dasher crashed with SIGSEGV in Dasher::CDasherModel::HandleOutput()

Bug #637747 reported by las
14
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Dasher
Expired
Critical
dasher (Ubuntu)
Incomplete
Medium
Unassigned

Bug Description

Binary package hint: dasher

tried it on gedit & it crashed

ProblemType: Crash
DistroRelease: Ubuntu 10.10
Package: dasher 4.11-1
ProcVersionSignature: Ubuntu 2.6.35-20.29-generic 2.6.35.4
Uname: Linux 2.6.35-20-generic i686
NonfreeKernelModules: nvidia
Architecture: i386
Date: Tue Sep 14 08:52:04 2010
ExecutablePath: /usr/bin/dasher
InstallationMedia: Ubuntu 10.04 "Lucid Lynx" - Alpha i386 (20100215)
ProcCmdline: dasher --appstyle=direct
ProcEnviron:
 SHELL=/bin/bash
 LANG=en_US.utf8
 LANGUAGE=en_GB:en
SegvAnalysis:
 Segfault happened at: 0x8097cb6 <_ZN6Dasher12CDasherModel12HandleOutputEPSt6vectorINS_10SymbolProbESaIS2_EEPi+102>: mov (%eax),%edx
 PC (0x08097cb6) ok
 source "(%eax)" (0x00000000) not located in a known VMA region (needed readable region)!
 destination "%edx" ok
SegvReason: reading NULL VMA
Signal: 11
SourcePackage: dasher
StacktraceTop:
 Dasher::CDasherModel::HandleOutput(std::vector<Dasher::SymbolProb, std::allocator<Dasher::SymbolProb> >*, int*) ()
 Dasher::CDasherModel::OneStepTowards(long long, long long, unsigned long, std::vector<Dasher::SymbolProb, std::allocator<Dasher::SymbolProb> >*, int*) ()
 Dasher::CDefaultFilter::Timer(int, Dasher::CDasherView*, Dasher::CDasherModel*, std::vector<Dasher::SymbolProb, std::allocator<Dasher::SymbolProb> >*, int*, Dasher::CExpansionPolicy**) ()
 Dasher::CDasherInterfaceBase::NewFrame(unsigned long, bool) ()
 CDasherControl::TimerEvent() ()
Title: dasher crashed with SIGSEGV in Dasher::CDasherModel::HandleOutput()
UserGroups: adm admin audio cdrom dialout dip fax floppy fuse lpadmin netdev plugdev pulse pulse-access sambashare tape vboxusers video
XsessionErrors: (polkit-gnome-authentication-agent-1:2404): GLib-CRITICAL **: g_once_init_leave: assertion `initialization_value != 0' failed

Revision history for this message
las (bandara-ls) wrote :
Revision history for this message
Apport retracing service (apport) wrote :

StacktraceTop:
 Dasher::CDasherModel::HandleOutput ()
 Dasher::CDasherModel::OneStepTowards ()
 Dasher::CDefaultFilter::Timer ()
 Dasher::CDasherInterfaceBase::NewFrame ()
 CDasherControl::TimerEvent ()

Revision history for this message
Apport retracing service (apport) wrote : Stacktrace.txt
Revision history for this message
Apport retracing service (apport) wrote : ThreadStacktrace.txt
Changed in dasher (Ubuntu):
importance: Undecided → Medium
tags: removed: need-i386-retrace
visibility: private → public
Revision history for this message
Charlie Kravetz (cjkgeek) wrote :

Thanks for reporting this bug and any supporting documentation. Since this bug has enough information provided for a developer to begin work, I'm going to mark it as confirmed and let them handle it from here.

This bug has been reported to the developers of the software. You can track it and make comments at: https://bugzilla.gnome.org/show_bug.cgi?id=640490

Thanks for taking the time to make Ubuntu better!

Changed in dasher (Ubuntu):
status: New → Triaged
Changed in dasher:
importance: Unknown → Medium
status: Unknown → New
Revision history for this message
Charlie Kravetz (cjkgeek) wrote :

From the developers:

Patrick Welche [dasher developer] 2011-01-25 10:29:49 UTC

I was unable to reproduce with the following attempt:

Ubuntu 10.04.2 LTS x86_64
grabbed dasher-4.11.tar.bz2 from download.gnome.org
./configure --prefix /tmp/foo
gedit 2.30.3
cd /tmp/foo; bin/dasher --appstyle=direct

I was able to write into gedit without problems with

System / Preferences / Assistive Technologies / assistive technologies

both enabled and disabled.

I'm not sure what to try next - maybe you could try building from that tarball
and try the same tests I did?

Changed in dasher (Ubuntu):
status: Triaged → Incomplete
Revision history for this message
las (bandara-ls) wrote :

First let me thank you for dasher. :)..I tried to build but i got this error..

 checking for GLIB... configure: error: Package requirements (glib-2.0 >= 2.6) were not met:

No package 'glib-2.0' found

so I tried to install libglib2.0-dev

It gave another error..

libglib2.0-dev:
  Depends: libglib2.0-0 (=2.26.0-0ubuntu1) but 2.26.1-0ubuntu1 is to be installed
 Depends: libglib2.0-bin but it is not going to be installed

What shall I do?

