hardy, locking assertion failure, xorg/libsdl
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Ubuntu Mobile Edition |
Confirmed
|
Undecided
|
Unassigned | ||
X.Org X server |
Fix Released
|
Medium
|
|||
libxcb (Debian) |
Fix Released
|
Unknown
|
|||
libxcb (Mandriva) |
Confirmed
|
Undecided
|
Unassigned | ||
libxcb (Ubuntu) |
Fix Released
|
High
|
Bryce Harrington | ||
Gutsy |
Invalid
|
Undecided
|
Unassigned | ||
Hardy |
Won't Fix
|
High
|
Unassigned | ||
Intrepid |
Won't Fix
|
High
|
Bryce Harrington | ||
openoffice.org (Ubuntu) |
Invalid
|
Undecided
|
Unassigned | ||
Gutsy |
Invalid
|
Undecided
|
Unassigned | ||
Hardy |
Invalid
|
Undecided
|
Unassigned | ||
Intrepid |
Invalid
|
Undecided
|
Unassigned | ||
pygtk (Ubuntu) |
Invalid
|
Low
|
Unassigned | ||
Gutsy |
Invalid
|
Undecided
|
Unassigned | ||
Hardy |
Invalid
|
Undecided
|
Unassigned | ||
Intrepid |
Invalid
|
Low
|
Unassigned |
Bug Description
Binary package hint: xorg
If I run Savage (a 3D non-open source game for Linux, available for download on the internet), I get loads of these messages:
silverback.bin: ../../src/
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 libs/libSDL-
#5 libs/libSDL-
#6 libs/libSDL-
#7 libs/libSDL-
#8 libs/libSDL-
#9 ./silverback.bin [0x80a95d3]
silverback.bin: ../../src/
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 libs/libSDL-
#5 libs/libSDL-
#6 libs/libSDL-
#7 libs/libSDL-
#8 libs/libSDL-
#9 ./silverback.bin [0x80a95d3]
silverback.bin: ../../src/
Locking assertion failure. Backtrace:
Segmentation fault (core dumped)
I searched the web for this error, and it seems like running another 3d game, neverwinter nights, causes the same problem on Hardy.
Here's a link to that:
http://
Alexander Rødseth (alexanro) wrote : | #1 |
Alexander Rødseth (alexanro) wrote : | #2 |
Timo Aaltonen (tjaalton) wrote : | #3 |
Thanks, I've forwarded this upstream.
Changed in libxcb: | |
importance: | Undecided → High |
status: | New → Confirmed |
Changed in xorg-server: | |
status: | Unknown → Confirmed |
Antoine Pairet (b-ly) wrote : Re: hardy, locking assertion failure: Java | #4 |
- backtrace_frostwire Edit (2.2 KiB, text/plain)
Hi!
I have a problem similar to this when I try to run Java based applications in Hardy Heron:
* frostwire
* matlab
"Locking assertion failure."
"java: xcb_xlib.c:82: xcb_xlib_unlock: Assertion `c->xlib.lock' failed."
In attachment, you 'll find complete backtrace of the problem.
Antoine Pairet (b-ly) wrote : | #5 |
Antoine Pairet (b-ly) wrote : | #6 |
Is this bug duplicate of: https:/
Nthalk (phocis) wrote : Re: [Bug 185311] Re: hardy, locking assertion failure, xorg/libsdl | #7 |
Oh, damnit. This isn't a bug at all, it's user error.
I installed the application as root.
I am very sorry for wasting your time, please delete this bug.
Thanks for a great distro!
On Jan 27, 2008 7:25 AM, b_ly <email address hidden> wrote:
> Is this bug duplicate of: https:/
>
> --
> hardy, locking assertion failure, xorg/libsdl
> https:/
> You received this bug notification because you are a direct subscriber
> of a duplicate bug.
>
Alexander Rødseth (alexanro) wrote : | #8 |
Nthalk, you're not talking about _this_ bug, right?
JeremyChapman (zz9plural9zalpha) wrote : | #9 |
I think I am seeing same bug in vmware-server on hardy amd64
----
Locking assertion failure. Backtrace:
#0 /usr/lib32/
#1 /usr/lib32/
#2 /usr/lib32/
#3 /usr/lib32/
#4 /usr/lib/
#5 /usr/lib/
#6 /usr/lib/
#7 /usr/lib/
#8 /usr/lib/
#9 /usr/lib/
#10 /usr/lib/
#11 /usr/lib/
#12 /usr/lib/
#13 /usr/lib/
#14 /usr/lib/
#15 /usr/lib/
#16 /usr/lib/
#17 /usr/lib/
#18 /usr/lib/
#19 /usr/lib/
vmware: ../../src/
Aymeric (mulx) wrote : | #10 |
I have the same bug when I want to run any java programs (netbeans 5.5 ; netbeans 6 ; java web start )
I am on hardy upgraded from a feisty. sun-java6 is installed
Runing java web start
petit@petit-
Java(TM) Web Start 1.6.0_04 Launching: /usr/lib/
/usr/lib/
-Xbootclasspat
-classpath
/usr/lib/
-Djava.
-DtrustProxy=true
-Xverify:remote
-Djnlpx.
-Djnlpx.
com.sun.
-viewer
petit@petit-
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 /usr/lib/
#5 /usr/lib/
#6 /usr/lib/
#7 [0xb4c943aa]
#8 [0xb4c8cf0d]
#9 [0xb4c8cf0d]
#10 [0xb4c8a249]
#11 /usr/lib/
#12 /usr/lib/
#13 /usr/lib/
#14 /usr/lib/
#15 /usr/lib/
#16 [0xb4c943aa]
#17 [0xb4c8cda7]
#18 [0xb4c8a249]
#19 /usr/lib/
java: xcb_xlib.c:82: xcb_xlib_unlock: Assertion `c->xlib.lock' failed.
petit@petit-
Runing netbeans 6 :
petit@petit-
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 /usr/lib/
#5 /usr/lib/
#6 /usr/lib/
#7 [0xb5b5eecd]
#8 [0xb5b57edd]
#9 [0xb5b57edd]
#10 [0xb5b55249]
#11 /usr/lib/
#12 /usr/lib/
#13 /usr/lib/
#14 ...
linovski (avelinorego) wrote : | #11 |
Same bug here
hardy alpha 4
jdk6 / netbeans
linovski (avelinorego) wrote : | #12 |
for now I solved the problem this way:
http://
(
Try this as root:
sed -i 's/XINERAMA/
The i686 users should use:
sed -i 's/XINERAMA/
)
linovski (avelinorego) wrote : | #13 |
by the way, on ubuntu:
sudo sed -i 's/XINERAMA/
Torsten Spindler (tspindler) wrote : | #14 |
The java problem is also discussed here: http://
A work-around seems to exist with
$ export LIBXCB_
Ferran Rius (frius64) wrote : | #15 |
$ export LIBXCB_
Didn't have any effect in my laptop, however,
sudo sed -i 's/XINERAMA/
solved the problem.
Alexander Rødseth (alexanro) wrote : | #16 |
export LIBXCB_
Alexander Rødseth (alexanro) wrote : | #17 |
(only for Java)
John Leach (johnleach) wrote : | #18 |
Confirmed here too with Netbeans 6. Worked fine with Gutsy.
export LIBXCB_
Greg A (etulfetulf) wrote : | #19 |
"sudo sed -i 's/XINERAMA/
This worked for me - geogebra had refused to load.
andrewpmk (andrewpmk) wrote : | #20 |
Confirmed with JOSM (an editor for openstreetmap.org). The workaround works.
Timo Aaltonen (tjaalton) wrote : | #21 |
Some of you have issues with java, which means bug 87947. This is about silverback
In freedesktop.org Bugzilla #14211, Josh Triplett (joshtriplett) wrote : | #22 |
Jamey and I just announced a set of changes to XCB and Xlib/XCB which, among other things, should address all the outstanding synchronization problems that we know of. Could anyone experiencing this bug please build XCB and Xlib with the patches found at http://
Jamey Sharp (sharpone) wrote : | #23 |
Josh and I just announced a set of changes to XCB and Xlib/XCB which, among other things, should address all the outstanding assertion failures that we know of. Could anyone experiencing this bug please build XCB and Xlib with the patches found at http://
Ryan Novosielski (novosirj) wrote : | #24 |
I get this also using the UPEK fingerprint X display. I forget the actual name of it, but pam_bioapi brings up a dialog which now crashes (during KDM, it crashes KDM too, generally it just prints an error to the offending console).
fabius (fabio-ballatore) wrote : | #25 |
I verify the same problem with lotus notes client v.8.0.1 installation, I supose e it's a eclipse problem with java 5 and 6.
I try to set LIBXCB_
solve the problem
I try to repet the installation with lotus notes 8.0.0. but the result not change.
No problem with ubuntu 7.10
Thanks
joe williams (joetify) wrote : | #26 |
I am having a similar issue with java applications, export LIBXCB_
cherry (dev-caltech) wrote : | #27 |
sed: can't read /usr/lib/
what to do ????
Slava (slava-slavix) wrote : | #28 |
same problem ubuntu 7.10, 8.04. tried all proposed workarounds
export LIBXCB_
export LIBXCB_
sudo sed -i 's/XINERAMA/
none worked for me. :(
Christian Zeller (chrizel) wrote : | #29 |
> sed: can't read /usr/lib/
> what to do ????
The directory name changed because of a newer java version on current Hardy systems. So change 04 to 05:
sudo sed -i 's/XINERAMA/
Worked for me.
Ryszard Szopa (ryszard-szopa) wrote : | #30 |
Apparently it has nothing to do with Java. I get the same error when trying to start Emacs. It seems to be related to today's Ubuntu update...
Prophet (prophet005) wrote : | #31 |
Just for conformance i get the error too. It appears whenever i start a java gui application that uses swing/awt. Applications that uses swt (e.g. eclipse) don't appear to have the error.
Here the error output:
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 /usr/lib/
#5 /usr/lib/
#6 /usr/lib/
#7 [0xb5cc5ecd]
#8 [0xb5cbeedd]
#9 [0xb5cbeedd]
#10 [0xb5cbc249]
#11 /usr/lib/
#12 /usr/lib/
#13 /usr/lib/
#14 /usr/lib/
#15 /usr/lib/
#16 [0xb5cc5ecd]
#17 [0xb5cbed77]
#18 [0xb5cbc249]
#19 /usr/lib/
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 /usr/lib/
#5 /usr/lib/
#6 /usr/lib/
#7 /usr/lib/
#8 [0xb5cc5ecd]
#9 [0xb5cbeedd]
#10 [0xb5cbeedd]
#11 [0xb5cbc249]
#12 /usr/lib/
#13 /usr/lib/
#14 /usr/lib/
#15 /usr/lib/
#16 /usr/lib/
#17 [0xb5cc5ecd]
#18 [0xb5cbed77]
#19 [0xb5cbc249]
Prophet (prophet005) wrote : | #32 |
SIDENOTE: Besides the error output i havn't noticed any crashes or something else going wrong within the applications.
fieldyweb (fieldyweb) wrote : | #33 |
I've had VMware server working previously, and built an FTP Server using a VMWare image, however after updating today, i too am now getting this error on VMWare
vmware: ../../src/
rerunning sudo vmware-config.pl didn't resolve this, but the modules did reload
I tried reinstalling VMware, still not working
8.04 AMD64
fieldyweb (fieldyweb) wrote : | #34 |
Managed to Fix this, it was the VMware console causing the problem, but running
sudo ln -s /usr/lib32 /usr/l32
sudo sed -i -e 's/usr\
sudo sed -i -e 's/usr\
Kjow (antispammoni) wrote : | #35 |
Thank you linovski!
Problem fixed with: sudo sed -i 's/XINERAMA/
:)
Prophet (prophet005) wrote : | #36 |
Does this have to be done manually or will the updates today/tomorrow fox it? Current update didn't...
zbharucha (zubinb) wrote : | #37 |
I've tried all the above suggestions and none of them seem to work. This still happens when invoking MATLAB:
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /opt/matlab74/
#4 /opt/matlab74/
#5 /opt/matlab74/
#6 /opt/matlab74/
#7 [0x7f5dbf345f9b]
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 /opt/matlab74/
#5 /opt/matlab74/
#6 /opt/matlab74/
#7 /opt/matlab74/
#8 [0x7f5dbf345f9b]
The MATLAB window comes up but none of the menus are displayed.
