scim-bridge crashed with SIGSEGV in scim::Module::unload()

Bug #199592 reported by Evan Rotert
76
This bug affects 7 people
Affects Status Importance Assigned to Milestone
scim-bridge (Ubuntu)
Confirmed
Medium
Arne Goetje
Hardy
Confirmed
Medium
Unassigned

Bug Description

Binary package hint: scim-bridge

Hardy Alpha 6- I get this error every time I start up. I don't know what causes it, as I've never touched SCIM's functions.

ProblemType: Crash
Architecture: i386
Date: Fri Mar 7 14:31:10 2008
DistroRelease: Ubuntu 8.04
ExecutablePath: /usr/bin/scim-bridge
NonfreeKernelModules: nvidia
Package: scim-bridge-agent 0.4.14-1ubuntu2
PackageArchitecture: i386
ProcCmdline: scim-bridge
ProcEnviron:
 PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games
 LANG=en_US.UTF-8
 SHELL=/bin/bash
Signal: 11
SourcePackage: scim-bridge
StacktraceTop:
 scim::Module::unload () from /usr/lib/libscim-1.0.so.8
 scim::Module::~Module () from /usr/lib/libscim-1.0.so.8
 ?? ()
 ?? ()
 ?? ()
Title: scim-bridge crashed with SIGSEGV in scim::Module::unload()
Uname: Linux 2.6.24-11-generic i686
UserGroups: adm admin audio cdrom dialout dip floppy fuse lpadmin plugdev video
SegvAnalysis:
 Segfault happened at: 0xb7e7c473 <_ZN4scim6Module6unloadEv+29>: mov (%edi),%eax
 PC (0xb7e7c473) ok
 source "(%edi)" (0x53454741) not located in a known VMA region (needed readable region)!
 destination "%eax" ok
SegvReason: reading unknown VMA

Tags: apport-crash
Revision history for this message
Evan Rotert (evan-rotert) wrote :
Revision history for this message
Apport retracing service (apport) wrote : Symbolic stack trace

StacktraceTop:~Module (this=0x53454741) at scim_module.cpp:132
ScimBridgeAgentImpl::finalize_scim (this=0x80752b8)
ScimBridgeAgentImpl::finalize (this=0x80752b8)
~ScimBridgeAgentImpl (this=0x80752b8)
ScimBridgeAgent::alloc () at scim-bridge-agent.cpp:186

Revision history for this message
Apport retracing service (apport) wrote : Symbolic threaded stack trace
Changed in scim-bridge:
importance: Undecided → Medium
Revision history for this message
Ming Hua (minghua) wrote :

Arne, it seems this crasher has dozens of duplicates now. Please investigate. Also please subscribe to scim-bridge's bugmail (or ask whoever requested scim-bridge's main inclusion to do so).

Arne Goetje (arnegoetje)
Changed in scim-bridge:
assignee: nobody → arnegoetje
status: New → In Progress
Revision history for this message
Arne Goetje (arnegoetje) wrote :

Why is this bug marked as private?

I cannot reproduce this crash at all. Neither on my local system with the latest packages installed, nor on the current Ubuntu Live CD from today.

What's going on here?

My system: Ubuntu Hardy with all latest packages, Locale is en_US, scim is enabled and scim-bridge selected as module. No crash at all.

Need more info, please.

Revision history for this message
Ming Hua (minghua) wrote :

Crasher bugs submitted by apport are marked as private by default, because there may be sensitive information (such as passwords) in the submitted data.

Revision history for this message
Tim (willc0de4food) wrote :

When it crashed on me, it was shortly after the default install. I downloaded the alpha 6 iso from the link on distrowatch yesterday (3/13)
I wasn't trying to use scim, but it was acting erratically, changing languages to arabic? The first language option available, but I use US english.
I submitted the bug report because I was asked to and then proceeded to uninstall scim.

Revision history for this message
Mikael Gerdin (mgerdin) wrote :

System: Kubuntu Hardy i386, upgraded today.
It seems that installing the "scim" package fixed this problem for me, at least partially.
Now when "scim-bridge" is first started it doesn't crash and it starts the gtk-based panel applet (since it's included in the "scim"-package.
However when scim-bridge is executed again, when the scim-processes are running, scim-bridge randomly crashes with SIGSEGV.

