openjdk problem

ajd · 2009-03-01 03:12:26

The latest OpenJDK6 (to 1.4.1-1) update has made all (i.e. the two I tested) java applets viewed in firefox to just show blank grey screens and have no content. A look at top reveals that "java" is using a lot of CPU (99% +) but nothing is happening in the window.

I've searched google to no avail...has anyone else seen this problem?

virusso80 · 2009-03-01 19:09:39

problem already noticed!!
http://bbs.archlinux.org/viewtopic.php?id=66460

It's a very annoying problem...let'hope it's gonna be fixed as soon as possible!

Runiq · 2009-03-01 20:36:01

This is a well-known toolkit problem and known to Java developers upstream. It happens with window managers which are not reparenting (dwm, compiz, etc.). If you have a 32-bit system, you can do

$ export AWT_TOOLKIT=MToolkit <run program>

to switch back to the old M Toolkit instead of using the newer X Toolkit. Unfortunately, using the M Toolkit on x86_64 usually gives you a segfault.

The M Toolkit is actually deprecated and won't be supported for much longer AFAIK.

AndyRTR · 2009-03-01 20:54:38

any upstream bugreport we/I can follow?

Runiq · 2009-03-01 23:36:39

AndyRTR wrote:

any upstream bugreport we/I can follow?

http://bugs.sun.com/bugdatabase/view_bu … id=6429775
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=504524

It's supposed to be closed as far as I see, though the grey blobs still exist here (dwm+jabref). The second link doesn't mention a solution either.

ajd · 2009-03-03 04:24:23

Thanks for the help guys!

imrehg · 2009-03-04 08:05:24

Is there a way to downgrade the package? The previous version of the openjdk6 package was working just fine for me (in fact much better/faster than the sun java package), and it would be great to have that back....

As far as I know, abs can be used only to get the current version of the package, and not sure if there's any repo with the old ones....

Yagi858 · 2009-03-04 13:57:18

imrehg wrote:

Is there a way to downgrade the package? The previous version of the openjdk6 package was working just fine for me (in fact much better/faster than the sun java package), and it would be great to have that back....
As far as I know, abs can be used only to get the current version of the package, and not sure if there's any repo with the old ones....

Same problem same request ... please help

Runiq · 2009-03-04 17:34:33

Is there a way to downgrade the package? The previous version of the openjdk6 package was working just fine for me (in fact much better/faster than the sun java package), and it would be great to have that back....

Of course there is. If you didn't do pacman -Sc since the update, the package is still in the cache and can be installed with pacman -U. If you did pacman -Sc since then, I guess you'll have to recompile after lowering pkgver in the openjdk6 PKGBUILD (which effectively downloads an older version of the package from the official sources and installs it, using the method provided in the PKGBUILD's build{} function).

EDIT: ABS itself really isn't that complicated, you might want to look into it.

Last edited by Runiq (2009-03-04 17:37:52)

imrehg · 2009-03-04 18:28:33

Too bad that I get rid of the cache after every update (8Gb hard drive is not much to keep those things around).

I know how to do package recompiling from PKGBUILD that I get from ABS. Do that all the time, as I did plenty of my own packages, or fixing up some broken ones from AUR.

My only concern was that the difference between the different versions usually more than a simple version change... So, I bet there is a version control system (if there's one for the "community" repo, surely makes even more sense to have one for the "higher ranking" repos....), just cannot have access to it. ABS is good if you want to modify something current. It feels pretty useless if you want to downgrade....

Please correct me if there's an accessible core/extra... version control site. That would be a goldmine (as this is not the first time an official package broke things that were working before).

shining · 2009-03-04 18:52:49

imrehg wrote:

Too bad that I get rid of the cache after every update (8Gb hard drive is not much to keep those things around).

You can use external drive, web storage, other box, etc.. Plenty of options

My only concern was that the difference between the different versions usually more than a simple version change... So, I bet there is a version control system (if there's one for the "community" repo, surely makes even more sense to have one for the "higher ranking" repos....), just cannot have access to it. ABS is good if you want to modify something current. It feels pretty useless if you want to downgrade....
Please correct me if there's an accessible core/extra... version control site. That would be a goldmine (as this is not the first time an official package broke things that were working before).

web interface is here : http://repos.archlinux.org/viewvc.cgi?view=roots
community : it was and still is a cvs repo
core/extra : cvs repo contains only the old pkgbuilds
packages : new svn repo contains the current core/extra pkgbuilds
unstable : dead

Instructions to get core/extra pkgbuilds from svn : http://wiki.archlinux.org/index.php/Get … S_From_SVN
I am sure you can also get old core/extra and community by using cvs directly.

AndyRTR · 2009-03-04 20:17:42

Runiq wrote:

AndyRTR wrote:
any upstream bugreport we/I can follow?
http://bugs.sun.com/bugdatabase/view_bu … id=6429775
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=504524
It's supposed to be closed as far as I see, though the grey blobs still exist here (dwm+jabref). The second link doesn't mention a solution either.