Hardy Beta, amd64, java-6-sun-1.6.0.05
sothis (janos-dev) wrote : | #38 |
This really seems like a libX11 problem, not a specific java one. I am currently developing an opengl application an stumbled on that problem while testing my program on ubuntu hardy amd64. In my case the problem occurs when calling glXSwapBuffers(), but not continuesly. therefore it's hard to track down, but i am on it at the moment. Until i found something specific here's the backtrace of my application so far:
#0 0x00007f48ea83a095 in raise () from /lib/libc.so.6
#1 0x00007f48ea83baf0 in abort () from /lib/libc.so.6
#2 0x00007f48ea8332df in __assert_fail () from /lib/libc.so.6
#3 0x00007f48ea54c17d in _XGetXCBBuffer () from /usr/lib/
#4 0x00007f48ea54d343 in _XSend () from /usr/lib/
#5 0x00007f48eb7c9625 in __glXFlushRende
#6 0x00007f48eb7cb35f in __glXSetupForCo
#7 0x00007f48eb7c7628 in glXSwapBuffers () from /usr/lib/libGL.so.1
#8 0x0000000000418795 in draw_scene (arg=0x7ffff3e4
at /home/laubej/
#9 0x00007f48eb30a3f7 in start_thread () from /lib/libpthread
#10 0x00007f48ea8dfb2d in clone () from /lib/libc.so.6
#11 0x0000000000000000 in ?? ()
Maybe someone else has an idea and can poke me into the right direction. It might be worth to notice that the rendering is done by MESA indirect glx, although i think that's not a problem with MESA at all.
bq (balique) wrote : | #41 |
The problem occurs also on Sun jdk j2sdk1.4.2_17 (not installed from .deb but from Sun installer, but info is important anyway).
The workaround seems to be very similar to other java versions. I tried only with simple gui program, so I cannot say if ti has any side effects.
sudo sed -i 's/XINERAMA/
Jørgen H. Seland (jorgen-fabeljet) wrote : | #42 |
Experiencing the bug while running JUnit tests on Eclipse 3.3.2 64-bit, sun-java6 6-03-0ubuntu2 (gutsy version, to avoid bug #174759), Hardy beta 5 AMD 64:
Does *not* work: export LIBXCB_
Does *not* work: export LIBXCB_
Works: sudo sed -i 's/XINERAMA/
For completeness, here is the backtrace:
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 /usr/lib/
#5 /usr/lib/
#6 /usr/lib/
#7 [0x7fa8249b9e77]
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 /usr/lib/
#5 /usr/lib/
#6 /usr/lib/
#7 /usr/lib/
#8 [0x7fa8249b9e77]
@les (alesstimec) wrote : | #43 |
Hey guys...
I saw somebody had this problem with matlab.. Me too.. But doing:
export MATLAB_
solved the problem for me.. By setting the MATLAB_JAVA variable you force matlab to use the specified java and it seems that 1.6.0.06 version that comes with Hardy does not cause this bug...
Hope this helps,
Ales
Scub (maxime-47) wrote : | #44 |
Neither "export LIBXCB_
Although the programms work, the notices remain.
Trying export the matlab java path later. This could fixed the error messages starting matlab.
Scub (maxime-47) wrote : | #45 |
Export the java path worked fine for matlab.
Just add export "MATLAB_
For maple11 it has to be done different:
sudo mv /opt/maple/
sudo ln -s /usr/bin/java /opt/maple/
(for 32bit Ubuntu)
Now both programms work just fine without giving any errors when started.
eagleflyer (wweagleflyer) wrote : | #46 |
Hi all,
I just updated to ubuntu 8.04, and started getting the same errors when I try to run MATLAB.
I tried all the suggestions above, and none worked except for
export MATLAB_
But this fixed the problem when launching MATLAB v2007a, but it did not fix the problem when launching MATLAB v2008a
The following:
sudo sed -i 's/XINERAMA/
sudo sed -i 's/XINERAMA/
export LIBXCB_
export LIBXCB_
all did not fix the problem with either version of MATLAB
Any ideas why MATLAB v2008 didn't work with export MATLAB_
Thanks
Sergei (msergei) wrote : | #47 |
I get this problem with my own Qt4 based application compiled against Ubuntu bundled Qt 4.3.4 library, thought this time around it looks like:
../../
The offending code should be something like
Never happened in Kubuntu 7.10.
MMarking (cpt-mocha) wrote : | #48 |
I'm also getting similar error messages running Tunapie out of the Hardy repositories. I talked to the author of Tunapie and he says it's related to the bug reported here.
python: ../../src/
I also get:
python: ../../src/
Aaron D. Campbell (aaroncampbell) wrote : | #49 |
- eclipse.log.txt Edit (14.1 KiB, text/plain)
These solutions don't work for me. You can see what look like related problems:
http://
http://
Here are the errors I get, and I attached a log that eclipse gave:
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /tmp/install.
#4 /tmp/install.
#5 /tmp/install.
#6 /tmp/install.
#7 [0xb28ceb28]
#8 [0xb28c8aeb]
#9 [0xb28c8aeb]
#10 [0xb28c61b4]
#11 /tmp/install.
#12 /tmp/install.
#13 /tmp/install.
#14 /tmp/install.
#15 /tmp/install.
#16 [0xb28ce458]
#17 [0xb28c8a14]
#18 [0xb28c61b4]
#19 /tmp/install.
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 /tmp/install.
#5 /tmp/install.
#6 /tmp/install.
#7 /tmp/install.
#8 [0xb28ceb28]
#9 [0xb28c8aeb]
#10 [0xb28c8aeb]
#11 [0xb28c61b4]
#12 /tmp/install.
#13 /tmp/install.
#14 /tmp/install.
#15 /tmp/install.
#16 /tmp/install.
#17 [0xb28ce458]
#18 [0xb28c8a14]
#19 [0xb28c61b4]
Warning: Cannot convert string "-b&h-lucidasan
Aaron D. Campbell (aaroncampbell) wrote : | #50 |
By removing GCJ, and then removing the directories in my ~ that had settings for the two programs, I was able to get them working. I still get the JAVA errors, but at least they work.
Michael T (michaeltandy) wrote : | #51 |
I've experienced this problem with MATLAB 7.4.1 (R2007a) installed in my user directory, on x86_64 - ashipika's suggestion of using
export MATLAB_
seems to fix the problem for me.
raboof (raboof) wrote : | #52 |
Adding export MATLAB_
tingri (parisochr) wrote : | #53 |
I've tried the different proposed solutions but none fixes the problem in my case. I still have sometimes the message:
../../src/
when calling wxWindow:
The most promising solution seems to be the one here:
http://
but it has still not be included in the last version of Xcb (http://
I hope the patches will be quickly tested Xcb updated...
pichulines (pichulines) wrote : | #54 |
this,
# sudo sed -i 's/XINERAMA/
solve the problem to me.
(java 1.6.0_03 on ubuntu hardy)
szemy (sz-tomika) wrote : | #55 |
Same here on xubuntu.
szemy@szemy-
name of display: :0.0
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/libGL.so.1 [0xb7d5e22b]
#4 /usr/lib/libGL.so.1 [0xb7d5d87a]
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 /usr/lib/libGL.so.1 [0xb7d5ddb1]
glxinfo: ../../src/
Aborted
kikujiro (lmfsreznor) wrote : | #56 |
I had the same problem. Neither solutions said here works so i keep searching and found this, that finally did it for me:
http://
Besides this is for Kubuntu It has worked for my Ubuntu 8.04
Good luck
Hunter_bt (hunterggl) wrote : | #57 |
(sorry for my English)
I use Kubuntu 8.04 (32-bit, DVD) and the problem is NDN (filemanager) crashes on startup.
NDN (Necromancer's Dos Navigator) is a freeware filemanager like Midnight Commander and Norton Commander.
Can anyone suggest what I should do? I'm still a newbie in Linux so I need detailed instructions. I appreciate any your help.
I guess I should:
1. Revert to the previous version of XCB and X11 which should work and use it till the bug is fixed. What should I revert?
2. Run some commands in console which will repair the bug for me. What commands?
3. I'd like to get the fixed version as soon as it's available. Where I should expect it? (I know how to compile source code and apply patches if necessary)
NDN crashes at startup when I run it konsole or xterm.
Runtime error 217 at B7F7B410 (Exception C0060000) TID=1.
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 /usr/lib/
#5 ./ndn [0x50b5cc]
#6 ./ndn [0x5080a2]
#7 ./ndn [0x5068a7]
#8 ./ndn [0x50f23f]
#9 ./ndn [0x50a8b4]
#10 [0xb7f7b420]
#11 /lib/tls/
#12 /lib/tls/
#13 /usr/lib/
#14 /usr/lib/
#15 /usr/lib/
#16 /usr/lib/
#17 ./ndn [0x50b627]
#18 ./ndn [0x509460]
#19 ./ndn [0x414327]
ndn: ../../src/
Runtime error 217 at B7F7B410 (Exception C0060000) TID=1.
Sean Sprouse (sprouse-sean) wrote : | #58 |
Matlab 2007b works if I use Metacity instead of Compiz. I have to start it in the terminal though and I get the same locking assertion failure message whether it works or not.
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/local/
#4 /usr/local/
#5 /usr/local/
#6 /usr/local/
#7 [0x7f99c7b00563]
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 /usr/local/
#5 /usr/local/
#6 /usr/local/
#7 /usr/local/
#8 [0x7f99c7b00563]
Ubuntu Hardy amd64
raynor512 (bowlingogre300-aim) wrote : | #59 |
Hi all,
Had the same prob with HH i386 trying to compile java progs with eclipse. All my problems where solved by installing the new 1.6 JRE from synaptic and setting the programs to use it in the build path.
Cheers
itsallcrap (itsallcrap-yahoo) wrote : | #60 |
- libSDL.UT2003 Edit (9.3 KiB, text/plain)
Lo there,
As if we need any further confirmation of this bug, I have the same problem with Unreal Tournament 2003 (native linux version).
Attached is the backtrace.
Halley (halley-zhao) wrote : | #61 |
I also found this issue for media/codecs
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 /usr/lib/
#5 /usr/lib/
#6 /usr/lib/
#7 /usr/lib/
#8 /usr/lib/
#9 /usr/lib/
#10 /usr/lib/
#11 /lib/libc.
#12 /usr/lib/
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 /usr/lib/
#5 /usr/lib/
#6 /opt/helix/
szemy (sz-tomika) wrote : | #62 |
Jessica Ji (jessica-ji) wrote : | #63 |
on ubuntu-mobile project, base on hardy image, also find this issue. scenario is full screen and window screen change, then easy video freeze and no response. so also affect "ubuntu-mobile" project.
itsallcrap (itsallcrap-yahoo) wrote : | #64 |
Thomas (szemy),
That tuxx-home topic seems to be specific to Matrox video cards.
Did this work for you on a non-Matrox card?
szemy (sz-tomika) wrote : | #65 |
No, but if you read it he says that it is not a driver problem...
Changed in ubuntu-mobile: | |
status: | New → Confirmed |
GrosBedo (gros-bedo) wrote : | #66 |
Tested on Mandriva 2008.1
Changed in libxcb: | |
status: | New → Confirmed |
GrosBedo (gros-bedo) wrote : | #67 |
- Panic_Button_Source_Code_2-6-1_13-buggy.tar.gz Edit (1.1 MiB, application/x-tar)
Tested on Mandriva 2008.1 with a simple Python program.
Here it is in the attachment. It happens when I try to open a wxMessageBox or wxDialog in a thread. You'll see the bug when trying to fill a report, at the moment the program should open a confirmation dialog. It's very random, sometimes it works, but most of the time no.
Spayder26 (spayder26) wrote : | #68 |
I also found using python 2.5.2 and pygtk in Ubuntu 8.04 amd64:
python: ../../src/
(dpy->request)) >= 0)' fallida.
