Bug 523700 - iccexamin fails to run with "Segmentation fault" message
Summary: iccexamin fails to run with "Segmentation fault" message
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: Fedora
Classification: Fedora
Component: icc_examin
Version: 12
Hardware: x86_64
OS: Linux
low
high
Target Milestone: ---
Assignee: Nicolas Chauvet (kwizart)
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2009-09-16 13:45 UTC by Alex
Modified: 2010-09-20 20:33 UTC (History)
3 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2010-08-28 14:24:08 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)
thread apply all bt full (2.96 KB, text/plain)
2009-11-18 04:55 UTC, Bill McGonigle
no flags Details

Description Alex 2009-09-16 13:45:19 UTC
Description of problem:

I have Acer laptop, Intel Core Duo, Intel965GM chipset, 4GB, 320GB hard, etc
I use Fedora 11 x86_64 fully updated up to date (Kernel 2.6.30.5).
I have installed via add/remove programs tool Cinepaint (has iccexamin as plugin) and iccexamin standalone packages. Neither the plugin nor the stanalone program works.
Precisely, iccesamin-0.46.rpm
Once started the frame of the graphic window is drawn (only frame, no any contents or even background) and then the program exits writing "Segmentation fault".
I tried to download the source and compile it myself. It does compile without problems but has the same net result ("Segmentation fault").

How reproducible:
Every time when the program is started

Actual results:
"Segmentation fault" output to the terminal

Expected results:
Graphical interface, etc

Additional info:
I do not remember have I tried it or not on the previous (2.6.29) Kernel but almost sure it is the graphics (i.e. Gnome) related problem.
I do not know about non-64 version.
I know that when I had openSuSE (about 2 weeks ago) I had downloaded the program, compiled it and  ran normally.

Comment 1 Alex 2009-09-16 14:03:44 UTC
Two changes:
I mistakenly put 'rawhide'. It is Fedora 11.
I assume the severity might be 'Medium' since workaround is not that easy.

Comment 2 Nicolas Chauvet (kwizart) 2009-09-17 22:04:15 UTC
I cannot reproduce this problem with nvidia driver on x86_64
So this indeed seems a driver problem.

I will try to reproduce it using radeon with x600se i686, but I may rely this problem with some weird other blender still experiences with F-11 (wasn't exerienced with early F-10).

Comment 3 Alex 2009-09-18 08:53:14 UTC
Hope, the following helps in the understanding.
Both, the standalone program and the cinepaint plugin exit with the same error but I have more info with the plugin.
1. I start cinepaint from the terminal
2. I crate a new file and go to menu "Image->Icc profile information->Image profile..."
3. At this moment iccexamin plugin starts with the following output to the terminal:
------------------------------------------------------------------
icc_cinepaint.cpp:121
/usr/lib64/cinepaint/0.22-1/plug-ins/icc_examin_cp
options are:
 -bg2 color
 -bg color
 -di[splay] host:n.n
 -dn[d]
 -fg color
 -g[eometry] WxH+X+Y
 -i[conic]
 -k[bd]
 -na[me] classname
 -nod[nd]
 -nok[bd]
 -not[ooltips]
 -s[cheme] scheme
 -ti[tle] windowtitle
 -to[oltips]

/usr/lib64/cinepaint/0.22-1/plug-ins/icc_examin_cp: Segmentation fault caught
/usr/lib64/cinepaint/0.22-1/plug-ins/icc_examin_cp (pid:2831): [E]xit, [H]alt, show [S]tack trace or [P]roceed:
------------------------------------------------------------------
4. Simultaneously ONLY THE FRAME of the ICC examin window is drawn but there is no context in it (just everything what was below). The title is written at the top of the window.
5. I type 's' and press enter in the terminal. This produces the following typical output:
------------------------------------------------------------------
#0  0x00000037496a3b9d in waitpid () from /lib64/libc.so.6
#1  0x000000374b2150c2 in g_on_error_stack_trace ()
#2  0x000000374b215293 in g_on_error_query () from /lib64/libglib-2.0.so.0
#3  0x00007f29d0d3afc7 in ?? () from /usr/lib64/libcinepaint.so.0
#4  <signal handler called>
#5  0x0000000000000021 in ?? ()
#6  0x000000000044caf8 in ?? ()
#7  0x0000000000430d44 in ?? ()
#8  0x00000000004111e9 in gimp_image_get_icc_profile_by_mem ()
#9  0x0000000000412a8a in gimp_image_get_icc_profile_by_mem ()
#10 0x0000000000413f71 in gimp_image_get_icc_profile_by_mem ()
#11 0x00007f29d0d3ae44 in gimp_main () from /usr/lib64/libcinepaint.so.0
#12 0x000000374961ea2d in __libc_start_main () from /lib64/libc.so.6
#13 0x000000000040f0b9 in gimp_image_get_icc_profile_by_mem ()
#14 0x00007fffe1016558 in ?? ()
#15 0x000000000000001c in ?? ()
#16 0x0000000000000006 in ?? ()
#17 0x00007fffe1018510 in ?? ()
#18 0x00007fffe1018543 in ?? ()
#19 0x00007fffe1018549 in ?? ()
#20 0x00007fffe101854c in ?? ()
#21 0x00007fffe101854f in ?? ()
#22 0x00007fffe1018552 in ?? ()
#23 0x0000000000000000 in ?? ()
/usr/lib64/cinepaint/0.22-1/plug-ins/icc_examin_cp (pid:2831): [E]xit, [H]alt, show [S]tack trace or [P]roceed: 
------------------------------------------------------------------
6. Typing 'p'+Enter just closes the ICC examin window (incomplete window, frame only)

Yes, when I wrote graphics I meant the window management related functions, not the hardware. Sorry for the confusion, I'm not so programmer, excuse me, please ;)

Comment 4 Alex 2009-11-14 14:24:55 UTC
 I'm still fighting from time to time with this bug. What I could find is that I cannot always compile from the source the icc_examin package.

