0.6.2 core dumps under jaunty x64

[danboid]

I've just downloaded the source tarball for 0.6.2, I followed the build instructions and it configured and compiled without any error, but when I try running pianobooster it just dumps core - it never gets as far as the main window unfortunately :(

Here's what pianobooster says when I run it:

dan@iatn-ati:~/src/pianobooster-src-0.6.2/build$ ./pianobooster
/home/dan/.themes/Dust/gtk-2.0/gtkrc:708: Murrine configuration option "highlight_ratio" will be deprecated in future releases. Please use "highlight_shade" instead.
/home/dan/.themes/Dust/gtk-2.0/gtkrc:709: Murrine configuration option "lightborder_ratio" will be deprecated in future releases. Please use "lightborder_shade" instead.
unknown chip id 0x94c9, can't guess.
Error: Zero tracks in SMF file
Opening song /home/dan/autosave/
Error: Zero tracks in SMF file
Using Pianist Channels 1 + 2
Segmentation fault (core dumped)

I updated Jaunty just before compiling pb and I've compiled both LMMS and Ardour on this laptop so its got more multimedia dev deps installed than you can shake a very geeky stick at :)

Any ideas?

Thanks

Dan

[danboid]

I have tried running the pb Linux 0.6.2 binary now too but all that says when I try to run that is:

./pianobooster: error while loading shared libraries: libQtOpenGL.so.4: cannot open shared object file: No such file or directory

Am I correct in thinking that the package I would require to get this working properly under Ubuntu would be libqt4-opengl-dev (and its non-dev counterpart of course) - both of which I have installed, to no avail.

I thought it might for you to see a dump of the build process as there were a few errors, come to think of it but nothing that interrupted compilation:

dan@iatn-ati:~/src/pianobooster-src-0.6.2/build$ cmake ../src/
-- Configuring done
-- Generating done
-- Build files have been written to: /home/dan/src/pianobooster-src-0.6.2/build
dan@iatn-ati:~/src/pianobooster-src-0.6.2/build$ make
-- Configuring done
-- Generating done
-- Build files have been written to: /home/dan/src/pianobooster-src-0.6.2/build
[  1%] Generating ui_GuiLoopingPopup.h
[  3%] Generating moc_QtWindow.cxx
[  5%] Generating moc_GlView.cxx
[  7%] Generating moc_GuiTopBar.cxx
[  9%] Generating moc_GuiSidePanel.cxx
[ 11%] Generating moc_GuiMidiSetupDialog.cxx
[ 13%] Generating moc_GuiKeyboardSetupDialog.cxx
[ 15%] Generating moc_GuiPreferencesDialog.cxx
[ 17%] Generating moc_GuiLoopingPopup.cxx
[ 19%] Generating qrc_application.cxx
[ 21%] Generating ui_GuiTopBar.h
[ 23%] Generating ui_GuiSidePanel.h
[ 25%] Generating ui_GuiMidiSetupDialog.h
[ 26%] Generating ui_GuiKeyboardSetupDialog.h
[ 28%] Generating ui_GuiPreferencesDialog.h
[ 30%] Building CXX object CMakeFiles/pianobooster.dir/QtMain.o
[ 32%] Building CXX object CMakeFiles/pianobooster.dir/QtWindow.o
[ 34%] Building CXX object CMakeFiles/pianobooster.dir/GuiTopBar.o
[ 36%] Building CXX object CMakeFiles/pianobooster.dir/GuiSidePanel.o
[ 38%] Building CXX object CMakeFiles/pianobooster.dir/GuiMidiSetupDialog.o
[ 40%] Building CXX object CMakeFiles/pianobooster.dir/GuiKeyboardSetupDialog.o
[ 42%] Building CXX object CMakeFiles/pianobooster.dir/GuiPreferencesDialog.o
[ 44%] Building CXX object CMakeFiles/pianobooster.dir/GuiLoopingPopup.o
[ 46%] Building CXX object CMakeFiles/pianobooster.dir/GlView.o
/home/dan/src/pianobooster-src-0.6.2/src/GlView.cpp: In member function ‘virtual void CGLView::resizeGL(int, int)’:
/home/dan/src/pianobooster-src-0.6.2/src/GlView.cpp:242: warning: unused variable ‘space’
[ 48%] Building CXX object CMakeFiles/pianobooster.dir/MidiFile.o
[ 50%] Building CXX object CMakeFiles/pianobooster.dir/MidiDevice.o
[ 51%] Building CXX object CMakeFiles/pianobooster.dir/MidiTrack.o
[ 53%] Building CXX object CMakeFiles/pianobooster.dir/Song.o
[ 55%] Building CXX object CMakeFiles/pianobooster.dir/Conductor.o
[ 57%] Building CXX object CMakeFiles/pianobooster.dir/Util.o
[ 59%] Building CXX object CMakeFiles/pianobooster.dir/Chord.o
[ 61%] Building CXX object CMakeFiles/pianobooster.dir/Tempo.o
[ 63%] Building CXX object CMakeFiles/pianobooster.dir/rtmidi/RtMidi.o
[ 65%] Building CXX object CMakeFiles/pianobooster.dir/StavePosition.o
[ 67%] Building CXX object CMakeFiles/pianobooster.dir/Score.o
[ 69%] Building CXX object CMakeFiles/pianobooster.dir/Cfg.o
[ 71%] Building CXX object CMakeFiles/pianobooster.dir/Piano.o
[ 73%] Building CXX object CMakeFiles/pianobooster.dir/Draw.o
[ 75%] Building CXX object CMakeFiles/pianobooster.dir/Scroll.o
/home/dan/src/pianobooster-src-0.6.2/src/Scroll.cpp: In member function ‘void CScroll::compileSlot(CScroll::CSlotDisplayList)’:
/home/dan/src/pianobooster-src-0.6.2/src/Scroll.cpp:50: warning: unused variable ‘av8Left’
[ 76%] Building CXX object CMakeFiles/pianobooster.dir/Notation.o
[ 78%] Building CXX object CMakeFiles/pianobooster.dir/TrackList.o
[ 80%] Building CXX object CMakeFiles/pianobooster.dir/Rating.o
[ 82%] Building CXX object CMakeFiles/pianobooster.dir/Bar.o
[ 84%] Building CXX object CMakeFiles/pianobooster.dir/moc_QtWindow.o
[ 86%] Building CXX object CMakeFiles/pianobooster.dir/moc_GlView.o
[ 88%] Building CXX object CMakeFiles/pianobooster.dir/moc_GuiTopBar.o
[ 90%] Building CXX object CMakeFiles/pianobooster.dir/moc_GuiSidePanel.o
[ 92%] Building CXX object CMakeFiles/pianobooster.dir/moc_GuiMidiSetupDialog.o
[ 94%] Building CXX object CMakeFiles/pianobooster.dir/moc_GuiKeyboardSetupDialog.o
[ 96%] Building CXX object CMakeFiles/pianobooster.dir/moc_GuiPreferencesDialog.o
[ 98%] Building CXX object CMakeFiles/pianobooster.dir/moc_GuiLoopingPopup.o
[100%] Building CXX object CMakeFiles/pianobooster.dir/qrc_application.o
Linking CXX executable pianobooster
[100%] Built target pianobooster

------------------

Have a good weekend all!

Dan

[Louis B.]

Hi Dan,

I finally got the looping working which you requested (It was much more tricky than I thought).

Regarding the core dump can you get me a back trace. in a new builddebug directory turn on debug with
 
  cmake -DCMAKE_BUILD_TYPE=Debug  ../src

then start pb with gdb, type

  gdb pianobooster

then type

  run

when it core dumps type

   bt

and then send me the output directly on

  l o u i s j b a r m a n [at] g o o g l e m a i l [dot] c o m

Does deleting the config file fix the problem (saving the original for me) ? type

    mv ~/.config/PianoBooster/Piano\ Booster.ini  ~/.config/PianoBooster/Piano\ Booster.ini.bug

Thanks

Louis

[danboid]

Hi LB!