Aborted
Spayder26 (spayder26) wrote : | #69 |
the freedesktop.org newsgroup ( http://
# [Xcb] Bug#481745: marked as done (libxcb: sloppy locking needs to be on by default) Debian Bug Tracking System
Does that mean the bug is fixed? If do, could any ubuntu maintainer update the repository's xcb packages?
Changed in libxcb: | |
status: | Unknown → Fix Released |
Fabian Ferreyra (fferreyra) wrote : | #70 |
On Ubuntu Hardy 8.04 i386, all patches applied, I had the same problem using the Java image libraries.
Using this command
sudo sed -i 's/XINERAMA/
it was fixed.
Changed in libxcb: | |
importance: | Undecided → High |
status: | New → Confirmed |
Stunts (f-pinamartins) wrote : | #71 |
I'm getting something similar when running openoffice.org in ubuntu 8.04 AMD64:
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 /usr/lib/
#5 /usr/lib/
#6 /usr/lib/
#7 /usr/lib/
#8 /usr/lib/
#9 /usr/lib/
#10 /usr/lib/
#11 /usr/lib/
#12 /usr/lib/
#13 /usr/lib/
#14 /usr/lib/
#15 /usr/lib/
#16 /usr/lib/
#17 /usr/lib/
#18 /usr/lib/
#19 /usr/lib/
However, as I stated in bug #240911, which was marked a duplicate of this one, I have managed a few work arounds, that turn out not to be perfect, but do the trick:
1. sudo apt-get remove openoffice.org-gtk
This will make openoffice look disgruntled from the rest of the theme in your desktop. It will also provoke other bugs such as oo.org impress in fulscreen staying behind the gnome panel
2. Create a new user account and run oo.org.
It will run fine on my machine, even with openoffice.org-gtk package installed.
3. Run as sudo
I know this should not be done, but for testing purposes, I tried it and it worked flawlessly...
Any better solutions?
Stunts (f-pinamartins) wrote : | #72 |
Forgot to mention - this only started to happen after today's (18-6-2008) updates!
madblueimp (madblueimp) wrote : | #73 |
I used to have proposed and backports software sources enabled, then had the problem with Open Office, which is marked as duplicate of this bug:
https:/
For this reason I decided to reinstall and never enable again proposed and backports as software sources for my productive system.
After todays updates, I have to uninstall openoffice.org-gtk to be able to use Open Office with my main user.
I wonder why the latest version of Open Office got into the stable updates?
I'm using another user profile as a workaround, but haven't found out yet which setting is associated with this problem. Neither of the config files in my home directory seem to be the source of the problem, as I did a complete sync (using rsync and then chown) of all configurations files and directories with another user profile and had no problem in using open office there.
Might be it has something to do with a system setting outside the home directory?
Currently, I'm using sux to be able to use Open Office without having to uninstall openoffice.org-gtk and without having to switch from my main users desktop:
-------
$ sudo apt-get install sux
$ sudo adduser newuser
$ chmod 777 document.odt
$ sux newuser openoffice document.odt
-------
Bryce Harrington (bryce) wrote : | #74 |
Fixed in intrepid with the libxcb 1.1-1.1 merge:
* Turn sloppy locking on by default (closes: #481745). It can be disabled
by setting the LIBXCB_
value).
Changed in libxcb: | |
status: | Confirmed → Fix Released |
Philip Belemezov (phible) wrote : | #75 |
Same problem here after today's updates (Hardy, amd64)
Mathieu Marquer (slasher-fun) wrote : | #76 |
Two users still encoutering the problem on ubuntu-fr.org : http://
Stunts (f-pinamartins) wrote : | #77 |
Any chance the interpid fix can be applied to hardy too?
I really don't know what I'm talking about here...
Adrian Bridgett (adrian-bridgett) wrote : | #78 |
looks like: https:/
In freedesktop.org Bugzilla #14211, Bryce Harrington (bryce) wrote : | #79 |
Several of those patches fail to apply to the last released libxcb. Are they in a git tree we can pull from?
W. Cravens (wcravens) wrote : | #80 |
This problem also started for me as of an update June 18th, 2008.
AMD64
NVidia video module installed via NVIDIA-
Using only restricted, universe and multiverse.
Sam_ (and-sam) wrote : | #81 |
Hi,
I've joined the bug #241324 and didn't know it's a duplicate, sorry.
My question now according to Hardy, should I wait for a fix or should I remove all the updated OpenOffice packages and reinstall? I mean reverse this:
(1:2.4.0-3ubuntu6) to 1:2.4.1-1ubuntu1
I guess my mistake was, not realizing that I had this source activated, which I deactivate now,
# deb http://
Thanks for a proposal.
Benjamin Drung (bdrung) wrote : | #82 |
- Locking_assertion_failure Edit (1.3 KiB, text/plain)
I get this error after updating openoffice.org to 1:2.4.1-1ubuntu1 (via hardy-updates). When I start for example oowriter I get this Locking assertion failure. But I can open some of my odt files within nautilus without this failure.
I saw that the bug is marked as fixed with libxcb 1.1-1.1. So I get the source and compiled it for hardy. But the assertion failure does not disappear.
William McKee (william-knowmad) wrote : | #83 |
I too have recently begun to get this error with some files. I have found that I can open one of the problematic files as the root user. I have no idea why I would get the libxcb-xlib.so.0 errors as a normal user but not as root. Let's hope this gets resolved quickly.
Adrian Bridgett (adrian-bridgett) wrote : | #84 |
started being bitten by this one yesterday after latest hardy updates.
setting LIBXCB_
purging openoffice.org-gtk openoffice.
I'm using the Crux theme with metacity which I think has been implicated before.
Andy Buchanan (andy-buchanans) wrote : | #85 |
I'm having the same problem with "Locking assertion failure".
uname -a: Linux abuchanan-desktop 2.6.24-19-generic #1 SMP Wed Jun 4 15:10:52 UTC 2008 x86_64 GNU/Linux
Package: openoffice.org 2.4.1-1ubuntu1
It only happens if I have openoffice.
Aleksander Demko (ademko) wrote : | #86 |
I still get issues when running MATLAB r2008 under Hard (32-bit and 64-bit). I've tried the sed thing, the SLOPPY_LOCK export and have all the updates installed as of today.
Basically, matlab over UNIX sockets (the default, DISPLAY=:0.0) seems fine, but over TCP sockets (DISPLAY=
Why can't we just remove the XCB infection/bridge layer from libX11 and revert it to the stock libX11 module? Let old libX11 programs use that, and new ones use XCB.
Thanks!
Aleksander Demko (ademko) wrote : | #87 |
Sorry, I meant to say "unusably slow". TCP/X11 transport of Matlab and other Java applications are unusably slow even with all the hacks and updates applied.
Aleksander Demko (ademko) wrote : | #88 |
Ok, here's a really ugly and large hack that seems (for me anyways) to get Matlab back to where it was before Hardy.
Basically, I took the libX11.so (and related) files from a different machine that didn't have a libxcb-backed libX11:
scp fedora3machine:
Then I placed this new set of X11 libs in my load path:
export LD_LIBRARY_
Then I ran Matlab with great success, even over TCP.
Note, make sure you use the proper (32-bit or 64-bit) libraries for you machine/
This hack is at best temporary to get you going. Ideally, the libX11-xcb library should emulate all the issues of the old libX11 for older apps, or it should be removed and let the old libX11 continue to serve those applications. Frameworks and applications can then migrate to libxcb at their own pace.
Bryce Harrington (bryce) wrote : | #89 |
The locking assertion failure sounds like bug 185311.
Since we're seeing a few issues with traces involving libxcb, I've prepared a non-xcb-enabled libx11 here: http://
Would someone mind testing with this noxcb libx11 and see if it makes any difference?
Bryce Harrington (bryce) wrote : | #90 |
er, ignore that first sentence in my reply. Obviously a paste-o.
Bryce Harrington (bryce) wrote : | #91 |
I suspect that this bug report is being used for more than one bug. Debian closed debbugs #481745, which identified one of these issues (in relation to Java and perhaps other packages), however other issues remain -- and are still present in intrepid.
In contacting upstream, they are aware of numerous problems similar to this one, which I gather is due to a deficiency in the design of the current Xlib / XCB interactions; they are in the process of re-architecting this code and will have it in the next release of Xlib/XCB.
The plan for Hardy is, if someone can validate the libx11~noxcb package above, to turn off XCB. For Intrepid hopefully this new Xlib/XCB will resolve the issues so we can continue shipping it; I'm re-opening the intrepid task for this work.
Changed in libxcb: | |
status: | Fix Released → Triaged |
Kirill Belokurov (kirill-belokurov) wrote : | #92 |
Bryce, I tested the non-xcb-enabled packages locally - the openoffice that was previously failing with such message https:/
~$ *** glibc detected *** /usr/lib/
======= Backtrace: =========
/lib/libc.
/lib/libc.
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/lib/libc.
/usr/lib/
I am not sure if the dependant packages (like openoffice.org-kde) should be rebuilt to fix this.
Stephen Jenuth (stephen-jenuth) wrote : | #93 |
I had the same problem on Kubuntu 8.04. I am running a AMD64 laptop with an nvidia card. I have installed the proprietary nvidia software which is distributed through its website.
The June 18 download of Openoffice crashed with the "Locking assertion failure" error.
I had previously been running NVIDIA-
NVIDIA-
Now, the only remaining problem is what on exiting, openoffice ends with the kde crash manager advising of the signal 11 which terminated the process. This is annoying, but openoffice is now usable.
Boudewijn (wankelwankel) wrote : | #94 |
As far as openoffice is concerned, a workaround that works for me, is to run via SSH:
me@gao:~$ ssh -Y localhost
me@localhost's password:
Linux gao 2.6.24-17-generic #1 SMP Thu May 1 13:57:17 UTC 2008 x86_64
The programs included with the Ubuntu system are free software;
the exact distribution terms for each program are described in the
individual files in /usr/share/
Ubuntu comes with ABSOLUTELY NO WARRANTY, to the extent permitted by
applicable law.
To access official Ubuntu documentation, please visit:
http://
Last login: Sat Jun 21 18:16:03 2008
me@gao:~$ soffice
me@gao:~$
I run Kubuntu; amd64 / ati fglrx.
Stephen Jenuth (stephen-jenuth) wrote : | #95 |
I took some time to look through the related bugs and fixes for this problem which have been identified.
They seem to relate to libxcb, but I tend to agree with the person who said that they are not really libxcb problems, but problems in the software which use the updated libraries (and which enforce a but of extra discipline on programs). The problem has been around a long time, but hit me on Thursday when I upgraded whatever was upgraded at that time.
Fixes have included installing a non-libxcb version of x11(I didn't try this), setting various environment variables to disable or modify the working of libxcb (this didn't work for me), modifying binaries of the libmawt.so library (this didn't work for me), modifying libxcb to be more forgiving (I didn't try this), running openoffice though a network x11 link (the ssh solution above, which also works for me), and installing a new (and presumably) better behaved version of the hardware libraries and modules (the nvidia solution I referenced above which mostly worked for me).
Since the problems seem to be outside the kernel, I doubt the new module I installed has anything to do with it. Instead, my betting in on the newer and better nvidia binaries which replaced the following libraries:
./usr/lib/
./usr/lib/
./usr/lib/
./usr/lib/
./usr/lib/
./usr/lib/
./usr/X11R6/
./usr/X11R6/
./usr/X11R6/
Some of these appear to be machine dependent libraries, but I would bet that the libGL.so, and libglx.so have something to do with the source of the problem. I suspect that until these problems are going to be around a long time until the various X11 libraries are all fixed and made consistent, and that people like me who run non-standard X11 libraries for specific cards are going to be in grief until they are all made consistent with the new discipline being imposed by libxcb.
Stephane Wirtel (OpenERP) (stephane-openerp) wrote : | #96 |
I get the same problem with openoffice.
