skencil crashes on startup with a SIGSEGV in free()

Bug #81567 reported by Niko
166
This bug affects 1 person
Affects Status Importance Assigned to Milestone
skencil (Ubuntu)
Invalid
Medium
Unassigned

Bug Description

Binary package hint: skencil

skencil is simply crashing ... clicking from the application menu.

ProblemType: Crash
Date: Fri Jan 26 08:44:44 2007
DistroRelease: Ubuntu 7.04
ExecutablePath: /usr/lib/skencil/skencil.py
InterpreterPath: /usr/bin/python2.5
Package: skencil 0.6.17-7build1
ProcCmdline: /usr/bin/python /usr/bin/skencil
ProcCwd: /home/niko
ProcEnviron:
 PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/bin/X11:/usr/games
 LANG=en_US.UTF-8
 SHELL=/bin/zsh
Signal: 11
SourcePackage: skencil
StacktraceTop:
 free () from /lib/tls/i686/cmov/libc.so.6
 ?? () from /usr/lib/skencil/Sketch/../Lib/paxmodule.so
 ?? ()
 ?? ()
Uname: Linux sunova 2.6.20-5-generic #2 SMP Sat Jan 6 14:50:47 UTC 2007 i686 GNU/Linux
UserGroups: adm admin audio cdrom dialout dip floppy lpadmin plugdev scanner video

Revision history for this message
Niko (niko-macnica) wrote : Dependencies.txt
Revision history for this message
Niko (niko-macnica) wrote : Disassembly.txt
Revision history for this message
Niko (niko-macnica) wrote : ProcMaps.txt
Revision history for this message
Niko (niko-macnica) wrote : ProcStatus.txt
Revision history for this message
Niko (niko-macnica) wrote : Registers.txt
Revision history for this message
Niko (niko-macnica) wrote : Stacktrace.txt
Revision history for this message
Niko (niko-macnica) wrote : ThreadStacktrace.txt
Revision history for this message
Stefano Costa (steko) wrote : Re: [apport] skencil.py crashed with SIGSEGV in free()

I think I have the same issue here.

Revision history for this message
Dave Gilbert (ubuntu-treblig) wrote :

Yep same here; Feisty skencil package 0.6.17-7build1

Revision history for this message
Daniel Holbach (dholbach) wrote :

Thanks for your bug report.

Changed in skencil:
assignee: nobody → motu
importance: Undecided → Medium
Revision history for this message
Apport retracing service (apport) wrote : Symbolic stack trace

StacktraceTop:_start () from /lib/ld-linux.so.2
?? ()
?? ()
?? ()
?? ()

Revision history for this message
Apport retracing service (apport) wrote : Symbolic threaded stack trace
Revision history for this message
Sebastien Bacher (seb128) wrote : Re: [apport] skencil.py crashed with SIGSEGV in free()

Thanks for your bug report. The crash looks like a memory corruption. Could you try to get a valgrind log for it (you can follow the instructions from https://wiki.ubuntu.com/Valgrind for that)?

Changed in skencil:
status: Unconfirmed → Needs Info
Revision history for this message
Christopher Barrington-Leigh (cpbl) wrote :

I've the same problem on Feisty.
But when I run it with the Valgrind instructions, it doesn't crash!!
c

Revision history for this message
Pauli Virtanen (pauli-virtanen) wrote :

Note that the origin of this bug may be the PyMem/PyObject changes in python2.5. [2]
Skencil built from Ubuntu's source package works fine under python2.4 but crashes under python2.5.
And indeed, lines with PyMem_DEL appear in the source.

Googling, I heard[2] that skencil SVN should already ported to python2.5, so patches could maybe be extracted from there.

[1] http://docs.python.org/whatsnew/ports.html
[2] http://lists.wald.intevation.org/pipermail/skencil-devel/2007-April/000018.html

Changed in skencil:
status: Incomplete → New
Revision history for this message
Serge Matveenko (lig) wrote :

python2.4 /usr/lib/skencil/skencil.py

Revision history for this message
thsths (thomas-steffen+ubuntu) wrote :

Hi

I have been rather annoyed with this bug for a while, because sketch/skencil is my favorite drawing program for more complex things. And it seems that I have found a work around short of recompiling, although it is certainly not for anyone faint of heart.

1. Remove skencil.
2. Install python2.4.
3. Download skencil_0.6.17-7_i386.deb from Ubuntu 6.10 and install it with dpkg. This shows a conflict in the python version.
4. Edit /var/lib/dpkg/status to delete the conflict: just remove the entry "python (<< 2.5), " in skencil.
5. Edit /usr/bin/skencil to start /usr/bin/python2.4 instead of /usr/bin/python. (skencil should now work!)
6. Use dselect to put a hold on skencil. (is there an easier way to do this?)

Note that this does make changes to your packaging system. If you make a mistake, you may have to reinstall your system. When the workaround is no longer needed, the changes should go away once you remove the hold flag and upgrade.