I've just e-mailed you the gdb backtrace a few minutes but I forgot to mention

mv ~/.config/PianoBooster/Piano\ Booster.ini  ~/.config/PianoBooster/Piano\ Booster.ini.bug

~/.config/PianoBooster/Piano\ Booster.ini didn't exist on my drive anyway so moving that made no difference

[danboid]

Hi LB!

I tried mailing this to you but apparently you didn't get it for some reason so here's the backtrace:

dan@iatn-ati:~/src/pianobooster-src-0.6.2/build$ gdb pianobooster
GNU gdb 6.8-debian
Copyright (C) 2008 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-linux-gnu"...
(gdb) run
Starting program: /home/dan/src/pianobooster-src-0.6.2/build/pianobooster
[Thread debugging using libthread_db enabled]
[New Thread 0x7f0b70512760 (LWP 4618)]
/home/dan/.themes/Dust/gtk-2.0/gtkrc:708: Murrine configuration option
"highlight_ratio" will be deprecated in future releases. Please use
"highlight_shade" instead.
/home/dan/.themes/Dust/gtk-2.0/gtkrc:709: Murrine configuration option
"lightborder_ratio" will be deprecated in future releases. Please use
"lightborder_shade" instead.
unknown chip id 0x94c9, can't guess.
Error: Zero tracks in SMF file
Opening song /home/dan/autosave/
Error: Zero tracks in SMF file
Using Pianist Channels 1 + 2

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7f0b70512760 (LWP 4618)]
0x0000000017cc5ab9 in ?? ()
(gdb) bt
#0  0x0000000017cc5ab9 in ?? ()
#1  0x00007f0b6f8ec7d1 in glXWaitX () from /usr/lib/libGL.so.1
#2  0x00007f0b6fe0a6bd in QGLWidget::resizeEvent ()
  from /usr/lib/libQtOpenGL.so.4
#3  0x00007f0b6eeb07ca in QWidget::event () from /usr/lib/libQtGui.so.4
#4  0x00007f0b6fdd9c51 in QGLWidget::event () from /usr/lib/libQtOpenGL.so.4
#5  0x00007f0b6ee5f83d in QApplicationPrivate::notify_helper ()
  from /usr/lib/libQtGui.so.4
#6  0x00007f0b6ee67a2a in QApplication::notify () from /usr/lib/libQtGui.so.4
#7  0x00007f0b6e47875c in QCoreApplication::notifyInternal ()
  from /usr/lib/libQtCore.so.4
#8  0x00007f0b6eeaf1e3 in QWidgetPrivate::sendPendingMoveAndResizeEvents ()
  from /usr/lib/libQtGui.so.4
#9  0x00007f0b6eeb56f3 in QWidgetPrivate::show_helper ()
  from /usr/lib/libQtGui.so.4
#10 0x00007f0b6eeb5dba in QWidget::setVisible () from /usr/lib/libQtGui.so.4
#11 0x00007f0b6eeb5b8e in QWidgetPrivate::showChildren ()
  from /usr/lib/libQtGui.so.4
#12 0x00007f0b6eeb570f in QWidgetPrivate::show_helper ()
  from /usr/lib/libQtGui.so.4
#13 0x00007f0b6eeb5dba in QWidget::setVisible () from /usr/lib/libQtGui.so.4
#14 0x00007f0b6eeb5b8e in QWidgetPrivate::showChildren ()
  from /usr/lib/libQtGui.so.4
---Type <return> to continue, or q <return> to quit---
#15 0x00007f0b6eeb570f in QWidgetPrivate::show_helper ()
  from /usr/lib/libQtGui.so.4
#16 0x00007f0b6eeb5dba in QWidget::setVisible () from /usr/lib/libQtGui.so.4
#17 0x0000000000421e03 in QWidget::show (this=0x7fff78546740)
   at /usr/include/qt4/QtGui/qwidget.h:473
#18 0x0000000000426035 in Window (this=0x7fff78546740)
   at /home/dan/src/pianobooster-src-0.6.2/src/QtWindow.cpp:94
#19 0x0000000000421d07 in main (argc=1, argv=0x7fff78546948)
   at /home/dan/src/pianobooster-src-0.6.2/src/QtMain.cpp:35
(gdb)