If I launch oowriter with xterm I get this backtrace and oowriter is stopped, but oowriter works if I use the gnome's launcher
$ oowriter
$ Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 /usr/lib/
#5 /usr/lib/
#6 /usr/lib/
#7 /usr/lib/
#8 /usr/lib/
#9 /usr/lib/
#10 /usr/lib/
#11 /usr/lib/
#12 /usr/lib/
#13 /usr/lib/
#14 /usr/lib/
#15 /usr/lib/
#16 /usr/lib/
#17 /usr/lib/
#18 /usr/lib/
#19 /usr/lib/
strace:
stat(".", {st_mode=
open("/
fcntl(3, F_DUPFD, 10) = 10
close(3) = 0
fcntl(10, F_SETFD, FD_CLOEXEC) = 0
rt_sigaction(
rt_sigaction(
rt_sigaction(
rt_sigaction(
rt_sigaction(
rt_sigaction(
read(10, "#!/bin/
clone(child_
wait4(4294967295, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], 0, NULL) = 17714
--- SIGCHLD (Child exited) @ 0 (0) ---
read(10, "", 8192) = 0
exit_group(0) = ?
Process 17713 detached
Benjamin Drung (bdrung) wrote : | #97 |
I have two machines running hardy amd64. One has an Intel GMA chip (intel xorg driver) and one an AMD chip (fglrx driver) and both have the same problem with the crashing openoffice.org. The intel driver uses the libGL.so.1.2 provided by the libgl1-mesa-dri package and the flgrx driver (Catalyst 8.6) uses its own version of libGL.so.1.2. So is it really a hardware depending problem?
I think this problem with openoffice.org provided by hardy-updates only affects amd64 systems or is there anybody using hardy i386 with the same problem? There is a thread about this bug on ubuntuusers.de and all of them are using the 64 bit version: http://
Stephane Wirtel (OpenERP) (stephane-openerp) wrote : | #98 |
I forgot my configuration
Linux atlantis 2.6.24-19-generic #1 SMP Wed Jun 4 15:10:52 UTC 2008 x86_64 GNU/Linux
model name : Intel(R) Core(TM)2 Quad CPU Q9300 @ 2.50GHz
Sam_ (and-sam) wrote : | #99 |
- OpenOffice.org Writer.png Edit (51.8 KiB, image/png)
Hi,
after removing the packages openoffice.
btw. before this ominous update there were never any problems with OpenOffice.
System is amd_64 with open source driver Radeon X300.
Stunts (f-pinamartins) wrote : | #100 |
According to this thread (http://
One other user says he's using i386, but he's not sure. I'm expecting his reply to the thread with the output of "uname -a".
Stephane Wirtel (OpenERP) (stephane-openerp) wrote : | #101 |
Hi,
after removing the packages openoffice.
the bug is in openoffice.org-gtk and not in openoffice.
Nizar Kerkeni (nizarus) wrote : | #102 |
reported for openoffice.org here https:/
Changed in openoffice.org: | |
status: | New → Confirmed |
jhyder79 (jhyder79) wrote : | #103 |
Just figured I needed to add this...
This bug appears to be causing weird issues with Pro/Engineer Wildfire 3.0 on Kubuntu 8.04. It appears that several people are having the same issues that I am:
http://
The menus in Pro/E become very sluggish and if you launch Pro/E from a terminal window, there are a ton of "Locking assertion failure" messages like the following:
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 /usr/lib/libXt.so.6 [0xb778fce6]
#5 /usr/lib/
#6 /usr/local/
#7 /usr/lib/
#8 /usr/local/
#9 /usr/local/
#10 /usr/local/
#11 /usr/local/
#12 /usr/local/
#13 /usr/local/
#14 /lib/tls/
#15 /usr/local/
If I can be of any service in helping to squash this bug, just let me know.
oschwa2s (oliver-schwaneberg) wrote : | #104 |
By the way, I'm not a Kubuntu user. I'm using Ubuntu for AMD64 architecture and the Bug was definitively patched into my systems by the last weeks online updates. I'm using "hardy-security" and "hardy-updates" repositories only (not hardy-proposed). A friend of mine got the same error after installing and updating a fresh Hardy on his AMD64 machine, so it's reproduceable easily.
With openoffice.org-gtk installed I get the following message:
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 /usr/lib/
#5 /usr/lib/
#6 /usr/lib/
#7 /usr/lib/
#8 /usr/lib/
#9 /usr/lib/
#10 /usr/lib/
#11 /usr/lib/
#12 /usr/lib/
#13 /usr/lib/
#14 /usr/lib/
#15 /usr/lib/
#16 /usr/lib/
#17 /usr/lib/
#18 /usr/lib/
#19 /usr/lib/
Marco Scholl (traxanos) wrote : | #105 |
I have the same bug.
I start ooo directly with /usr/lib/
It only a workaround.
madblueimp (madblueimp) wrote : | #106 |
Thanks for the workaround, Marco!
It's much more convenient than my approach (using sux and another user account) and will do until this bug is fixed.
oschwa2s (oliver-schwaneberg) wrote : | #107 |
Starting /usr/lib/
A little (constructive) criticism at this point:
Updates which corrupt Ubuntu's basic packages (like Open Office) should never make it from "hardy-proposed" to "hardy-updates". I believe this bug is unnecessary and annoying because it is very obvious and it occurred on every up-to-date AMD64 Hardy so far. It even may scare new users back to Windows. It's simply not acceptable for productive machines that Office won't work for a week or so. I always believed (or hoped) that Canonical had a better quality assurance...
Stunts (f-pinamartins) wrote : | #108 |
The workaround worked for me too.
On a side note if i start /usr/lib/
Maybe what got bug is some permission inside the module(?) that launches an app directly. It apparently got blocked for the user that made the update only...
Maybe someone more experienced then me could look deeper into this?
I think the solution could be really close.
Karl (kzawoy) wrote : Re: [Bug 185311] Re: hardy, locking assertion failure, xorg/libsdl | #109 |
I agree, I have stopped using Hardy with this bug. It's taken too long to
fix. It's back to Win XT.
On Mon, Jun 23, 2008 at 6:23 AM, oschwa2s <email address hidden>
wrote:
> Starting /usr/lib/
> too, tank you!
>
> A little (constructive) criticism at this point:
> Updates which corrupt Ubuntu's basic packages (like Open Office) should
> never make it from "hardy-proposed" to "hardy-updates". I believe this bug
> is unnecessary and annoying because it is very obvious and it occurred on
> every up-to-date AMD64 Hardy so far. It even may scare new users back to
> Windows. It's simply not acceptable for productive machines that Office
> won't work for a week or so. I always believed (or hoped) that Canonical had
> a better quality assurance...
>
> --
> hardy, locking assertion failure, xorg/libsdl
> https:/
> You received this bug notification because you are a direct subscriber
> of a duplicate bug.
>
--
Karl R. Zawoy
386-344-1226
Chris Cheney (ccheney) wrote : | #110 |
oschwa2s,
Whatever is causing this bug for openoffice.org users is hard to nail down. If someone who is being affected by this bug wants to try using the intrepid version of libxcb or intrepid entirely to see if it still occurs for them that would be useful to know. I can not reproduce this problem on my machines on either i386 or amd64 hardy with hardy-updates. According to several users who have reported this bug it only occurs for them on the user who originally installed the system. If they create a test user it works fine. Removing their ~/.openoffice.org2 directory does not help and neither does rebooting (which removes all files under /tmp). So this bug isn't easily reproducible even by the users who see this bug, except under their exact same user account.
Chris
Matthew L. Dailey (matthew-l-dailey) wrote : | #111 |
I have tested Bryce's non-xcb libs on an amd64 hardy machine, and can verify that they solve the Matlab problem noted by Alexsander Demko above. I have reported this previously to MathWorks and will point them at this thread. If it's possible to roll these out for 8.04.1, that would be great.
I expect this will fix the console messages from ProE Wildfire 3 noted by jhyder79, but this runs in 32-bit compatibility mode, so I'll need to test on a "real" 32-bit machine with Bryce's libs.
This brings up the point that these libs are also in ia32-libs for amd64 machines, so the non-xcb libs should be put here, too.
Aleksander Demko (ademko) wrote : | #112 |
I've also tested Bryce's non-xcb libs and they also fix my Matlab issues completely. They also fix my Maple issues, which is reasonable as both packages share a Java foundation in their GUI. I bet this fixes bugs with many other software packages too.
I hope the non-xcb backed X11 libraries go back into Hardy, especially as it's a LTS release. Installing Hardy and finding many packages break is quite disconcerting. I'm not sure how many applications use xcb and X11 together, but surely sacrificing the stability of older applications isn't worth it just to run a few xcb+x11 mixed programs.
Sam_ (and-sam) wrote : | #113 |
Hi,
do I miss here something, there will be no fix for Hardy LTS, but for Intrepid?
What else are we supposed to do, except showing backtraces and sysdetails?
Speaking of the libx11~noxcb package above, kirill-belokurov said
>>The only thing that is not OK - now the openoffice crashes when closing it<<
is there an unspectacular alternative?
I run Ubuntu productive since edgy without reinstalling, only upgrading, that's a mistake?
Agree
https:/
oschwa2s (oliver-schwaneberg) wrote : | #114 |
Hi! I see I'm a little late. Anyways, I also installed the following packages from Bryce's site:
libx11-
libx11-
libx11-
And now everything works fine again!
I strongly agree with Aleksander Demko. I don't want to blame anybody at all, but I just wanted to state that this bug is on the simple user, who just installed the system from CD to use it for work or whatever (maybe even with the user who originally installed the system). These users are simply not interested on the systems insides, they just want to have a stable and fast system to work with. I'm a computer scientist myself and nevertheless it is pain in the ass for me to hunt such bugs during work time, because I have a lot of other things to do. So my own productiveness depends on the reliability of my Ubuntu powered system. So bugs like these can make me sad. I just want to warn that fast development on systems which were already modified from the standard installation can cause bugs which are bad for Ubuntu's image and also for the image of Linux and OpenSource in general.
Cheers,
oschwa2s
oschwa2s (oliver-schwaneberg) wrote : | #115 |
Hi again,
I just realized that compiz-fusion stopped working since I have installed the libx11~noxcb libs.
Cheers,
oschwa2s
Matthew L. Dailey (matthew-l-dailey) wrote : | #116 |
I just tested Bryce's packages on an i686 hardy install, and I can confirm that this fixed the issues with Matlab (7.5 - R2007B and 7.6 - R2008a), ProEngineer Wildfire 3.0, and Maple 11.02. All the java libxcb-xlib console messages are gone in Matlab 7.5, ProE and Maple, and the slow rendering of Matlab and Maple over tcp is gone as well.
Sam_ (and-sam) wrote : | #117 |
@oschwa2s
Confirmed so far, because when I was about to remove libx11-xcb in Synaptic, it was about to remove compiz also, so I stopped and kept the package.
Thanks to Marco Scholl I created a new starter with the mentioned command and I'm fine and will not install any noxcb. To be clear I use compiz not for fun, no fire, water etc., but for a better view and smoother working.
Peter Soetens (peter-soetens) wrote : | #118 |
For the record, the Kubuntu users need to remove the openoffice.org-kde package to work around this bug when their open office program crashes.
Chris Cheney (ccheney) wrote : | #119 |
Peter,
I have heard that just logging in as a different user will correct this problem as well without needing to remove openoffice.org-kde, openoffice.org-gtk, or openoffice.
Chris
Stunts (f-pinamartins) wrote : | #120 |
Chris,
Is there any possibility that another file outside the home folder could be causing this?
Could it be some file that get a wrong no read access on a specific user - something that could have happened during the update while the user had oo.org running for instance (I know I had an I have the bug).
I know I keep insisting on the same issue, but despite being relatively unexperienced, it makes sense to me...
You sound like you really know your way around the system... Any chance you can look this up?
Check my last post to see what tipped me off on this...
Benjamin Drung (bdrung) wrote : | #121 |
Chris,
I have tried following (tested with amd64 on Core 2 Duo):
1. back port libxcb 1.1-1.1 from intrepid -> oo.org does not work
2. back port libx11 2:1.1.4-2 from intrepid -> oo.org does not work
3. test back ported libxcb and libx11 -> oo.org does not work
4. testing intrepid (libxcb 1.1-1.1, libx11 2:1.1.4-2, openoffice.org 1:2.4.1-1ubuntu1) -> oo.org works
You can find the packages on my PPA. I can confirm that this bug only occurs for the user who originally installed the system. An other user created before the upgrade to the buggy version has no problems to start openoffice.org.
