Bug 651248

Summary: [abrt] gpodder-2.8-2.fc14: node_get_last: Process /usr/bin/python was killed by signal 11 (SIGSEGV)
Product: [Fedora] Fedora Reporter: Jon Dufresne <jon.dufresne>
Component: gtk2Assignee: Matthias Clasen <mclasen>
Status: CLOSED DUPLICATE QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: medium Docs Contact:
Priority: low    
Version: 14CC: dmalcolm, ivazqueznet, james.antill, jonathansteffan, jspaleta, mclasen, vpvainio
Target Milestone: ---   
Target Release: ---   
Hardware: x86_64   
OS: Unspecified   
Whiteboard: abrt_hash:b7c260807d594f761726560faf0ebdf6d09a9c96
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2010-11-10 09:16:23 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
File: backtrace none

Description Jon Dufresne 2010-11-09 04:20:15 UTC 
abrt version: 1.1.13
architecture: x86_64
Attached file: backtrace
cmdline: /usr/bin/python /usr/bin/gpodder
component: gpodder
crash_function: node_get_last
executable: /usr/bin/python
kernel: 2.6.35.6-48.fc14.x86_64
package: gpodder-2.8-2.fc14
rating: 4
reason: Process /usr/bin/python was killed by signal 11 (SIGSEGV)
release: Fedora release 14 (Laughlin)
time: 1289275953
uid: 500

How to reproduce
-----
It is not clear to me what I was doing when this happend. If I recall,  I was clicking around in gpodder quickly changing values and states of a large selection of a podcast's episodes.
Comment 1 Jon Dufresne 2010-11-09 04:20:18 UTC 
Created attachment 458964 [details]
File: backtrace
Comment 2 Jon Dufresne 2010-11-09 04:23:09 UTC 
Although gpodder was running when this bug occurred, the backtrace would suggest that python itself segfaulted so changing component.

$ rpm -qa python
python-2.7-8.fc14.1.x86_64
Comment 3 Dave Malcolm 2010-11-09 16:31:00 UTC 
Thank you for reporting this bug.

How reproducible is this problem?  If you run the program from a terminal, is an error message printed?

What's the output of running:
  $ rpm -qa gtk2 pygtk2

Looking at the backtrace, it looks like the problem occurred in thread #1 in node_get_last, deep within GTK's tree view code.  From frame #20 it looks like it was responding to a row deletion event, which called a handler written in Python (frame #19 down to frame #6), which called back into GTK to get a value for column 3 from a GtkListStore (frame #5 down to frame #0).

The "iter" value might be corrupt, but unfortunately, it's been optimized out so that debugger can't tell us what it was.

The disassembly shows the SIGSEGV happened at:

  => 0x00000030e5c5cbbb <+43>:	mov    0x18(%rdi),%rax
with %rdi = 0xffffffe700000000

My hunch is that we're seeing a int32 vs int64 casting mismatch of an GtkTreeIter somewhere in either gtk2, pygtk2 or gpodder (does the latter contain any .c code?) - though this _is_ a hunch.

Reassigning component from "python" to "gtk2";  hopefully the gtk2 maintainer will be able to figure this out further or reassign as necessary.
Comment 4 Jon Dufresne 2010-11-09 16:53:54 UTC 
$ rpm -qa gtk2 pygtk2
pygtk2-2.17.0-7.fc14.x86_64
gtk2-2.22.0-1.fc14.1.x86_64


I have not been able to reproduce the bug through normal use. I recall when the bug happened I had gpodder open on a podcast with many episodes. In an effort to remove the "delete" status from all the episodes I selected the entire list, set it to "new" status then selected the entire list and unselected "new" status.

Doing the above isn't enough to reproduce the bug.
Comment 5 Ville-Pekka Vainio 2010-11-10 09:16:23 UTC 
This seems to be another report of bug #583450.

Dave: gPodder is pure Python. Thanks for looking into this. I'll mark this bug as a duplicate.

*** This bug has been marked as a duplicate of bug 583450 ***