 Namely, if I just do ./configure, make, it does compile but if I want to do
./configure --enable-debug in order to build the debug version the 'make' command stops at some moment with the following statement:

-------------------------------------
Compiling threads.cpp ...
c++ -I/usr/local/include -I/usr/X11R6/include -I. -I/usr/include/g++ -I/usr/include  -Wall -g -DDEBUG --pedantic -Wall -g -DDEBUG --pedantic -I/usr/local/include -I/usr/X11R6/include -I. -I/usr/include/g++ -I/usr/include  -I/usr/include/freetype2 -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_THREAD_SAFE -D_REENTRANT -I/usr/X11R6/include -I/usr/include      -I/usr/include/oyranos  -I/usr/include/freetype2   -I./Flmm -c threads.cpp
threads.cpp: In function ‘int fl_create_thread(Fl_Thread&, void* (*)(void*), void*)’:
threads.cpp:62: error: ‘PTHREAD_STACK_MIN’ was not declared in this scope
threads.cpp:64: error: ‘PTHREAD_STACK_MIN’ was not declared in this scope
make: *** [threads.o] Error 1
-------------------------------------

I presume that this is the origin of my problem.
Hope this helps.

Comment 5 Bug Zapper 2009-11-16 12:29:08 UTC
This bug appears to have been reported against 'rawhide' during the Fedora 12 development cycle.
Changing version to '12'.

More information and reason for this action is here:
http://fedoraproject.org/wiki/BugZappers/HouseKeeping

Comment 6 Bill McGonigle 2009-11-18 04:55:10 UTC
Created attachment 370008 [details]
thread apply all bt full

looks like perhaps a GL issue?

2.6.30.8-64.fc11.x86_64

Comment 7 Kai-Uwe Behrmann 2009-12-01 05:50:43 UTC
The compile error is fixed in ICC Examin git #5447c0b9.

Would a new release help for Fedora to bring in the current state of
ICC Examin. (It would require as well a new Oyranos.) 
I think to have fixed some of the x86_64 issues. 

Alex, thanks for reporting to SF.

Comment 8 Alex 2009-12-02 22:44:10 UTC
Now I have Fedora 12 updated fully just few minutes ago. No luck with ICCexamin yet. Still 'segmentation fault' I'll try using git to compile with debug enabled in few days. Will report the result.

Comment 9 Nicolas Chauvet (kwizart) 2009-12-20 21:48:51 UTC
(In reply to comment #7)
> The compile error is fixed in ICC Examin git #5447c0b9.
> 
> Would a new release help for Fedora to bring in the current state of
> ICC Examin. (It would require as well a new Oyranos.) 
> I think to have fixed some of the x86_64 issues. 
If new release can be made, I guess it will be nice to have, last release was one year agao.
But the more difficult patch I need to rebase is oyranos-0.1.9-shared_NVCtrl.patch
We cannot bundle such library:
https://fedoraproject.org/wiki/Packaging:No_Bundled_Libraries

Is there a way I can build oyranos using the system version without permanently have a patch on the buildsys ?

Comment 10 Kai-Uwe Behrmann 2009-12-27 20:19:18 UTC
ICC Examin git commit #d664369e applies and corrects
icc_examin-0.46-default.patch and
icc_examin-0.46-fixes.patch

Thanks.

Comment 11 Kai-Uwe Behrmann 2009-12-27 21:55:02 UTC
Oyranos git commit #9bdbcc44 detects and uses a
external libXNVCtrl. Hope it works on Fedora.

Comment 12 Nicolas Chauvet (kwizart) 2010-08-28 14:24:08 UTC
This bug is fixed in icc-examin-0.47-2 that will hit updates-testing on next push.

Comment 13 Alex 2010-09-18 10:47:49 UTC
Today icc_examin 0.47 appeared in updates. It now starts, but it crashes when I try to open the 'windows->gamut' if some profile is already loaded.
Or, I can open this window empty when NO profile is loaded but then program crashes if I load the profile.
The issue is systematic.

I use Fedora 13 fully updated up to now, Sep 18 2010.

Also one aesthetic issue: all windows are opened with their zero edge at the top of the desktop, i.e. UNDER the top gnome panel if such panel presents. You then have to use key bindings to move the window but not mouse. This is just inconvenient.

Any ideas how to fix at least first?

Comment 14 Kai-Uwe Behrmann 2010-09-20 20:33:31 UTC
I can repeat when no XYZ.icc profile is installed.
The same will be true for no Lab.icc profile. They are
used as system default profiles.
Both come within the original Oyranos package.

After installing these profiles the crashes should go 
away.


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