Only users with multicore systems (e.g. AMD Athlon X2, Core 2 Duo, Core 2 Quad) are affected from this bug! I found nobody using a single core.
Ian Ellis (ianmailstuff) wrote : | #122 |
Actually, I have a single core Athlon 64 system that I'm getting this bug on, but I copied the OS over from my dual-core laptop by making a tarball of it since I had made several changes since the install that I didn't want to go through again. I'm not sure if that makes a difference or not.
s1300045 (s1300045) wrote : | #123 |
I have a single core Athlon 64 system and I am getting this bug, too. My system was updated from Gutsy, but that was a while back, and I never had any problem at all until yesterday I tried to open a document.
Kees Cook (kees) wrote : | #124 |
I've encounted the locking bug now. No clue what the problem is, but downgrading all of OOo fixed it for me:
sudo apt-get install $(for i in openoffice.org openoffice.org-core openoffice.org-base openoffice.org-calc openoffice.org-draw openoffice.
Ralf Naujokat (ubuntu-ralf-naujokat) wrote : | #125 |
Hi.
Some of my observations, don't know if helpful:
I found, that i get the locking error on some (not all) files on a NFS-mounted directory.
So i copied that folder to a local disc. Now some files are ok, some other files still
have the error. When i copy and past the file (with nautilus) the error is gone on the copy.
So i deleted the old file and renamed the copy to the old filename. Now the error is there
again. This is reproducable: Copy is ok, error is back again after renaming.
So i went back to the NFS-mounted directory:
Copy / Renaming does not help here.
After playing a bit more, i found, that copying the files to a subdirectory
does help sometimes, sometimes not. In one case the copy to ./tmp had
the error, the copy to ./123 not.
One NFS-file with the locking error was special: It had a german umlaut (ö) in
its name. Copying didn't help, but replacing the ö with o did help. But don't
concentrate on that, i found an other example, where adding an ö 'fixed' the
error.
So now my view into that big crystal ball:
Is there some caching / hashing mechanism in the file locking system,
that depends on filenames and that does its work ..uhm.. sub optimal?
greetings
Ralf
Is there some
Lukas Hejtmanek (xhejtman) wrote : | #126 |
> 4. testing intrepid (libxcb 1.1-1.1, libx11 2:1.1.4-2, openoffice.org 1:2.4.1-1ubuntu1) -> oo.org works
I'm using Ubuntu/intrepid and having exactly the same version of the packages above. document editor from oo.org does not work. spreadsheet editor works.
W. Cravens (wcravens) wrote : | #127 |
Is before:
This problem also started for me as of an update June 18th, 2008.
AMD64
NVidia video module installed via NVIDIA-
Using only restricted, universe and multiverse.
I'm also -
- Intel(R) Core(TM)2 Duo CPU T7700 @ 2.40GHz
- Error occurs for the initial user added at installation
- ssh localhost "ooffice path/to/file" works for both spreadsheets and writer files
- Yesterday (20080623 8am CDT) after an update I could open spreadsheets from 'link icons' but not from command line.
- Today (No further updates installed) I can no longer open any without ssh.
Sam_ (and-sam) wrote : | #128 |
For the record, one core and OpenOffice or any other app have been open during update.
Sam_ (and-sam) wrote : | #129 |
Sorry, correct above: ... OpenOffice nor any other app have been open during update.
Age Jan Kuperus (ajk-tt) wrote : | #130 |
I am using Ubuntu 8.04 x86_64 on a HP Compaq nc6320. Since last weeks update (including proposed and backports) I got the aforementioned openoffice bug. I also used a new user as a workaround. Today I completely removed the two packages mentioned earlier, and I could run openoffice with my normal account again. Then I reinstalled those two packages:
Commit Log for Wed Jun 25 10:32:37 2008
Installed the following packages:
openoffice.
openoffice.org-gtk (1:2.4.1-1ubuntu1)
and it is still working.
oschwa2s (oliver-schwaneberg) wrote : | #131 |
Sorry, I cannot confirm that. I removed and reinstalled openoffice.
Lars Noodén (larsnooden) wrote : openoffice.org | #132 |
This is moved from bug 242950:
#1 & #3 : openoffice.org never gets past the splash screen. This occurs when running it from the Kmenu, or from the shell. That, I guess, would be when making a new file.
#3 Even when running it with an explicit filename from the shell (e.g. soffice foo.ods) the same thing happens.
#2: How would I go about identifying the version of the JRE?
from dpkg -s openoffice.
Below is the error output:
QSettings: failed to open file '/etc/qt3/
lars@macbook:~$ Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 /usr/lib/
#5 /usr/lib/
#6 /usr/lib/
#7 /usr/lib/
#8 /usr/lib/
#9 /usr/lib/
#10 /usr/lib/
#11 /usr/lib/
#12 /usr/lib/
#13 /usr/lib/
#14 /usr/lib/
#15 /usr/lib/
#16 /usr/lib/
#17 /usr/lib/
#18 /usr/lib/
#19 /usr/lib/
Onno Steenbergen (osteenbergen) wrote : | #133 |
Installed the noxcb package. No luck, I now get a sequence lost error when i run glx or fglrxinfo.
But some program like ume-launcher are running with the patch.
Does anyone know how to fix these sequence errors or is that just because there is a new version of glxinfo which uses xcb?
Chris Cheney (ccheney) wrote : | #134 |
Lars,
To find out what JRE openoffice.org is attempting to use look in this file, it will be named slightly differently depending on arch.
~/.openoffice.
Chris
Bryce Harrington (bryce) wrote : | #135 |
Thanks osteenbeergen for testing that, although your comments are ambiguous - does it function better, worse, the same as, or just differently than before?
Since we're seeing a few issues with traces involving libxcb, I've prepared a non-xcb-enabled libx11 here: http://
Would others mind testing with this noxcb libx11 and see if it makes any difference?
Abinoam Jr (abinoam) wrote : | #136 |
I'm using 8.04 amd64 with backports enabled.
openoffice from another user's account --> worked for me!
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
And, if I edit the /usr/lib/
# test for availability of the fast external splash
# if [ "$sd_binary" = "soffice.bin" -a -x "$sd_prog/
# sd_binary=
# fi
... and type "openoffice" at the terminal, openoffice opens!
But, "openoffice -writer" --> bugs.
Abinoam Jr (abinoam) wrote : | #137 |
I mean... COMMENTING the lines in /usr/lib/
# test for availability of the fast external splash
# if [ "$sd_binary" = "soffice.bin" -a -x "$sd_prog/
# sd_binary=
# fi
chewearn (chewearn) wrote : | #138 |
Confirmed working when following Abinoam Jr post, by commenting out the indicated lines in /usr/lib/
Now, the following (in terminal) work:
openoffice
ooffice
ooffice --writer
ooffice --calc
ooffice --impress
The followng does not work:
ooffice -writer
ooffice -calc
ooffice -impress
It's weird that one additional dash - can be the difference; anyone knows why?
chewearn (chewearn) wrote : | #139 |
Oops, scratch that. The double dashes don't do anything, except open writer.
Benjamin Drung (bdrung) wrote : | #140 |
- strace-f_oowriter Edit (982.7 KiB, text/html)
Bryce,
the noxcb libx11 works fine but Compiz does not work anymore. Compiz depends on libx11-xcb1. The noxcb libx11-xcb1 package contains nothing (except the normal doc files). So compiz.real complains that it cannot find libX11-xcb.so.1. I have tried to recompile compiz. But it complains about the missing xcb lib. There is no option to disable xcb. I tried to remove xcb dependency in debian/control and configure.ac and Xlib-xcb.h in include/
So we should fix the bug in openoffice.org-gtk (providing libvclplug_
PS: This bug has rank 33 (sorted by number of comments) and rank 76 (sorted by number of duplicates) in Ubuntu.
Jonathan Ernst (jonathan.ernst) wrote : | #141 |
This is getting ridiculous... Why don't we simply revert this update and try to find a solution in testing or proposed repository. This bug affects too many people who are supposed to use a production version of Ubuntu. I have at least twelve of my users that are complaining because of this bug.
Chris Cheney (ccheney) wrote : | #142 |
For the openoffice.org part of this bug I am currently building a set of test openoffice.org debs with some new patches and will be checking to see if it fixes the bug. Hopefully we will know by tomorrow and can get it into hardy-updates as well.
Chris
F. Kooman (fkooman) wrote : | #143 |
(Woops, missed that the other bug was closed and this one is the tracker of this issue now)
It seems it has something to do with locale settings. With locale set to en_US.UTF-8 it works, with locale setting to nl_NL.UTF-8 (I also tried pl_PL, but also doesn't work)
openoffice.org 1:2.4.1-1ubuntu1
Linux fkooman-laptop 2.6.24-19-generic #1 SMP Wed Jun 18 14:15:37 UTC 2008 x86_64 GNU/Linux (64 bit as well)
fkooman@
fkooman@
fkooman@
javaldx: Could not find a Java Runtime Environment!
<<< OpenOffice.org works here!! >>>
fkooman@
fkooman@
fkooman@
javaldx: Could not find a Java Runtime Environment!
fkooman@
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 /usr/lib/
#5 /usr/lib/
#6 /usr/lib/
#7 /usr/lib/
#8 /usr/lib/
#9 /usr/lib/
#10 /usr/lib/
#11 /usr/lib/
#12 /usr/lib/
#13 /usr/lib/
#14 /usr/lib/
#15 /usr/lib/
#16 /usr/lib/
#17 /usr/lib/
#18 /usr/lib/
#19 /usr/lib/
jhyder79 (jhyder79) wrote : | #144 |
Just wanted to add that Bryce's libx11 files completely fixed my problems with Pro/Engineer Wildfire 3.0. I no longer have the issues with the menus reacting slowly and everything is extremely responsive.
Sebastian Meyer (wastl) wrote : bug also depends on filename and pwd | #145 |
Wanted to comment on the fact, that ooffice opens with failure depending on the command line argument.
The following commands work fine for me, assuming I'm in my home directory
pwd # /home/sebastian
soffice --anything
touch helloworld.odt; soffice helloworld.odt
touch hellobigbigworl
touch hellobigbigbigw
NOW, by increasing the length of the filename by 1
touch hellobigbigbigw
it does not work any longer and throws the locking assertion failure!
Also
soffice /home/sebastian
does NOT work, NOR does
cd /tmp
soffice --anything
touch helloworld.odt; soffice helloworld.odt
And this is really strange... If it works or not, not only depends on the filename of the document (or the command line argument in general), but also interacts with the working directory `pwd`!
Final note: I'm using ubuntu hardy with kernel release 2.6.24-19-generic on x86_64 AMD architecture and latest updates, especially those concerning the openoffice suite version 1:2.4.1-1ubuntu1, published on 2008-06-17. (I did not change anything in the soffice starting script.)
mesatchornug (mesatchornug) wrote : | #146 |
having the same problem with openoffice
also running hardy amd64 2.6.24-19-generic
when i boot into a secondary test account that i made since OO.o went bad, everything's fine
from my primary account i can "sudo soffice" and it works, but if i "sudo soffice -write" i get the same locking assertion failure
when i boot into 2.6.22-15-generic [leftover after upgrading from gutsy] OO.o works, though my video drivers are screwy
a few minutes ago there was a new series of updates, including one for the kernel, and two others that sounded related. i applied them, rebooted, no luck. sorry i'm not terribly helpful on what these last three were, i was too busy being excited to see what looked like fixes to write down what they were.
thanks to everyone working on this.
Chris Cheney (ccheney) wrote : | #147 |
After testing it appears I have a fix for the openoffice.org problem and hopefully will be uploading a new version tomorrow to hardy-proposed. I have a few other bugs to fix at the same time which I am still working on.
Chris
Chris Cheney (ccheney) wrote : | #148 |
I marked the openoffice.org part of this bug as invalid since it is an openoffice.org specific issue and I split out the duplicate OOo related bugs into their own separate bug 236676.
