballs/VTK/Common/vtkObjectBase.cxx:86
#3  0x000000000040bebc in ~GGCMInterpolateDataSet (this=0x52c960) at
GGCMInterpolateDataSet.cc:39
#4  0x000000000040aa8d in GGCMAnimator::Interpolate (this=0x7fbfffef70,
time=0.02) at GGCMAnimator.cc:207
#5  0x000000000040a404 in GGCMAnimator::Run (this=0x7fbfffef70) at
GGCMAnimator.cc:132
#6  0x000000000040ed59 in main (argc=3, argv=0x7fbffff368) at
main.cc:58
(gdb) up
#1  0x0000002a958285b2 in vtkSource::UnRegister (this=0x52c980, o=0x0)
at /home/tfogal/tarballs/VTK/Common/vtkSource.cxx:740
740             if (this->Outputs[idx]->GetNetReferenceCount() != 1)
(gdb) list
735             {
736             if (this->Outputs[idx] == o)
737               {
738               done = 0;
739               }
740             if (this->Outputs[idx]->GetNetReferenceCount() != 1)
741               {
742               done = 0;
743               }
744             }
(gdb) print this
$1 = (vtkSource * const) 0x52c980
(gdb) print idx
$2 = 0
(gdb) print this->Outputs[0]
$3 = (class vtkDataObject *) 0x5b380d0
(gdb) print this->Outputs[0]->GetNetReferenceCount()
warning: can't find linker symbol for virtual table for `vtkDataObject'
value

Program received signal SIGSEGV, Segmentation fault.
0x0000002a9686ac18 in main_arena () from /lib64/tls/libc.so.6
The program being debugged was signaled while in a function called from
GDB.
GDB remains in the frame where the signal was received.
To change this behavior use "set unwindonsignal on"
Evaluation of the expression containing the function (main_arena) will
be abandoned.

------- Forwarded Message
Subject: [vtkusers] segfault Delete()ing vtkInterpolateDataSetAttributes

I'm getting a weird segfault in libc when I try to delete an object of
this type:

(gdb) run --config jxr.cfg 
Starting program: /home/tfogal/ggcm/ggcm_anim/ggcm_animator --config
jxr.cfg
[Thread debugging using libthread_db enabled]
[New Thread 182953122592 (LWP 17616)]
rendering [4] streamlines
rendering... 0% complete
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 182953122592 (LWP 17616)]
0x0000002a9686ac18 in main_arena () from /lib64/tls/libc.so.6
(gdb) bt
#0  0x0000002a9686ac18 in main_arena () from /lib64/tls/libc.so.6
#1  0x0000002a958285b2 in vtkSource::UnRegister (this=0x52c980, o=0x0)
    at /home/tfogal/tarballs/VTK/Common/vtkSource.cxx:740
#2  0x0000002a957d704f in vtkObjectBase::Delete (this=0x52c980)
    at /home/tfogal/tarballs/VTK/Common/vtkObjectBase.cxx:86
#3  0x000000000040bebc in ~GGCMInterpolateDataSet (this=0x52c960)
    at GGCMInterpolateDataSet.cc:39
#4  0x000000000040aa8d in GGCMAnimator::Interpolate (this=0x7fbfffef70, 
    time=0.02) at GGCMAnimator.cc:207
#5  0x000000000040a404 in GGCMAnimator::Run (this=0x7fbfffef70)
    at GGCMAnimator.cc:132
#6  0x000000000040ed15 in main (argc=3, argv=0x7fbffff368) at
main.cc:58

The line in vtkSource::UnRegister is 740, or

if (this->Outputs[idx]->GetNetReferenceCount() != 1)

The 'o' is a NULL pointer, but I should think this is expected. In
vtkObjectBase::Delete there is the single line:

this->UnRegister((vtkObjectBase *)NULL);

Should the NULL be a 'this' pointer? I don't understand how the above
code would work for any object, but I have been using this version of
vtk (4.2) for a while now.

The stranger thing is that this is a rewrite of a poorly written
program which performs the same task, and that app works (well...)
fine.

Any ideas? I'm lost...

- -tom
_______________________________________________
This is the private VTK discussion list. 
Please keep messages on-topic. Check the FAQ at: <http://public.kitware.com/cgi-bin/vtkfaq>
Follow this link to subscribe/unsubscribe:
http://www.vtk.org/mailman/listinfo/vtkusers

------- End of Forwarded Message