[Louis B.]

Hi Dan,

Try commenting line 94 in QtWindow.cpp so it reads

    //show();

Does that fix the problem?

Also I think you had this running before under Linux. Do the old versions of PB work? Or is this a new machine to PB?

Louis

[danboid]

Hi Louis!

Yes, you're right, I did have an earlier version of PB running OK under Linux, but I was running either intrepid or hardy then, intrepid I think it was- x86-64. Since then I've upgraded to Jaunty but it hasn't been going all that well and I think I'll go back to intrepid - wish I'd mirrored my install before jumping to jaunty. It was the pull of ext4- I couldn't resist it any longer. ext4 does indeed fsck MUCH faster than ext3 and seems visibly faster overall but Jaunty is still shaky on my laptop.

After commenting that line, pb gets a bit closer to starting as the keyboard/MIDI setup window successfully appears now, I can choose my MIDI IN/OUTs but then as soon as I click on OK it dumps on me again! Bah! Here's the new backtrace:

dan@iatn-ati:~/src/pianobooster-src-0.6.2/build$ gdb pianobooster
GNU gdb 6.8-debian
Copyright (C) 2008 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-linux-gnu"...
(gdb) run
Starting program: /home/dan/src/pianobooster-src-0.6.2/build/pianobooster
[Thread debugging using libthread_db enabled]
[New Thread 0x7fa5b3e7a760 (LWP 4556)]
/home/dan/.themes/Dust/gtk-2.0/gtkrc:708: Murrine configuration option "highlight_ratio" will be deprecated in future releases. Please use "highlight_shade" instead.
/home/dan/.themes/Dust/gtk-2.0/gtkrc:709: Murrine configuration option "lightborder_ratio" will be deprecated in future releases. Please use "lightborder_shade" instead.
unknown chip id 0x94c9, can't guess.
Error: Zero tracks in SMF file
Opening song /home/dan/autosave/
Error: Zero tracks in SMF file
Using Pianist Channels 1 + 2
[New Thread 0x7fa5a4601950 (LWP 4559)]

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7fa5b3e7a760 (LWP 4556)]
0x00007fa5b32547cf in glXWaitX () from /usr/lib/libGL.so.1
(gdb) bt
#0  0x00007fa5b32547cf in glXWaitX () from /usr/lib/libGL.so.1
#1  0x00007fa5b37726bd in QGLWidget::resizeEvent () from /usr/lib/libQtOpenGL.so.4
#2  0x00007fa5b28187ca in QWidget::event () from /usr/lib/libQtGui.so.4
#3  0x00007fa5b3741c51 in QGLWidget::event () from /usr/lib/libQtOpenGL.so.4
#4  0x00007fa5b27c783d in QApplicationPrivate::notify_helper () from /usr/lib/libQtGui.so.4
#5  0x00007fa5b27cfa2a in QApplication::notify () from /usr/lib/libQtGui.so.4
#6  0x00007fa5b1de075c in QCoreApplication::notifyInternal () from /usr/lib/libQtCore.so.4
#7  0x00007fa5b28171e3 in QWidgetPrivate::sendPendingMoveAndResizeEvents () from /usr/lib/libQtGui.so.4
#8  0x00007fa5b281d6f3 in QWidgetPrivate::show_helper () from /usr/lib/libQtGui.so.4
#9  0x00007fa5b281ddba in QWidget::setVisible () from /usr/lib/libQtGui.so.4
#10 0x00007fa5b281db8e in QWidgetPrivate::showChildren () from /usr/lib/libQtGui.so.4
#11 0x00007fa5b281d70f in QWidgetPrivate::show_helper () from /usr/lib/libQtGui.so.4
#12 0x00007fa5b281ddba in QWidget::setVisible () from /usr/lib/libQtGui.so.4
#13 0x00007fa5b281db8e in QWidgetPrivate::showChildren () from /usr/lib/libQtGui.so.4
#14 0x00007fa5b281d70f in QWidgetPrivate::show_helper () from /usr/lib/libQtGui.so.4
#15 0x00007fa5b281ddba in QWidget::setVisible () from /usr/lib/libQtGui.so.4
#16 0x0000000000421e03 in QWidget::show (this=0x7fffbbeb10b0) at /usr/include/qt4/QtGui/qwidget.h:473
#17 0x0000000000421d13 in main (argc=1, argv=0x7fffbbeb12b8) at /home/dan/src/pianobooster-src-0.6.2/src/QtMain.cpp:36
(gdb)

