Bug 1074 – Segmentation fault at start-up

PySide Bugzilla Closed for New Bugs

PySide is now a Qt Add-on and uses the Qt Project's JIRA Bug Tracker instead of this Bugzilla instance. This Bugzilla is left for reference purposes.

Bug 1074 - Segmentation fault at start-up

Summary:

Segmentation fault at start-up

Status:	REOPENED

Product:	PySide
Classification:	Unclassified
Component:	PySide
Version:	1.0.8
Platform:	Macintosh Mac OS

Importance:	P5 normal
Assigned To:	Hugo Parente Lima

URL:

Depends on:
Blocks:
	Show dependency tree / graph

Reported:	2011-11-30 06:55 EET by JC Briar
Modified:	2012-03-08 16:57 EET (History)
CC List:	9 users (show)

See Also:

Attachments
sample program that seg faults on startup (1.04 KB, text/x-python-script) 2011-11-30 06:55 EET, JC Briar	Details
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.

Description JC Briar 2011-11-30 06:55:32 EET

Created attachment 459 [details]
sample program that seg faults on startup

After upgrading (via MacPorts) from PySide 1.0.2 to 1.0.8, my app started
crashing at start-up: the Python interpreter fails with a segmentation fault.
The top of the stack is always as follows:

0   libSystem.B.dylib                 0x93763e10 strlen + 16
1   org.python.python                 0x0022b459 PyDict_GetItemString + 25
2   org.python.python                 0x0024d91f PyType_Ready + 927
3   org.python.python                 0x00251847 type_new + 2551
4   ...hiboken-python2.6.1.0.dylib    0x00104033 SbkObjectTypeTpNew + 243
5   org.python.python                 0x0024aafc type_call + 60
6   org.python.python                 0x001e4695 PyObject_Call + 85
7   org.python.python                 0x001e47db PyObject_CallFunctionObjArgs +
187
8   org.python.python                 0x002935b2 PyEval_EvalFrameEx + 4162
9   org.python.python                 0x002992cd PyEval_EvalCodeEx + 2109
10  org.python.python                 0x00299457 PyEval_EvalCode + 87

Note that I'm not sure if the problem is within PySide itself, or Shiboken, or
some other component.

I've distilled my app down to a small sample that exhibits the same problem;
see attachment.

Other vital stats:
Mac OS X 10.5.8
Python 2.6
Shiboken 1.0.9
Qt 4.7.4

Comment 1 JC Briar 2011-12-01 21:38:26 EET

If I use the MacPorts 'port activate' command to back out of the latest
revisions of PySide and all related software, then the segmentation faults go
away. Specifically, the provided sample program seg faults when the following
are installed:

apiextractor@0.10.8_0
apr@1.4.5_3
apr-util@1.3.12_0
bison@2.5_0
cairo@1.10.2_4+x11
cmake@2.8.6_0
dbus@1.4.16_0
freetype@2.4.8_0
gdbm@1.10_0
generatorrunner@0.6.14_0
ghostscript@9.04_1
glib2@2.30.2_1
gperf@3.0.4_2
graphviz@2.28.0_2+no_pangocairo+no_x11+python26
gsed@4.2.1_2
help2man@1.40.4_1
jasper@1.900.1_8
libiconv@1.14_0
libpixman@0.24.0_0
libpng@1.4.8_0
libtool@2.4.2_0
ncurses@5.9_1
neon@0.29.6_0
netpbm@10.47.31_0
openssl@1.0.0e_1
pango@1.28.4_2+x11
pcre@8.12_1
perl5.12@5.12.3_2
pkgconfig@0.26_0
py26-pyside@1.0.8_0
py26-shiboken@1.0.9_0
py26-tkinter@2.6.7_0
python26@2.6.7_3
python27@2.7.2_3
qt4-mac@4.7.4_1+quartz
sqlite3@3.7.9_0
subversion@1.7.1_0
swig@2.0.4_0
swig-python@2.0.4_0
tcl@8.5.11_0
tiff@3.9.5_0
tk@8.5.11_0
xorg-inputproto@2.0.99.1_0
xorg-libX11@1.4.4_0
xorg-libXext@1.3.0_0
xorg-util-macros@1.15.0_0
xorg-xproto@7.0.22_0
xrender@0.9.6_1

And it doesn't seg fault when the following are installed:

apiextractor@0.10.2_0
apr@1.4.2_1
apr-util@1.3.10_0
bison@2.4.3_0
cairo@1.10.2_3
cmake@2.8.4_0
dbus@1.2.24_1
freetype@2.4.4_0
gdbm@1.8.3_3
generatorrunner@0.6.9_0
ghostscript@9.02_0
glib2@2.28.6_0
gperf@3.0.4_0
graphviz@2.28.0_0+no_pangocairo+no_x11+python26
gsed@4.2.1_1
help2man@1.39.3_0
jasper@1.900.1_7
libiconv@1.13.1_0
libpixman@0.22.0_0
libpng@1.4.7_0
libtool@2.4_0
ncurses@5.9_0
neon@0.29.5_0
netpbm@10.47.26_0
openssl@1.0.0d_0
pango@1.28.4_0
pcre@8.12_0
perl5.12@5.12.3_1
pkgconfig@0.25_2
py26-pyside@1.0.2_0
py26-shiboken@1.0.2_0
py26-tkinter@2.6.6_0
python26@2.6.6_3
python27@2.7.1_3
qt4-mac@4.7.3_0+quartz
sqlite3@3.7.6_0
subversion@1.6.16_0
swig@2.0.3_0
swig-python@2.0.3_0
tcl@8.5.9_0
tiff@3.9.4_0
tk@8.5.9_0
xorg-inputproto@2.0.1_0
xorg-libX11@1.4.3_0
xorg-libXext@1.2.0_0
xorg-util-macros@1.13.0_0
xorg-xproto@7.0.21_0
xrender@0.9.6_0

While this serves as a workaround for now, it clearly isn't a long-term
solution to the problem.

Comment 2 Hugo Parente Lima 2011-12-02 14:34:01 EET

It works for me, but I'm on Linux.

Comment 3 JC Briar 2011-12-02 17:19:40 EET

(In reply to comment #2)
> It works for me, but I'm on Linux.

I'm not surprised it works on Linux. It's a rather subtle sort of bug: remove a
few innocuous-seeming lines of code, and the distilled sample program will run;
remove the identical lines from the real app, and it continues to fail. 

If you could find someone capable of confirming the bug on another Mac, I would
be most grateful.

Comment 4 Hugo Parente Lima 2012-01-09 19:36:46 EET

I tested the code on our mac build bot and it works, I used the latest revision
from git.

Comment 5 JC Briar 2012-01-09 20:48:48 EET

Which version of Python did you use? Due to dependencies on other packages, I'm
using Python 2.6.

Comment 6 Matti Airas 2012-03-08 16:57:53 EET

PySide is now a Qt-addon and uses Qt Project's JIRA tool for tracking bugs.
Please verify that the bug is still valid and re-submit it in the address
below:

https://bugreports.qt-project.org/

Sorry for the inconvenience!