Revision history for this message
Wenzhuo Zhang (wenzhuo) wrote :

Arne, you can reproduce the segfault by setting /FrontEnd/X11/Dynamic = true.

Revision history for this message
Wenzhuo Zhang (wenzhuo) wrote :

Setting /FrontEnd/X11/Dynamic to true is a workaround provided by Ming Hua for bug #66104.

Revision history for this message
Wenzhuo Zhang (wenzhuo) wrote :

I applied the above workaround in Gutsy. After upgrading to Hardy beta, scim-bridge segfaulted in my old account. Having noticed that it does not segfault in a newly created account in Hardy, I changed /FrontEnd/X11/Dynamic back to false in my old account, and scim-bridge no longer segfaulted. However, changing the option back to true can no longer reproduce the segfault.

Revision history for this message
Wenzhuo Zhang (wenzhuo) wrote :

I just saw a segfault even though /FrontEnd/X11/Dynamic is set to false. Now I cannot reliably reproduce the crash. Here are my findings:

1. scim-bridge never crashes upon Gnome login in a newly created account.
2. If a scim-bridge process is already running, start a second instance always segfaults.
3. scim-bridge always crashes with configuration left behind by Gutsy in an old account, especifically with /FrontEnd/X11/Dynamic being set to true.
4. Changing /FrontEnd/X11/Dynamic back to false can make the segfaults happen less often.
5. After copying /etc/scim/config to replace the .scim/config (in a virtual console and with Gnome logged out), it seems scim-bridge no longer crashes (restarted Gnome for several times already).

Revision history for this message
Wenzhuo Zhang (wenzhuo) wrote :

Attaching the scim config file left behind by Gutsy. However, it has already been touched by scim in Hardy.

Revision history for this message
Wenzhuo Zhang (wenzhuo) wrote :

I renamed .scim to .scim.bak three days ago, and scim-bridge crashes reappeared yesterday. Rebooted my computer a couple of time in between. Therefore, scim-bridge crashes in a newly created account.

Revision history for this message
PeterShinners (pete-shinners) wrote :

I'm getting this on startup after converting from xubuntu to kubuntu

Revision history for this message
hardhu (qzerty) wrote :

I have also this problem in Intrepid beta when I start a kde session; I have also Gnome installed and no problems instead with it

Revision history for this message
Adric Riedel (adric-riedel) wrote :

I get this error every time I log on now; I had xubuntu-desktop installed at one point. This error seems to drag down all of KDE with it; I get hundreds of scim instances, and if I try to kill any of them Konqueror crashes... but nothing else will load until I do. As soon as all the scim instances are killed everything I clicked on abruptly loads. I have no real idea what's going on, except that I was trying to manually remove XFCE and Gnome apps so I could have a pure-er KDE install.

Arne Goetje (arnegoetje)
Changed in scim-bridge:
assignee: arnegoetje → nobody
status: In Progress → Confirmed
assignee: arnegoetje → nobody
status: In Progress → Confirmed
Revision history for this message
Ostien (ostiencf) wrote :

Adric, I have had a slimier issue. I get the mass amount of scim segfaults (although for me its like 20) I wanted to try GNOME so I installed it along side KDE and found I didin't like it so I manually uninstalled gnome to get back to a pure KDE install. For me the problem is that programs load much slower, especially any terminal program.

Also what is the scim package I'm not sure if I need it or not.

Revision history for this message
Ostien (ostiencf) wrote :

Adric, try reinstalling the scim package, if you ran into the problem the same way I did this could be your issue. I did that and it worked! no more delay, everything is faster now like it used to be.

Revision history for this message
ZhengPeng Hou (zhengpeng-hou) wrote :
Kees Cook (kees)
description: updated
Revision history for this message
Rick Spencer (rick-rickspencer3) wrote :

What shall we do with this bug? Is there a fix available?

Changed in scim-bridge (Ubuntu):
assignee: nobody → Arne Goetje (arnegoetje)
Revision history for this message
Rolf Leggewie (r0lf) wrote :

Rick, I'm currently only aware of "rm -Rf ~/.scim/" that the user needs to invoke to reset scim configuration. I don't think there is anything that the packagers could release to deal with this. Unfortunately, the chances of something appearing are not too bright, either.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.