Revision history for this message
Patrick Welche (prlw1) wrote :

Does

aptitude search libglib

give you other -dev options? Hopefully one of the ubuntu specialists can answer - on my old 10.04, I just have
libglib2.0-dev

Revision history for this message
las (bandara-ls) wrote :

This is my o/p10:28:43-las@las-eME640:~$ aptitude search libglib
i libglib-perl - interface to the GLib and GObject librarie
p libglib2-ruby - Glib 2 bindings for the Ruby language
p libglib2-ruby1.8 - Glib 2 bindings for the Ruby language
p libglib2-ruby1.8-dbg - Glib 2 bindings for the Ruby language
i libglib2.0-0 - The GLib library of C routines
p libglib2.0-0-dbg - The GLib libraries and debugging symbols
p libglib2.0-0-refdbg - The GLib library of C routines - refdbg li
p libglib2.0-bin - The programs for the GLib library
i libglib2.0-cil - CLI binding for the GLib utility library 2
p libglib2.0-cil-dev - CLI binding for the GLib utility library 2
i libglib2.0-data - Common files for GLib library
p libglib2.0-dev - Development files for the GLib library
p libglib2.0-doc - Documentation files for the GLib library
i libglibmm-2.4-1c2a - C++ wrapper for the GLib toolkit (shared l
p libglibmm-2.4-dbg - C++ wrapper for the GLib toolkit (debug sy
p libglibmm-2.4-dev - C++ wrapper for the GLib toolkit (developm
p libglibmm-2.4-doc - C++ wrapper for the GLib toolkit (document
p libglibmm-utils-dev - utility functions, classes and widgets wri
p libglibmm-utils2 - utility functions, classes and widgets wri
10:28:54-las@las-eME640:~$

Changed in dasher:
importance: Medium → Critical
status: New → Incomplete
Revision history for this message
Patrick Welche (prlw1) wrote :

I'm afraid this is one for the ubuntu specialists. Rephrased: Why can't I install a -dev version of glib on <insert your version of ubuntu here>?
(I see the same list on my old 10.04.2, and have libglib2.0-dev installed)

Revision history for this message
Charlie Kravetz (cjkgeek) wrote :

Sorry, was hoping a developer would jump in. las is attempting to compile in Natty Narwhal, which is a development version of Ubuntu. The original error using libglib2.0-dev tells us:

libglib2.0-dev:
  Depends: libglib2.0-0 (=2.26.0-0ubuntu1) but 2.26.1-0ubuntu1 is to be installed
 Depends: libglib2.0-bin but it is not going to be installed

Since the package in Natty is newer than the package in the -dev file, it will have to either be forced to install, probably breaking things, or the test has to be done in Ubuntu 10.10.

Revision history for this message
las (bandara-ls) wrote :

looks like my system is gone crazy,Im using 10.10..

 lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 10.10
Release: 10.10
Codename: maverick

i tried to install 11.04 to a diffrent partion but at the end it had changed my 10.10 repos to 11.04.so i manually edited it but looks like damage is done already..What can be done to remove 11.04 packs and install10.10 's?

Revision history for this message
Charlie Kravetz (cjkgeek) wrote :

@las:
I don't know how to repair that. I triage bugs, getting them ready for developers as much as I can. I will attempt to run this on 10.10 today and report back here. Don't try to do anything until we say if it works for me. I will have install Ubuntu 10.10 on a test system and give this a go.

Revision history for this message
Charlie Kravetz (cjkgeek) wrote :

Okay, I installed Ubuntu 10.10, updated it, then installed dasher. I can not reproduce this issue with

System / Preferences / Assistive Technologies / assistive technologies

enabled. With assistive technologies disabled, I can not make dasher write to gedit or any other application.

Revision history for this message
Charlie Kravetz (cjkgeek) wrote :

Errors from terminal when running dasher without Assistive Technologies enabled:

vbadmin@number70:~$ dasher --appstyle=direct
** Message: Opening GUI file: /usr/share/dasher/dasher.direct.ui
** Message: Opening GUI file: /usr/share/dasher/dasher.preferences.ui

(dasher:2928): Gtk-CRITICAL **: IA__gtk_widget_hide: assertion `GTK_IS_WIDGET (widget)' failed

(dasher:2928): Gtk-CRITICAL **: IA__gtk_widget_hide: assertion `GTK_IS_WIDGET (widget)' failed

(dasher:2928): Gtk-CRITICAL **: IA__gtk_widget_hide: assertion `GTK_IS_WIDGET (widget)' failed

(dasher:2928): Gtk-CRITICAL **: IA__gtk_widget_hide: assertion `GTK_IS_WIDGET (widget)' failed

(dasher:2928): Gtk-CRITICAL **: IA__gtk_widget_hide: assertion `GTK_IS_WIDGET (widget)' failed