[Louis B.]

Try disabling compiz, do other open gl apps run ok?

[danboid]

Hi Louis!

I've had compiz disabled the whole time I've been trying to get 0.6.2 working. Yes, I can use opengl apps (3D games like mu-cade and foobillard) fine, then run smoothly and accelerated as I've got the latest, official (fglrx) ATI driver installed. To test I just tried running mu-cade - no problem. Piano booster still segfaults after the midi options settings screen, as per my last backtrace.

Are you using QT 4.5? That's the standard in Jaunty? I won't be the only one wanting to use pb under Jaunty as it will be released soon. It might be a good idea to install virtualbox and install Jaunty under a virtual machine if you've got 10/15GB or whatever HD space to spare for such a venture?

Dan

[Louis B.]

Do the old version of PB segfaults which are still available on source forge?


On Sun, Apr 5, 2009 at 6:42 PM, danboid (via Nabble)
<[hidden email]> wrote:

> Hi Louis!
>
> I've had compiz disabled the whole time I've been trying to get 0.6.2
> working. Yes, I can use opengl apps (3D games like mu-cade and foobillard)
> fine, then run smoothly and accelerated as I've got the latest, official
> (fglrx) ATI driver installed. To test I just tried running mu-cade - no
> problem. Piano booster still segfaults after the midi options settings
> screen, as per my last backtrace.
>
> Are you using QT 4.5? That's the standard in Jaunty? I won't be the only one
> wanting to use pb under Jaunty as it will be released soon. It might be a
> good idea to install virtualbox and install Jaunty under a virtual machine
> if you've got 10/15GB or whatever HD space to spare for such a venture?
>
> Dan
>
> ________________________________
> View message @
> http://n2.nabble.com/0.6.2-core-dumps-under-jaunty-x64-tp2585538p2589349.html
> To start a new topic under Piano Booster, email
> [hidden email]
> To unsubscribe from Piano Booster, click here.
>

[danboid]

Yes- I just compiled 0.5.1 and that segfaults before any Window gets displayed too. Jaunty doesn't like PB :(

[danboid]

Hi Louis!

I've just finished installing Debian Lenny (5.0) x86-64 on my laptop and thankfully 0.6.2 compiles and runs without problem - or at least I've not encountered any so far, touch wood! ;)

Repeat mode is working fine and after having a quick test of it under Debian there doesn't seem to be any noticeable jump at the end of repeated sequences, even when there is a note on the first beat of the next bar after the loop.

Installing Debian took longer than I'd hoped so I don't have much time to test it now but I've got to have a quick play at least. Its looking very promising though - how about getting this into the Debian/Ubuntu repos?

Which distro are you running LB?

[Louis B.]