Chris
Changed in openoffice.org: | |
status: | Confirmed → Invalid |
status: | New → Invalid |
Alexander Rødseth (alexanro) wrote : | #149 |
I tried downloading and running Savage again, and it no longers produces the original output/error. It now tells me that I need GLSL-support instead.
Bryce Harrington (bryce) wrote : | #150 |
I'm marking the libxcb bug won't fix for hardy because a) the only libxcb "fix" we can deploy at this point would be to disable libxcb entirely - with the side effect of disabling compiz for everyone (c.f. comment #137), and b) a viable workaround exists (c.f. comment #144). See bug #236676 for details.
Leaving it open for Intrepid, as we expect to see a new libX11/libxcb with the relevant fixes within a month or so.
Changed in libxcb: | |
status: | Confirmed → Won't Fix |
status: | New → Invalid |
Changed in openoffice.org: | |
status: | New → Invalid |
cparg (cparg) wrote : Re: [Bug 185311] Re: hardy, locking assertion failure, xorg/libsdl | #151 |
Hello,
this is really bad because it means the LTS Ubuntu Hardy is not usable
for serious applications.
Is there any summary about this problem? Anything one could pass to the
programmers of broken applications to fix it on the application side ?
(BTW, the subject refers libsdl. My application that fails does not bind
any sdl library...)
Thanks
Chris
Bryce Harrington wrote:
> I'm marking the libxcb bug won't fix for hardy because a) the only
> libxcb "fix" we can deploy at this point would be to disable libxcb
> entirely - with the side effect of disabling compiz for everyone (c.f.
> comment #137), and b) a viable workaround exists (c.f. comment #144).
> See bug #236676 for details.
>
> Leaving it open for Intrepid, as we expect to see a new libX11/libxcb
> with the relevant fixes within a month or so.
>
> ** Changed in: libxcb (Ubuntu Hardy)
> Status: Confirmed => Won't Fix
>
> ** Changed in: libxcb (Ubuntu Gutsy)
> Status: New => Invalid
>
> ** Changed in: openoffice.org (Ubuntu Gutsy)
> Status: New => Invalid
>
>
Aleksander Demko (ademko) wrote : | #152 |
Dang, is compiz the only major application that needs both X11 and xcb? Because if it is, wouldn't it make more sense to link compiz against "libX11-xcb.so" and leave "libX11.so" be the non-xcb, stock/old/
I suggest this, as I hope you guys rethink the magnatude of the wontfix status. Hardy essentially doesn't work out of the box with all Java applications and countless other closed & open, current & old applications. I think this is especially important as Hardy is supposed to be a LTS/stable release, that one can count on for larger, more long term deployments (such as one I'm trying to do, but this bug is holding up).
Thanks for your consideration.
thunter (thunter0512) wrote : | #153 |
This bug affects a large number of programs written using xlib. Why is the status "won't fix"?
Thanks
Tom
cparg (cparg) wrote : | #154 |
I totally agree. I can not run my main Application to create IC-Layouts
and Schematics...
LTS Hardy is therefor not a platform for such professional
applications.. :-( ... what a pitty...
Chris
Alex Demko wrote:
> Dang, is compiz the only major application that needs both X11 and xcb? Because if it is, wouldn't it make more sense to link compiz against "libX11-xcb.so" and leave "libX11.so" be the non-xcb, stock/old/
>
> I suggest this, as I hope you guys rethink the magnatude of the wontfix
> status. Hardy essentially doesn't work out of the box with all Java
> applications and countless other closed & open, current & old
> applications. I think this is especially important as Hardy is supposed
> to be a LTS/stable release, that one can count on for larger, more long
> term deployments (such as one I'm trying to do, but this bug is holding
> up).
>
> Thanks for your consideration.
>
>
thunter (thunter0512) wrote : | #155 |
Please note that the problem is triggered in stable applications which have been running correctly for many years on many platforms including *BSD, Solaris and Linux. The problem is clearly not in the application, but in XCB. One example of a non-fatal fault in Desktop CYBER is below:
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 /usr/lib/
#5 /usr/lib/
#6 /usr/lib/
#7 /usr/lib/
#8 /usr/lib/
#9 ./dtcyber [0x8066ec5]
#10 /lib/tls/
#11 /lib/tls/
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 ./dtcyber [0x8066ec5]
#5 /lib/tls/
#6 /lib/tls/
Martin West (martin-objectgizmos) wrote : | #156 |
I am getting the problem in eclipse ide. with jdk5 and 6, I knocked up this quick script to zap all ...
#!/bin/bash
sos=`locate libmawt.so`
for SO in $sos ; do
dir=`dirname $SO`
# make copy of original
if [ ! -f $dir/libmawt.so.old ]; then
cp $SO $dir/libmawt.so.old
fi
sed -i 's/XINERAMA/
done
Martin West (martin-objectgizmos) wrote : | #157 |
PS re WONT FIX - this must be a severity 1/2 problem since it breaks some many applications.
cparg (cparg) wrote : | #158 |
Hi folks,
I tried to install the libraries from
http://
... so far everything is working again !!! Great !
My OS is Ubuntu 8.04 (Hardy Heron) x86_64. All packages are up to date.
What does it need to verify this as a fix ?
Actually I did not have trouble with SDL libraries or Java, as many here.
A particular electronic CAD software was not running anymore in my case. Now it's working again...
Man thanks to ~bryce !!!!!
Regards,
Chris
Bryce Harrington (bryce) wrote : | #159 |
To follow up to the above comments, as I mentioned, the specific issue characterized on this bug - OpenOffice crashing - was resolved (it turned out to be a xrandr heap corruption rather than a libxcb issue).
However, regarding the libxcb locking issue, I checked with upstream about it the other week. The original report saw it via libsdl, but over time the bug report has accumulated other reports of problems with different software stacks but similar symptoms. I believe it is likely they do all share the same root cause - an issue in libxcb's locking architecture that upstream knows about and is actively working on. Unfortunately, fixing it requires a redesign of libxcb's locking architecture (which I assume means a huge code delta - which would probably be nigh impossible to get an SRU accepted for; a backport is definitely more likely, but I don't know if that would really be considered a "fix" - regardless, I'll put it on my todo list to get the backport in once it becomes available).
The other option is to remove libxcb support from libx11 entirely - the noxcb packages I included above take this approach. However, compiz has a hard dependency on the libxcb-enabled libx11, so uploading that package would cause compiz to stop working for everyone. In last week's platform team meeting we discussed the option of shipping two versions of lib11, one for compiz, one for legacy apps, however the xcb-enabled libX11.so.6 doesn't use symbol versions today, so trying to clear up symbol collisions between compiz and its libs would require a rebuild of the whole dep chain, /after/ the work to actually add symbol versions, and this whole plan would also need to be passed by upstream to verify that it's even possible to use both X11 libs from the same process (they might try to overrun each other's state via env vars or something). So that seems pretty infeasible to do for Hardy. (See https:/
Fortunately, so far each of the libxcb issues that have been reported in detail and fully investigated were able to be sorted out without resulting to such drastic measures. While the solutions were hacky workarounds, they could be implemented without risking regressions like the noxcb change would do. As a reminder, (and as confirmed by many others in comments above) we've found for Java applications you can work around the issue through one of these two approaches:
#http://
$ export LIBXCB_
or (Java apps only):
$ sudo sed -i 's/XINERAMA/
If neither of those work, or if you really, really want to get rid of xcb and don't care about compiz, you're welcome to use my noxcb libx11 package as a third way to work around it. I'll keep them at that location so they're available until we find a better solution.
All that said, if anyone can find an issue that is provably *not* fixed by any of the above workarounds, that can be traced to this libxcb locking issue, and can provide a DETAILED bug report against libxcb on it - that includes the failing application name/v...
Changed in libxcb: | |
assignee: | nobody → bryceharrington |
Lourenço Marinho (lopima) wrote : | #160 |
i was with jre 1.5, and i tried to do those things... and nothing worked for me, after i removed 1.5 and installed 1.6 everything is working fine.
anyway thanks for the advices.
Aleksander Demko (ademko) wrote : | #161 |
Bryce, thanks for the update and your good work on the issue. Keeping the community in the loop is one reason why I'm glad I switched over to Ubuntu back when I did, even if it means tripping over an odd bug or two.
I really hope Canonical decides to unroll the xcb stuff in libX11 (and perhaps release an updated Compiz that uses its own private xcb+X11). Even if you guys think you understand all the broken cases, I'm sure there will be piles of older software that will simply break (and won't be reported here), leaving a lot of frustrated users and their applications. Even if your updates expose bugs in the user client application itself, the user will still blame Ubuntu, and might not have any recourse for updating their old/unmaintaine
Having two seperate libX11 and libX11xcb libraries would allow a gradual (and distinct) migration route for applications (well, more for toolkits). Toolkits can switch to the X11+xcb one (or ideally, pure xcb) at their own, hopefully well tested, pace.
Sorry to sound like a broken record on this topic :)
knarf (launchpad-ubuntu-f) wrote : | #162 |
- A script to install libx11-noxcb outside of package management's reach Edit (1.0 KiB, text/x-sh)
A good temporary solution to these problems is to copy Bryce's noxcb version of libx11 to a separate directory which you then prepend to LD_LIBRARY_PATH before starting problematic apps. A simple if somewhat hackish way to do that goes like this:
# go to your home directory, get the package, create directory
# and change ownership to normal user for now... makes it safer
# for hamfisted types who might rm -rf /usr accidentally when given the chance...
TMPDIR=$(mktemp -d)
cd $TMPDIR
wget -P $TMPDIR http://
sudo mkdir -p /opt/LIBx11-noxcb
sudo chown $USER /opt/LIBx11-noxcb
# this produces a data.tar.gz, control.tar.gz and debian-binary file.
# The data.tar.gz contains the interesting bits, the rest is unimportant for this purpose
cd /opt/LIBx11-noxcb
ar x $TMPDIR/
# unpack data.tar.gz, move usr/lib -> /opt/LIBx11-
# remove everything else which is not needed
tar zxf data.tar.gz
mv usr/lib .
rm -rf control.tar.gz data.tar.gz debian-binary usr
#revert ownership of the directory back to root)
cd
sudo chown -R root:root /opt/LIBx11-noxcb
rm -rf $TMPDIR
# that's all folks!
(this script is attached to this posting - as always check before you run to assure yourself that there are no nasty surprises hiding in there...)
You now have a directory called /opt/LIBx11-noxcb which only contains a lib directory with libX11.so.6.2.0 (and libX11.so.6 symlinked to this file). This directory is outside of package management so it will not be messed up by any updates, nor will installing the library this way mess up your system. If you want to run an app which needs this library you just prepend it to LD_LIBRARY_PATH before running the app. Like so:
LD_LIBRARY_
If you have to do this very often you might want to create or change a startup script for said app...
Once the problems have been solved and xcb no longer gives locking problems you can just rm -rf /opt/LIBx11-
Marcos (deflagmator) wrote : | #163 |
OK, tested and work perfect.
This solves all my problems with Cinelerra CV. (and I can enable Compiz)
:-)
Andy (andy-xillean) wrote : | #164 |
Is this bug Fixed yet. It like most other high visibilty bugs in Hardy have been going on for way too long. This is pathetic. One cannot run vmware console on HARDY.. What gives? Does no one at Canonical CARE? This bug along with the DVD read bug and the pulse audio bugs have been since the release of Hardy. Does Canonical not care about these critical SHOW STOPPING BUGS. Why should people continue to use this dismal HARDY if the don't even care to make SHOW STOPPING BUGS in an LTS release a PRIORITY. This is NOT ready for prime time sorry. Dismally unusable.
Andy (andy-xillean) wrote : | #165 |
I tried the solution knarf posted
here is the command i type.