Revision history for this message
Emmet Hikory (persia) wrote :

Confirming that skencil just doesn't work as installed. There are some apport stack traces in some of the duplicates.

Changed in skencil:
assignee: motu → nobody
status: New → Confirmed
Revision history for this message
Guy K. Kloss (guy.kloss) wrote :

One of the (former) coders of Skencil (Tobias Hilbricht) told me that Skencil depends on execution with Python 2.4, however Python 2.5 is the default for quite a few versions of Ubuntu already. But even executing it with python2.4 does lead to the crash. Apparently Skencil's development is somewhat abandoned as well.

Apparently there is a successor (fork of Skencil) called sK1, with a few enhancements. It can be found here: http://www.sk1project.org

Maybe that should replace Skencil?

Revision history for this message
January (january-weiner) wrote :

I looked into the sk1project, but I am afraid that it will require a lot of work to make it run.

First of all, trying to install from their svn fails -- the mentioned "sk1_re" repository cannot be found on the sourceforge. I did manage to compile the sk1 and sk1_re packages provided as tar.gz, though. They seem to compile everything short of the linux kernel itself :-) (own python, own tcl/tk libs, ...). And yet they fail.

I got something which looked like a nice, spiced up skencil. After installing kdebase-bin I even got the open file dialog :-) (they are using kdialog heavily). Some simple operations that I tried seem to work as well, like creating a rectangle (duh). What did not work, however, was for example:

* adding any sort of text (no fonts available at all. A hilarious message: "font Slim not found. Using Slim instead")
* importing graphics (I tried EPS and jpeg; various error messages on the command line, some cryptic popup windows in the case of JPEG)

The good side is that they seem to be alive (as opposed to the sketch/skencil development team).

All in all, I don't like that route so much. I liked sketch because it was really lean and fast. If I wanted to go for a heavy graphics package (sK1 installs around 100MB in the "sk1_Apps" folder) that needs a huge GUI library (like Cairo), I'd choose inkscape. At least I have a popular, relatively stable program; where would be the advantage of sk1, I ask?

The only valuable bit for me is the uniconverter, which, if it fulfils its promises, will save my sk files from oblivion or re-drawing using inkscape.

j.

Revision history for this message
Peter Clifton (pcjc2) wrote :

Try the following.. appologies this has an -ubuntu2 name on the version. BEWARE.. this is NOT an official Ubuntu version, just a proposal for one.

Revision history for this message
Peter Clifton (pcjc2) wrote :

The .dsc file

Revision history for this message
Peter Clifton (pcjc2) wrote :

Here is the debdiff.

Please be aware, that the patch I wrote superficially works for me (no more crash, and I can use skconvert with inkscape again), but I have not tested it that thoroughly.

Revision history for this message
Dave Gilbert (ubuntu-treblig) wrote :

The error seems to be different in interpid's 0.6.17-16:

Traceback (most recent call last):
  File "/usr/bin/skencil", line 34, in <module>
    Sketch.main.main()
  File "/usr/lib/skencil/Sketch/Base/main.py", line 148, in main
    run_script = options.run_script)
  File "/usr/lib/skencil/Sketch/UI/skapp.py", line 183, in __init__
    self.build_window()
  File "/usr/lib/skencil/Sketch/UI/skapp.py", line 223, in build_window
    self.run_script)
  File "/usr/lib/skencil/Sketch/UI/mainwindow.py", line 98, in __init__
    self.build_menu()
  File "/usr/lib/skencil/Sketch/UI/mainwindow.py", line 901, in build_menu
    self.update_mru_files()
  File "/usr/lib/skencil/Sketch/UI/mainwindow.py", line 380, in update_mru_files
    self.file_menu.RebuildMenu()
  File "/usr/lib/skencil/Sketch/UI/tkext.py", line 367, in RebuildMenu
    self.menu.delete(0, END)
  File "/usr/lib/python2.5/lib-tk/Tkinter.py", line 2683, in delete
    if c in self._tclCommands:
TypeError: argument of type 'NoneType' is not iterable

Dave

Revision history for this message
Pietro Battiston (toobaz) wrote :

Dave, thank you for your report, however I think it has nothing to do with this bug, but instead with

https://bugs.launchpad.net/ubuntu/+bug/285955

Revision history for this message
Daniel T Chen (crimsun) wrote :

I've attempted to reproduce this symptom in an i386 maverick vm using virtualbox-ose (running on an amd64 host) and cannot do so. skencil appears to start up successfully and is "quite usable." Can you attempt to reproduce this symptom using a current daily live cd of maverick?

Revision history for this message
Jason McMullan (jason-mcmullan) wrote :

No longer reproducible with the default AMD64 Lucid installation (using python 2.6)

I think this can be closed for Lucid and higher.

Revision history for this message
xteejx (xteejx-deactivatedaccount) wrote :

Thank you for the update. Closing bug report.

Changed in skencil (Ubuntu):
status: Confirmed → Invalid
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.