zooming cause crash of Xorg

[Fabio Marconi]

Hello
I'm trying a4-0.03, md5sum verified
Step to reproduce:
open the image A4_nested_transform_svg, maximize, click forward till three appear, then continuosly click on zoom+ till Xorg crash rebooting the session.
Is reproducible only with this image, with roi.svg it works as expected.

Backtrace:
[ 4982.804] 0: /usr/bin/X (xorg_backtrace+0x28) [0x45c448]
[ 4982.804] 1: /usr/bin/X (0x400000+0x5a71d) [0x45a71d]
[ 4982.804] 2: /lib/libpthread.so.0 (0x7f918c149000+0xfb40) [0x7f918c158b40]
[ 4982.804] 3: /usr/lib/xorg/modules/drivers/intel_drv.so (0x7f91892cc000+0x24fec) [0x7f91892f0fec]
[ 4982.804] 4: /usr/lib/xorg/modules/drivers/intel_drv.so (0x7f91892cc000+0x25382) [0x7f91892f1382]
[ 4982.804] 5: /usr/lib/xorg/modules/drivers/intel_drv.so (0x7f91892cc000+0x34759) [0x7f9189300759]
[ 4982.804] 6: /usr/bin/X (0x400000+0xdc560) [0x4dc560]
[ 4982.804] 7: /usr/bin/X (0x400000+0xd19c8) [0x4d19c8]
[ 4982.805] 8: /usr/bin/X (0x400000+0x3f949) [0x43f949]
[ 4982.805] 9: /usr/bin/X (0x400000+0x2184b) [0x42184b]
[ 4982.805] 10: /lib/libc.so.6 (__libc_start_main+0xfe) [0x7f918b0b4d8e]
[ 4982.805] 11: /usr/bin/X (0x400000+0x213d9) [0x4213d9]
[ 4982.805] Segmentation fault at address (nil)
[ 4982.805]
Caught signal 11 (Segmentation fault). Server aborting
[ 4982.805]
Please consult the The X.Org Foundation support

[Fabio Marconi (fabiomarconi)]

Reproduceable in Virtualbox in a fresh updated Maverick

[Andrea Corbellini (andrea.corbellini)]

I suspect this is a bug in PyGTK, GTK or X.org itself, because A4 does not have control on the segments that are executed. I'll try to reproduce the bug and do some debugging with the -dbg packages in a virtual machine and report back the results here.

[Fabio Marconi (fabiomarconi)]

video of the issue

[Andrea Corbellini (andrea.corbellini)]

Hi. Unfortunately, I'm unable to reproduce the bug. I've tried both with metacity and compiz. Could you please try to reproduce the bug with Xephyr? Here's what you have to do (in three different terminal windows):

$ Xephyr :1
$ DISPLAY=:1 compiz
$ DISPLAY=:1 ./a4

Then, do the steps to reproduce the bug. This way, we'll know what exactly is failing and where to look for more detailed information.

[Fabio Marconi (fabiomarconi)]

njin@Lucid:~/a4$ Xephyr :1
[dix] Could not init font path element /usr/share/fonts/X11/cyrillic, removing from list!
XIO: fatal IO error 11 (Resource temporarily unavailable) on X server ":0.0"
      after 102 requests (102 known processed) with 0 events remaining.
njin@Lucid:~/a4$

Can you plase suggest what to do?
When I launch Xephyr only black screen appear and nothing more grabbing or not the mouse/keyboard

[Andrea Corbellini (andrea.corbellini)]

It's OK if you see a black window. After running Xephyr, open an another terminal and run `DISPLAY=:1 compiz`, then in a new terminal run `DISPLAY=:1 ./a4`.

After you have run this three commands, you should see the window of A4 inside the window of Xephyr. You have to use it to try to reproduce the problem. If the Xephyr window disappears, than it means that you have successfully reproduced the bug. The output on the three terminals will help us detect the cause of the problem.

[Fabio Marconi (fabiomarconi)]

Hello
Under Xephyr A4 don't crash nor in compiz or metacity.
I've installed on another machine with or without video drivers and it don't crash too, it is a bit slower (amd64 2800GHz singlecore/1.5Gbram), but consumption never goes to hight levels.
In the new machine it is fast but with a large use of resources, the Virtualbox run in the same machine that crash.
So seems depending on this machine( i3 2.93/2Gram), i will investigate.
Fabio

[Fabio Marconi (fabiomarconi)]

Jan 1 23:07:02 Lucid kernel: [44768.669142] [drm:i915_gem_mmap_gtt_ioctl] *ERROR* Attempting to mmap a purgeable buffer