soffice.bin crashed with SIGSEGV in *__GI___libc_free()

[Mikko Rantalainen]

Binary package hint: openoffice.org

Looking at the stacktrace, this may be double free() somewhere.

ProblemType: Crash
DistroRelease: Ubuntu 10.04
Package: openoffice.org-core 1:3.2.0-7ubuntu4.1
ProcVersionSignature: Ubuntu 2.6.31-11.154-rt
Uname: Linux 2.6.31-11-rt x86_64
Architecture: amd64
CheckboxSubmission: f0bf0101e3df07a87acfbc156f0db03d
CheckboxSystem: b5acb6c9ca4017b1d44043910f45329d
Date: Wed Sep 22 15:02:58 2010
Disassembly: => 0x7f31c32fced9: Cannot access memory at address 0x7f31c32fced9
ExecutablePath: /usr/lib/openoffice/program/soffice.bin
InstallationMedia: Ubuntu 9.10 "Karmic Koala" - Release Candidate amd64 (20091020.3)
ProcCmdline: /usr/lib/openoffice/program/soffice.bin -writer /tmp/webli_pedanet_interface_suggestion_v1.3.odt -splash-pipe=5
SegvAnalysis:
 Segfault happened at: 0x7f31c32fced9: Cannot access memory at address 0x7f31c32fced9
 PC (0x7f31c32fced9) not located in a known VMA region (needed executable region)!
SegvReason: executing unknown VMA
Signal: 11
SourcePackage: openoffice.org
StacktraceTop:
 ?? ()
 ?? ()
 ?? ()
 ?? ()
 *__GI___libc_free (mem=<value optimized out>)
Title: soffice.bin crashed with SIGSEGV in *__GI___libc_free()
UserGroups: adm admin cdrom dialout lpadmin plugdev sambashare

[Apport retracing service (apport)]

StacktraceTop:
 ?? ()
 ?? ()
 ?? ()
 ?? ()
 free () from /lib/libc.so.6

[Rolf Leggewie (r0lf)]

looks very similar to bug 627365. Does this happen in Maverick as well?

[João Pinto (joaopinto)]

Most of the times I was able to reproduce the crash quitting oocal after starting it with:
oocalc "-accept=socket,host=localhost,port=8100;urp;StarOffice.ServiceManager" -norestore -nofirststartwizard -nologo

I didn't experienced it on Maverick yet.

[Rolf Leggewie (r0lf)]

well, the log and backtrace isn't really helpful. So, we need a sure-fire way to reproduce this to proceed. I can't set this ticket to confirmed with the little valuable data there currently is.

[João Pinto (joaopinto)]

If you need to set it to incomplete then please provide instructions on what needs to be done to make it "complete".
Have you tried to reproduce the crash per my instructions on Lucid ?

[Rolf Leggewie (r0lf)]

I just came back to this ticket. Sorry, my bad, the status change was actually unintentional. Not sure, but maybe our messages crossed. I don't recall seeing your instructions for how you usually manage to reproduce this, for example.

FWIW, when I start oocalc the way you describe and then quit it, the program does not crash but it actually does not close, either.

[penalvch (penalvch)]

Mikko Rantalainen, thank you for reporting this bug and helping make Ubuntu better. This bug is unreproducible in Ubuntu 10.10 LibreOffice Calc at the Terminal:

localc "-accept=socket,host=localhost,port=8100;urp;StarOffice.ServiceManager" -norestore -nofirststartwizard -nologo

A blank Calc file opens with no crash. Does this work for you?

lsb_release -rd
Description: Ubuntu 10.10
Release: 10.10

sudo add-apt-repository ppa:libreoffice/ppa && sudo apt-get update && sudo apt-get -y upgrade && sudo apt-get -y install libreoffice-calc

 apt-cache policy libreoffice-calc
libreoffice-calc:
  Installed: 1:3.3.0-1maverick1
  Candidate: 1:3.3.0-1maverick1
  Version table:
 *** 1:3.3.0-1maverick1 0
        500 http://ppa.launchpad.net/libreoffice/ppa/ubuntu/ maverick/main i386 Packages
        100 /var/lib/dpkg/status

[Launchpad Janitor (janitor)]

[Expired for openoffice.org (Ubuntu) because there has been no activity for 60 days.]