Hey Dan,

That is great that it is working OK for you now. Have you wiped your old system now? I guess the problem could be with the  jaunty x64 distro. I don't think PB is doing anything wrong enough to make it core dump. We should really file a bug report to jaunty x64.

It would be great if you or some one else could post a request about getting this into the Debian/Ubuntu repos.


Just posted my distro details to the sub forum.

L.

[danboid]

Hibernate and suspend didn't work on my laptop under Debian Lenny so I apt-get dist-upgrade 'd to sid (squeeze, i think its called now) x64 and that fixed my hibernate and suspend. The upgrade went very smoothly and I can boot from grub to Firefox under GNOME in about 15s now. The reason I'm telling you is just so you know that pb 0.6.2 compiles and runs fine under squeeze (Debian unstable/sid amd64) too.

Anybody tried compiling pb under jaunty x86 yet?

[danboid]

Don't know what I'm on about with these Debian release codenames - according to Debian multimedia squeeze is the codename for the current Debian Testing. I'm running unstable which is squeeze+1 and I don't know what the codename for that is, if anyone is really bothered?

Bad thing about sid is its using xorg 1.6 and the current, latest ati linux drivers don't support xorg 1.6 yet and I've been unable to get the open source radeon / radeonhd drivers to give smooth playback when playing a fast song with lots of chords in it- think i'll have to re-install intrepid somewhere and try under that as the official ati drivers are supported on that.

[danboid]

I'm back on Intrepid amd64 now but I've still got my Debian sid install so I felt the need to return here to complete the tale of my pb testing saga with the Debian family and a Radeon HD2400 laptop.

I can play back fast songs with lots of chords smoothly no problems when I have the fglrx (non-free official ati driver) activated and compiz (Visual FX) turned off. If I enable compiz then pb is still just as smooth but you get an intermittent flickering of the piano roll display area.

I think it would be good to add a "minimum display requirements" section to the 'Requirements' page on the pianobooster site. People using VESA or an unaccelerated video driver or with a screen resolution less than, what, 1024x600, aren't really going to be able to use pb except under limited, not much fun circumstances.

[Frank Phillips]

I think I can explain this.

Intrepid uses mesa 7.2
Jaunty uses mesa 7.4
Lenny uses mesa 7.0
Squeeze uses mesa 7.0
Sid uses mesa 7.4

Here on Archlinux we just upgraded from mesa 7.2 to 7.4. We split out libGL into its own package, and I can roll it back to libgl-7.2 and PB-0.6.2 works, but with 7.4 it segfaults.

[fp@viron pianobooster]$ pacman -Ql libgl
libgl /usr/
libgl /usr/lib/
libgl /usr/lib/libGL.so
libgl /usr/lib/libGL.so.1
libgl /usr/lib/libGL.so.1.2
libgl /usr/lib/xorg/
libgl /usr/lib/xorg/modules/
libgl /usr/lib/xorg/modules/dri/
libgl /usr/lib/xorg/modules/dri/libdricore.so
libgl /usr/lib/xorg/modules/dri/swrast_dri.so
libgl /usr/lib/xorg/modules/extensions/
libgl /usr/lib/xorg/modules/extensions/libglx.so

The debian family splits it into libgl1-mesa-swx11.

https://launchpad.net/ubuntu/+source/mesa
https://launchpad.net/ubuntu/jaunty/i386/libgl1-mesa-swx11/7.4-0ubuntu1
http://packages.debian.org/lenny/libgl1-mesa-swx11

[Louis B.]

Thanks for the info

I don't believe that I am actually doing anything wrong (but if I am i'll fix it)

Open GL does not come with any window manger and so I have chosen to use the QT one. I am guessing that now that libGL has it's own library it is not being correctly initialise since it core dumps as soon as the glwindow is shown.