(dasher:2928): Gtk-CRITICAL **: IA__gtk_widget_hide: assertion `GTK_IS_WIDGET (widget)' failed

(dasher:2928): Gtk-CRITICAL **: IA__gtk_widget_hide: assertion `GTK_IS_WIDGET (widget)' failed

(dasher:2928): Gtk-CRITICAL **: IA__gtk_widget_hide: assertion `GTK_IS_WIDGET (widget)' failed

(dasher:2928): Gtk-CRITICAL **: IA__gtk_widget_hide: assertion `GTK_IS_WIDGET (widget)' failed

(dasher:2928): Gtk-CRITICAL **: IA__gtk_widget_hide: assertion `GTK_IS_WIDGET (widget)' failed

** (dasher:2928): WARNING **: AT_SPI_REGISTRY was not started at session startup.

** (dasher:2928): WARNING **: Could not locate registry
** Message: Could not initialise SPI - accessibility options disabled

(dasher:2928): Gtk-CRITICAL **: IA__gtk_paned_set_position: assertion `GTK_IS_PANED (paned)' failed

(dasher:2928): Gtk-CRITICAL **: IA__gtk_paned_get_position: assertion `GTK_IS_PANED (paned)' failed
Segmentation fault

Revision history for this message
Charlie Kravetz (cjkgeek) wrote :

and, now, as for installing libglib2.0-dev in my clean maverick updated installation. It worked without problems. Unfortunately, the only fix I know of for removing an upgraded ubuntu and trying to downgrade is to reinstall the version you wanted. Sometimes, downgrading all the packages works, and other times, it fails.

Revision history for this message
Charlie Kravetz (cjkgeek) wrote :

and a way to reproduce the issue:

1. start dasher from the terminal, using
    dasher --appstyle=direct
2. Instead of learning to use dasher slowly, move the cursor all the way to the right and allow it type
3. here is the stacktrace from the failure:

#0 0x080973ae in Dasher::CDasherModel::RecursiveOutput(Dasher::CDasherNode*, std::vector<Dasher::SymbolProb, std::allocator<Dasher::SymbolProb> >*) ()
#1 0x08098487 in Dasher::CDasherModel::OneStepTowards(long long, long long, unsigned long, std::vector<Dasher::SymbolProb, std::allocator<Dasher::SymbolProb> >*, int*) ()
#2 0x0809fdc1 in Dasher::CDefaultFilter::Timer(int, Dasher::CDasherView*, Dasher::CDasherModel*, std::vector<Dasher::SymbolProb, std::allocator<Dasher::SymbolProb> >*, int*, Dasher::CExpansionPolicy**) ()
#3 0x08092bd3 in Dasher::CDasherInterfaceBase::NewFrame(unsigned long, bool)
    ()
#4 0x0808c149 in CDasherControl::TimerEvent() ()
#5 0x0809009d in timer_callback(void*) ()
#6 0x009bffcc in ?? () from /lib/libglib-2.0.so.0
#7 0x009bf855 in g_main_context_dispatch () from /lib/libglib-2.0.so.0
#8 0x009c3668 in ?? () from /lib/libglib-2.0.so.0
#9 0x009c3ba7 in g_main_loop_run () from /lib/libglib-2.0.so.0
#10 0x00e611d9 in gtk_main () from /usr/lib/libgtk-x11-2.0.so.0
#11 0x08078f96 in main ()

Revision history for this message
Patrick Welche (prlw1) wrote :

Comment #14 makes sense: the dasher package was compiled to use the accessibility bus. It communicates with other programmes using it, so that is the way it delivers characters to e.g. gedit. If the accessibility bus is switched off, it can't. dasher can be built after configuring with --disable-a11y to not use the accessibility bus, but to use libXtst instead.

Comment #17 is worrying - and of course, I can't reproduce it: I just had dasher running with the cursor all the way to the right and left it running for over 20minutes outputting into gedit. This was on a core duo laptop - just mentioning that given "TimerEvent" in your backtrace. In the source head we have since made simplifications in that area, so RecursiveOutput has actually already been removed, but AFAIK those changes weren't actually bug fixes, just refactorings. What hardware have you got? (Still assuming there is some sort of event race condition...)

Revision history for this message
las (bandara-ls) wrote :

I got the sane results as #17 rain iy with valgrind.im attaching log.and my h/w info

Revision history for this message
las (bandara-ls) wrote :
Revision history for this message
Patrick Welche (prlw1) wrote :

Something happened to your attachement :-(

(I still can't reproduce on 10.04... will try other computers running 10.04 in the lab next week)

Revision history for this message
las (bandara-ls) wrote :

attachments again

Revision history for this message
las (bandara-ls) wrote :
tags: added: a11y
Changed in dasher:
status: Incomplete → Expired
Changed in dasher:
status: Expired → New
Changed in dasher:
status: New → Incomplete
Changed in dasher:
status: Incomplete → Expired
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.