LD_LIBRARY_
I still get this
LD_LIBRARY_
Locking assertion failure. Backtrace:
#0 /usr/lib32/
#1 /usr/lib32/
#2 /usr/lib32/
#3 /usr/lib/
#4 /usr/lib/
#5 /usr/lib/
#6 /usr/lib/
#7 /usr/lib/
#8 /usr/lib/
#9 /usr/lib/
#10 /usr/lib/
#11 /usr/lib/
#12 /usr/lib/
#13 /usr/lib/
#14 /usr/lib/
#15 /usr/lib/
#16 /usr/lib/
#17 /usr/lib/
#18 /usr/lib/
#19 /usr/lib/
Locking assertion failure. Backtrace:
#0 /usr/lib32/
#1 /usr/lib32/
#2 /usr/lib32/
#3 /usr/lib32/
#4 /usr/lib/
#5 /usr/lib/
#6 /usr/lib/
#7 /usr/lib/
#8 /usr/lib/
#9 /usr/lib/
#10 /usr/lib/
#11 /usr/lib/
#12 /usr/lib/
#13 /usr/lib/
#14 /usr/lib/
#15 /usr/lib/
Bryce Harrington (bryce) wrote : | #166 |
I spoke with Bart Massey about this the other day, and unfortunately he says the re-architecture work to eliminate the locking issues is delayed and doesn't have an ETA.
However, I've word that the recent java fixes in Intrepid pretty much resolve this issue now. I'd love hearing any confirms/denies of this.
Regarding unrolling xcb from libx11, it's impossible to justify it without knowing of specific issues in specific apps. At this point I have only heard the generality that "older proprietary apps might break", which isn't enough information to make a decision on. Several apps have been mentioned by name earlier, however many (like soffice) are either Java or contain large amounts of Java, so may have just been afflicted by the jdk issue now resolved; other apps I just don't know if they're in java or C/C++.
Cinelerra is C afaik, but it's also an open source app, so in that case fixing the issue in the app is the preferred way to go.
Aleksander Demko (ademko) wrote : | #167 |
Bryce, thanks for the update.
I'm glad that some fixes are being deployed in Intrepid, but I just wish some of them would be backported to this LTS release. There doesn't seem to be to much confidence about these Java fixes, never mind not even knowing (ETA) about when the real fixes could be applied. As it stands now, I can't deploy hardy with confidence, and have to wait to Intrepid and then follow the non-LTS releases.
Older, proprietary apps *do* break, this is fact. Hardy is still relatively young, and most people don't care/can't be bothered to file bugs, so they won't show up here. I have an install of IDL and UT2004 that do not work under hardy, but fine under previous versions.
I love Ubuntu for its polish and fixing all the corner cases in typical user experiences. I hope Canonical doesn't feel they have to re-architect core systems (libx11, pulse audio) just to stay "hip and modern." "Just works" is a really strong point for Ubuntu.
Or at the very least, keep the experiments to the non-LTS releases.
Thanks again.
Bryce Harrington (bryce) wrote : | #168 |
The last several comments is doing nothing at all to help the case of those wishing to revert libx11. Still no data about *specific* apps with *specific* problems (backtraces or at least stderr msgs) are NOT being posted to this bug report. If problems are so pervasive, then I would expect it to be a simple matter to reproduce and post some factual details (screenshots, error logs, *something*). Simply ranting more loudly and being insulting doesn't help - just makes me question all the more if there actually are any facts behind the assertions. I'll give it a few more weeks but without solid evidence will close this as wontfix.
Ripping out xcb-enabled libx11 from hardy would be a HUGE amount of work that will impact a number of people, and not something I'm willing to entertain without unquestionably clear, detailed benefits that I can show to Canonical management and those people who will be affected by this and that will need to sign off on it. Many won't care so much about proprietary software to begin with, so convincing them to make changes simply to improve proprietary software is going to be hard to justify to begin with. Non-CoC comments like "remove the XCB infection!", "Does no one at Canonical CARE?" etc. just serve to harden opinions against making any further changes, and invest any further time beyond the good deal we've put in already.
Alejandro R. Mosteo (mosteo) wrote : | #169 |
I hit this bug recently using the player/stage robot simulator [1]. Although the main GUI works okay, the amcl driver debug GUI will cause player to crash with the familiar locking assertions. I think that player/stage uses XCD through the C binding.
I know the above description is not totally helpful to someone not familiar with player/stage. I could produce a small testcase configuration after the 15th September if there is interest in reproducing it. For now I just wanted to contribute an actual instance in which this issue has been problematic, and not solvable with the options offered in this thread. In my case, I had to use a Gutsy installation within VirtualBox (which I already had for another regression, this time in kdenlive) to workaround the issue.
I don't really know if the problem is in libxcd or in the use that player makes of it, however.
Bryce Harrington (bryce) wrote : | #170 |
Btw, this bug is about XCB not XCD.
Rafael Sachetto (rsachetto) wrote : | #171 |
I'm having this bug while trying to run a small pyqt application that uses threads.
python: ../../src/
cparg (cparg) wrote : | #172 |
- strace.dbg.gz Edit (87.1 KiB, application/x-gzip; name="strace.dbg.gz")
Hello Bryce,
this is (again) the transcript from the proprietary CAD Application I
can not use anymore:
ic: ../../src/
((xcb_req) - (dpy->request)) >= 0)' failed.
Aborted
Attached is the complete strace -f transcript.
Hope this helps.
Let me know if there is anything else I can do to support you.
Thanks
Christian
Bryce Harrington wrote:
> The last several comments is doing nothing at all to help the case of
> those wishing to revert libx11. Still no data about *specific* apps
> with *specific* problems (backtraces or at least stderr msgs) are NOT
> being posted to this bug report. If problems are so pervasive, then I
> would expect it to be a simple matter to reproduce and post some factual
> details (screenshots, error logs, *something*). Simply ranting more
> loudly and being insulting doesn't help - just makes me question all the
> more if there actually are any facts behind the assertions. I'll give
> it a few more weeks but without solid evidence will close this as
> wontfix.
>
> Ripping out xcb-enabled libx11 from hardy would be a HUGE amount of work
> that will impact a number of people, and not something I'm willing to
> entertain without unquestionably clear, detailed benefits that I can
> show to Canonical management and those people who will be affected by
> this and that will need to sign off on it. Many won't care so much
> about proprietary software to begin with, so convincing them to make
> changes simply to improve proprietary software is going to be hard to
> justify to begin with. Non-CoC comments like "remove the XCB
> infection!", "Does no one at Canonical CARE?" etc. just serve to harden
> opinions against making any further changes, and invest any further time
> beyond the good deal we've put in already.
>
>
Alejandro R. Mosteo (mosteo) wrote : | #173 |
@Bryce: Of course you're right, the xcd/xcb thing was an unfortunate mental glitch on my part.
thunter (thunter0512) wrote : | #174 |
- x11test.tar.gz Edit (5.9 KiB, application/x-gzip; name="x11test.tar.gz")
Dear Bryce,
Thanks for your hard work and the temporary workaround with the non-xcb libx11.
Unfortunately the problem is real and hurting and needs a proper fix for Hardy. I have come across a number of trusted old applications which fail in a similar way when running with the xcb based libx11 under Hardy.
I have created a minimal X11 program which demonstrates at least one aspect of the problem. It should help you to debug the problem and verify a solution.
I have attached the source, makefile, binary and a sample error log as a gzip-ed tar file.
To run the program from bash:
1) extract the files: tar -xzf x11test.tar.gz
2) build: make
3) run: ./x11test
4) note the error messages (Locking assertion failure ...)
Please note that this program (and the original from which this was cut&pasted from) runs fine on a large number of other OS platforms including earlier versions of Ubuntu, Redhat, FreeBSD, Mac OS X, Solaris etc.
Let me know if you need any more supporting material.
Best regards from Perth/Australia
Tom Hunter
--- On Sun, 7/9/08, Bryce Harrington <email address hidden> wrote:
The last several comments is doing nothing at all to help the case of
those wishing to revert libx11. Still no data about *specific* apps
with *specific* problems (backtraces or at least stderr msgs) are NOT
being posted to this bug report. If problems are so pervasive, then I
would expect it to be a simple matter to reproduce and post some factual
details (screenshots, error logs, *something*).
...
Win a MacBook Air or iPod touch with Yahoo!7. http://
Adam Fonseca (adam-fonseca) wrote : | #175 |
The libx11-noxcb library does fix this problem for vmware. The vmware startup scripts destroy any changes to LD_LIBRARY_PATH though, so it is not sufficient to add the LD_LIBRARY_PATH override to libx11-noxcb. You have to add the path later in the vmware startup scripts. I do it in /usr/lib/
Colin Watson (cjwatson) wrote : | #176 |
Could those people reporting problems over the last half-dozen comments or so please confirm whether the workaround Bryce posted works? Many people have not mentioned this at all.
export LIBXCB_
(Note that of course this has to be run in a terminal before starting the application in question in the same terminal.)
Alejandro R. Mosteo (mosteo) wrote : | #177 |
I tried the export LIBXCB_
As for the -noxcb packages, I got a violation fault with them.
Timo Aaltonen (tjaalton) wrote : | #178 |
the hardy package already uses sloppy locking by default, so the environment value does nothing.
Aleksander Demko (ademko) wrote : | #179 |
Colin, for me, only the non-xcb libx11 packages seemed to help.
MikeMc (mikecmcleod) wrote : | #180 |
I would like to fix vmware in /usr/lib/
Andy (andy-xillean) wrote : | #181 |
Here is what I found out that was causing VMWARE console to crash with this bug in Hardy. I had a broken icon set. The default folder icons in gnome had broken links. After fixing the links this bug disappeared. If you run into this bug while starting the VMWARE Console then switch to the default Human theme and try again. If it works then you have a broken link in your custom icon theme that is triggering this bug some how. Here is what was causing this bug for me. In /usr/share/
ls -l
gnome-fs-
instead of folder.png
so i fixed the link to
gnome-fs-
Now vmware console starts up without issues.
MikeMc (mikecmcleod) wrote : | #182 |
No this is not my problem. I switched to Human with no difference. Also, I do not have custom icons and no directory /usr/share/icons like Andy's.
Andy (andy-xillean) wrote : | #183 |
I used this to install VMware on a fresh install
http://
When I installed the custom icon set I got the bug starting vmware both on
64bit and 32bit systems. On all the systems when I switch back to human
VMware Console starts up fine. Now that I have fixed the link in my custom icon theme
the VMware Console starts up fine with the custom icons on both 32 and 64bit systems
Not sure why you don't have a /usr/share/icons folder. Are you running Hardy Mike?
thunter (thunter0512) wrote : | #184 |
Hi Colin,
The "export LIBXCB_
Let me know if you haven't received the test program and I will send you a copy.
Regards
Tom
--- On Thu, 11/9/08, Colin Watson <email address hidden> wrote:
From: Colin Watson <email address hidden>
Subject: [Bug 185311] Re: hardy, locking assertion failure, xorg/libsdl
To: <email address hidden>
Received: Thursday, 11 September, 2008, 1:56 AM
Could those people reporting
problems over the last half-dozen comments
or so please confirm whether the workaround Bryce posted works? Many
people have not mentioned this at all.
export LIBXCB_
(Note that of course this has to be run in a terminal before starting
the application in question in the same terminal.)
--
hardy, locking assertion failure, xorg/libsdl
https:/
You received this bug notification because you are a direct subscriber
of the bug.
Status in Ubuntu Mobile Edition: Confirmed
Status in X.Org X server: Confirmed
Status in “libxcb” source package in Ubuntu: Triaged
Status in “openoffice.org” source package in Ubuntu: Invalid
Status in “pygtk” source package in Ubuntu: New
Status in libxcb in Ubuntu Gutsy: Invalid
Status in openoffice.org in Ubuntu Gutsy: Invalid
Status in pygtk in Ubuntu Gutsy: New
Status in libxcb in Ubuntu
Hardy: Won't Fix
Status in openoffice.org in Ubuntu Hardy: Invalid
Status in pygtk in Ubuntu Hardy: New
Status in libxcb in Ubuntu Intrepid: Triaged
Status in openoffice.org in Ubuntu Intrepid: Invalid
Status in pygtk in Ubuntu Intrepid: New
Status in “libxcb” source package in Debian: Fix Released
Status in “libxcb” source package in Mandriva: Confirmed
Bug description:
Binary package hint: xorg
If I run Savage (a 3D non-open source game for Linux, available for download on
the internet), I get loads of these messages:
silverback.bin: ../../src/
((xcb_req) - (dpy->request)) >= 0)' failed.