Question is it just PB that core dumps with  mesa 7.4/libgl-7.4 or do any other programs do this? In particular does Stellarium: 0.10.2 work? as they are also using the QT/GL combination.

ideally the fix should go into the whatever program or library that is causing the problem. (if any one can see any problems in the PB code then please let me know) could you/dan report this bug to mesa, thanks

By the way is there any chance of getting PB packaged for Archlinux? It should be very straight forward as there are so few dependencies. please let me know if there are any changes need to be made CMakeLists.txt


Thanks

Louis

[Frank Phillips]

To clarify, libgl has had its own package, not just with this latest update. Since Jaunty is seeing the same problem, it's probably not a packaging problem.

Stellarium 0.10.2 works with libgl-7.4 - confirmed. Maybe they have already patched a change that occurred between 7.2 and 7.4. There is a myth user who claims a libgl problem on jaunty:
http://www.gossamer-threads.com/lists/mythtv/users/378205#378205
I am on Intel 965 graphics here. Sounds like danboid is on ATI hardware, I don't know if they replace libgl.

We did update QT at the same time (just a point release), but I'm not discounting it yet.
Here is my backtrace:

Using Pianist Channels 2 + 3

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb6aa6710 (LWP 7946)]
0x00207075 in ?? ()
(gdb) bt
#0  0x00207075 in ?? ()
#1  0xb7d222d2 in glXWaitX () from /usr/lib/libGL.so.1
#2  0xb7e29bff in QGLWidget::resizeEvent () from /usr/lib/libQtOpenGL.so.4
#3  0xb74c2006 in QWidget::event () from /usr/lib/libQtGui.so.4
#4  0xb7df6d4c in QGLWidget::event () from /usr/lib/libQtOpenGL.so.4
#5  0xb746bdec in QApplicationPrivate::notify_helper ()
   from /usr/lib/libQtGui.so.4
#6  0xb7474162 in QApplication::notify () from /usr/lib/libQtGui.so.4
#7  0xb702b1bb in QCoreApplication::notifyInternal ()
   from /usr/lib/libQtCore.so.4
#8  0xb74c099c in QWidgetPrivate::sendPendingMoveAndResizeEvents ()
   from /usr/lib/libQtGui.so.4
#9  0xb74c7626 in QWidgetPrivate::show_helper () from /usr/lib/libQtGui.so.4
#10 0xb74c7d53 in QWidget::setVisible () from /usr/lib/libQtGui.so.4
#11 0xb74c7b28 in QWidgetPrivate::showChildren () from /usr/lib/libQtGui.so.4
#12 0xb74c7651 in QWidgetPrivate::show_helper () from /usr/lib/libQtGui.so.4
#13 0xb74c7d53 in QWidget::setVisible () from /usr/lib/libQtGui.so.4
#14 0xb74c7b28 in QWidgetPrivate::showChildren () from /usr/lib/libQtGui.so.4
#15 0xb74c7651 in QWidgetPrivate::show_helper () from /usr/lib/libQtGui.so.4
#16 0xb74c7d53 in QWidget::setVisible () from /usr/lib/libQtGui.so.4
#17 0x0805c5e5 in Window::Window ()
#18 0x08059bbe in main ()

PS:
Only Arch Trusted Users (TUs) can post binary packages. It seems no TU has a need to package it, but any user can post a PKGBUILD, which is used to build (just untar and makepkg: easy enough).
http://aur.archlinux.org/packages.php?ID=24276

[Frank Phillips]

OK, Louis, you can stop debugging your code and reading API documentation now. I found the problem. What did it turn out to be?.......a permissions problem. That's right, with libgl 7.2 I was able to run PB without special privilege, but with 7.4 I have to have access to the dri device, and with DRI2, GEM is in the kernel now, etc.

[fp@viron ~]$ ls -l /dev/dri/card0 
crw-rw---- 1 root video 226, 0 2009-04-13 22:53 /dev/dri/card0

Add myself to the video group and I'm all set. To Dan: check your groups on jaunty and/or add the "DRI" Section to your xorg.conf, if you have one (I run without an xorg.conf).