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4
libs/libSDL-
#5 libs/libSDL-
#6 libs/libSDL-
#7 libs/libSDL-
#8 libs/libSDL-
#9 ./silverback.bin [0x80a95d3]
silverback.bin: ../../src/
((xcb_req) - (dpy->request)) >= 0)' failed.
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 libs/libSDL-
#5 libs/libSDL-
#6...
cparg (cparg) wrote : | #185 |
Hello,
this setting had no visible influence in my case.
(Hardy, 64 Bit)
Christian
Colin Watson wrote:
> Could those people reporting problems over the last half-dozen comments
> or so please confirm whether the workaround Bryce posted works? Many
> people have not mentioned this at all.
>
> export LIBXCB_
>
> (Note that of course this has to be run in a terminal before starting
> the application in question in the same terminal.)
>
>
MikeMc (mikecmcleod) wrote : | #186 |
Sorry to add to the confusion Yes I have a /usr/share/icons/ folder but I do not have a /usr/share/
Setting the export LIBXCB_
produces:
Unable to load image-loading module: /usr/l3232/
Unable to load image-loading module: /usr/l3232/
Unable to load image-loading module: /usr/l3232/
Unable to load image-loading module: /usr/l3232/
Unable to load image-loading module: /usr/l3232/
Unable to load image-loading module: /usr/l3232/
Unable to load image-loading module: /usr/l3232/
Unable to load image-loading module: /usr/l3232/
Unable to load image-loading module: /usr/l3232/
Unable to load image-loading module: /usr/l3232/
Unable to load image-loading module: /usr/l3232/
Unable to load image-loading module: /usr/l3232/
Unable to load image-loading module: /usr/l3232/
Unable to load image-loading module: /usr/l3232/
Unable to load image-loading module:...
Andy (andy-xillean) wrote : | #187 |
OK that l3232 is a typo. A result of you running the "sed fix" for 64bit. I did not have to run that fix on hardy 64bit fresh install. that l3232 is supposed to be l32 in the "sed fix" so it should be /usr/l32/gtk.....
DanTheta (dr-circlesquares) wrote : | #188 |
Hiya -
Just to confirm, the noxcb deb files above allowed python-pyxine to work properly on hardy, where previously it had suffered a hard lockup just after creating the display visual object and spawning the X event thread on roughly 95% of attempts. This was tested on machines with Intel and NVidia graphics cards (failing ~95%), but strangely worked almost every time on Via boards (using the openchrome driver).
I would certainly like to have this module working on a long-term support release, but I understand the situation you're in. Thank you for your continuing efforts!
Of course, I realize I'm probably one of relatively few people using pyxine in favour of gstreamer ...
Bernhard Reiter (ockham-razor) wrote : | #189 |
Is anyone else experiencing libxcb related crashes of evolution on x86_64?
See: https:/
Sebastien Bacher (seb128) wrote : | #190 |
how is this issue a pygtk one?
Changed in pygtk: | |
importance: | Undecided → Low |
status: | New → Incomplete |
Bryce Harrington (bryce) wrote : | #191 |
We had been anticipating a fix for this issue in time for intrepid from libxcb upstream, but none was forthcoming. We did not have time during intrepid to explore the option of de-enabling xcb in libx11 (it would require changes to compiz and other things that now require it).
This bug has gotten long and difficult to follow, but we finally have some hard data in comments #169, #171, #181 on the remaining issues. It will require some additional analysis to determine if these bugs can be individually addressed or would only be fixed by a libxcb revert. I'll leave this bug open (but the intrepid task closed) in case of the latter, but if the former looks realistic we should break those bugs out into separate bugs so we can close this one.
Changed in libxcb: | |
status: | Triaged → Won't Fix |
Changed in pygtk: | |
status: | Incomplete → Invalid |
Bryce Harrington (bryce) wrote : | #192 |
Closing out the pygtk tasks; there was never evidence presented that it has any relation to this bug. If you feel there's an xcb issue in pygtk, please file a new bug.
Changed in pygtk: | |
status: | New → Invalid |
status: | New → Invalid |
charly4711 (karl-h-beckers) wrote : | #193 |
Bryce,
as far as fixing the applications are concerned, I don't quite follow why applications that used to work need to be fixed because libxcb apparently one way or another changes an API contract that it has no business of changing.
That being said, if somebody claims its the application's fault, I'd really appreciate some information about what I need to fix around somthing that wasn't even apparently broken with Xlib.
I can attach 3 backtraces of various ways xvidcap deadlocks with libxcb (variant 3 resulting in a complete X11 lockup, so I need to ssh into the box and kill xvidcap). The only thing I can see is that in one thread I'm trying to do XShmGetImage on a locked X display while in the other thread GTK does its stuff. Have I been doing this completely wrong for years without noticing? And I've just recorded more than an hour of screencasts without a single hang with your non-xcb libraries.
This issue keeps me from upgrading to intrepid because first I'll have to find a way of compiling the required X11 libs without xcb.
charly4711 (karl-h-beckers) wrote : | #194 |
charly4711 (karl-h-beckers) wrote : | #195 |
Aleksander Demko (ademko) wrote : | #196 |
Ok, it seems the Intrepid upgrade didn't do anything to fix this issue for me.
If I ssh from an Intrepid (32-bit) desktop to an Intrepid server (64-bit) and run Matlab (which utilizes Java), it still runs "slow" (screen drawing is unnaturally sluggish). If I LD_LIBRARY_PATH in the non-xcb libX11 libraries on the server, it works fine.
Am I missing the fixes that where applied?
charly4711 (karl-h-beckers) wrote : | #197 |
Alex,
can you make an intrepid non-xcb libX11 available somewhere?
Aleksander Demko (ademko) wrote : | #198 |
charly4711, I'm just using libX11 I have from a Fedora 3/64-bit install I have on another machine (a machine I'm trying to replace with an Ubuntu one). I've also had success with Bryce's non-x11 .deb files (see my first posts on this bug, and his posts after that).
Aleksander Demko (ademko) wrote : | #199 |
The Marble Blast game demo also exhibits this bug. I've only checked this under hardy, however.
The demo can be downloaded from http://
Backtrace:
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 ./lib/libSDL-
#4 ./lib/libSDL-
#5 ./lib/libSDL-
#6 ./lib/libSDL-
#7 ./lib/libSDL-
#8 ./lib/libSDL-
#9 ./marbleblastgo
#10 ./marbleblastgo
Francesco "Zeirus" Sorrentino (zeirus) wrote : | #200 |
I want to change the latest edition of libx11-6 with the "non xcb" version (I've the problem of jhyder79 in Pro/Engineer with very slowy slid-up windows). I've downloaded libx11-
Francesco "Zeirus" Sorrentino (zeirus) wrote : | #201 |
My actual installed version of libx11-6 is: 2.1.1.5-2ubuntu1.1
In freedesktop.org Bugzilla #14211, Bryce Harrington (bryce) wrote : | #202 |
Ping? Any new status on this?
Bryce Harrington (bryce) wrote : | #203 |
I've pinged upstream. Has anyone had a chance to test if this still occurs on Jaunty yet?
Francesco "Zeirus" Sorrentino (zeirus) wrote : | #204 |
Hi Bryce,
ok. I've upgraded to jaunty now (from 8.10 with update-manager -d).
Now in ProE the refresh of windows is more fast that in previous version. There's only a "slight hiccups" initially when i open menu or parameter window, but for the rest it's perfect! Now it's possible to working again with ProE in Ubuntu! :D
In the terminal now i've only this message:
get fences failed: -1
param: 6, val: 0
but i don't know if it's relative to libxcb or to another library.
Bryce Harrington (bryce) wrote : | #205 |
Thanks, sounds like it's resolved and can be closed for karmic.
Changed in libxcb (Ubuntu): | |
status: | Triaged → Fix Released |
In freedesktop.org Bugzilla #14211, Jamey-minilop (jamey-minilop) wrote : | #206 |
The socket handoff patches have been shipping for a while now. This bug shouldn't be relevant any more. Sorry for the lack of feedback in the meantime, though.
Changed in xorg-server: | |
status: | Confirmed → Fix Released |
Jake Cobb (error404) wrote : | #207 |
I just started having this problem in hardy (manifested by Java) after an update today (2009-07-14). The sed/XINERAMA workaround suggested by others earlier worked for now.
Here are the packages I upgraded, it began immediately afterward:
2009-07-14 09:12:52 upgrade libdbus-1-3 1.1.20-1ubuntu3.2 1.1.20-1ubuntu3.3
2009-07-14 09:12:52 upgrade dbus 1.1.20-1ubuntu3.2 1.1.20-1ubuntu3.3
2009-07-14 09:12:52 upgrade dbus-x11 1.1.20-1ubuntu3.2 1.1.20-1ubuntu3.3
2009-07-14 09:12:52 upgrade libtiff4 3.8.2-7ubuntu3.2 3.8.2-7ubuntu3.4
2009-07-14 09:12:52 upgrade libtiff-tools 3.8.2-7ubuntu3.2 3.8.2-7ubuntu3.4
Changed in xorg-server: | |
importance: | Unknown → Medium |
Changed in xorg-server: | |
importance: | Medium → Unknown |
Changed in xorg-server: | |
importance: | Unknown → Medium |
Binary package hint: xorg
If I run Savage (a 3D non-open source game for Linux, available for download on the internet), I get loads of these messages:
silverback.bin: ../../src/ xcb_lock. c:77: _XGetXCBBuffer: Assertion `((int) ((xcb_req) - (dpy->request)) >= 0)' failed. libxcb- xlib.so. 0 [0xb6def767] libxcb- xlib.so. 0(xcb_xlib_ lock+0x2e) [0xb6def81e] libX11. so.6 [0xb76ca5f8] libX11. so.6(XUnmapWind ow+0x25) [0xb76c0fb5] 1.2.so. 0 [0xb7e76f01] 1.2.so. 0 [0xb7e7840e] 1.2.so. 0(SDL_VideoQuit +0x52) [0xb7e6e99a] 1.2.so. 0(SDL_QuitSubSy stem+0x8c) [0xb7e50298] 1.2.so. 0(SDL_Quit+ 0x20) [0xb7e502ec] xcb_lock. c:77: _XGetXCBBuffer: Assertion `((int) ((xcb_req) - (dpy->request)) >= 0)' failed. libxcb- xlib.so. 0 [0xb6def767] libxcb- xlib.so. 0(xcb_xlib_ lock+0x2e) [0xb6def81e] libX11. so.6 [0xb76ca5f8] libX11. so.6(XUnmapWind ow+0x25) [0xb76c0fb5] 1.2.so. 0 [0xb7e76f01] 1.2.so. 0 [0xb7e7840e] 1.2.so. 0(SDL_VideoQuit +0x52) [0xb7e6e99a] 1.2.so. 0(SDL_QuitSubSy stem+0x8c) [0xb7e50298] 1.2.so. 0(SDL_Quit+ 0x20) [0xb7e502ec] xcb_lock. c:77: _XGetXCBBuffer: Assertion `((int) ((xcb_req) - (dpy->request)) >= 0)' failed.
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 libs/libSDL-
#5 libs/libSDL-
#6 libs/libSDL-
#7 libs/libSDL-
#8 libs/libSDL-
#9 ./silverback.bin [0x80a95d3]
silverback.bin: ../../src/
Locking assertion failure. Backtrace:
#0 /usr/lib/
#1 /usr/lib/
#2 /usr/lib/
#3 /usr/lib/
#4 libs/libSDL-
#5 libs/libSDL-
#6 libs/libSDL-
#7 libs/libSDL-
#8 libs/libSDL-
#9 ./silverback.bin [0x80a95d3]
silverback.bin: ../../src/
Locking assertion failure. Backtrace:
Segmentation fault (core dumped)
I searched the web for this error, and it seems like running another 3d game, neverwinter nights, causes the same problem on Hardy. ubuntuforums. org/archive/ index.php/ t-666726. html
Here's a link to that:
http://