I doubt these are the right reports.

I've seen only reports about high load with the browser plugin. I tested OpenOffice wizards and help and both work well. Also we have i686 affected too.

Please all check/report the plugin with "ICEDTEAPLUGIN_DEBUG=true firefox" and look if it's the same as https://bugzilla.redhat.com/show_bug.cgi?id=473565

Check if it's already reported there: http://icedtea.classpath.org/bugzilla/b … alue0-0-0=

Last edited by AndyRTR (2009-03-04 20:26:32)

imrehg · 2009-03-05 00:45:49

Shining: I'm not going to carry an external drive with my Eee just to keep backups of packages. And even if I would, I think there is some cleverly named "law" about your data will always fill the available space...

Cheers for the addresses! I only knew the comunity cvs before... Now let's check it out...

imrehg · 2009-03-05 02:29:33

I checked the "ICEDTEAPLUGIN_DEBUG=true firefox" thing with the site http://www.time.gov/timezone.cgi?Pacific/d/-8/java and it seems that it's a different problem from the bug report...

ICEDTEA PLUGIN: Factory::IcedTeaPluginFactory
CONSTRUCTING FACTORY
ICEDTEAPLUGIN_DEBUG = true
ICEDTEA PLUGIN: Factory::IcedTeaPluginFactory return
ICEDTEA PLUGIN: Factory::Initialize
ICEDTEA PLUGIN: Factory::Initialize: using /usr/lib/jvm/java-1.6.0-openjdk/jre/lib/i386/../../bin/java
ICEDTEA PLUGIN: thread manager
ICEDTEA PLUGIN: current thread
Initializing JVM...
ICEDTEA PLUGIN: get component manager
ICEDTEA PLUGIN: liveconnect
ICEDTEA PLUGIN: thread manager
ICEDTEA PLUGIN: Instance::StartAppletviewer
ICEDTEA PLUGIN: get component manager
ICEDTEA PLUGIN: create local file
ICEDTEA PLUGIN: init with path
ICEDTEA PLUGIN: create process
ICEDTEA PLUGIN: init process
ICEDTEA PLUGIN: clearing old input fifo (if any): /home/greg/.icedteaplugin/icedtea-appletviewer-to-plugin
ICEDTEA PLUGIN: creating input fifo: /home/greg/.icedteaplugin/icedtea-appletviewer-to-plugin
ICEDTEA PLUGIN: created input fifo: /home/greg/.icedteaplugin/icedtea-appletviewer-to-plugin
ICEDTEA PLUGIN: got confirmation that appletviewer is running
ICEDTEA PLUGIN: clearing old output fifo (if any): /home/greg/.icedteaplugin/icedtea-plugin-to-appletviewer
ICEDTEA PLUGIN: creating output fifo: /home/greg/.icedteaplugin/icedtea-plugin-to-appletviewer
ICEDTEA PLUGIN: created output fifo: /home/greg/.icedteaplugin/icedtea-plugin-to-appletviewer
ICEDTEA PLUGIN: run process
Listening for transport dt_socket at address: 8787
ICEDTEA PLUGIN: Instance::StartAppletviewer return
ICEDTEA PLUGIN: started appletviewer
ICEDTEA PLUGIN: Factory::Initialize return
ICEDTEA PLUGIN: Factory::CreateSecureEnv
ICEDTEA PLUGIN: JNIEnv::IcedTeaJNIEnv
ICEDTEA PLUGIN: JNIEnv::IcedTeaJNIEnv return
CREATESECUREENV
ICEDTEA PLUGIN: Factory::CreateSecureEnv return
ICEDTEA PLUGIN: Factory::CreateInstance
ICEDTEA PLUGIN: Instance::IcedTeaPluginInstance
ICEDTEA PLUGIN: Instance::IcedTeaPluginInstance return
ICEDTEA PLUGIN: Factory::CreateInstance return
ICEDTEA PLUGIN: Instance::Initialize
TAG FROM BROWSER = instance 1 tag http://www.time.gov/timezone.cgi?Pacific/d/-8/java <applet code="utcnist4.class" codebase="/" ignore="" width="350" height="100">
   <param name="MINUTESOFFSET" value="-480">
   <param name="BACKGROUNDCOLOR" value="0x00695E">
   <param name="FOREGROUNDCOLOR" value="0xFFFFEE">
  </applet></embed>
ICEDTEA PLUGIN: Instance::SendMessageToAppletViewer
Writing to JVM: instance 1 tag http://www.time.gov/timezone.cgi?Pacific/d/-8/java <applet code="utcnist4.class" codebase="/" ignore="" width="350" height="100">
   <param name="MINUTESOFFSET" value="-480">
   <param name="BACKGROUNDCOLOR" value="0x00695E">
   <param name="FOREGROUNDCOLOR" value="0xFFFFEE">
  </applet></embed>
Wrote 328 bytes to pipe
ICEDTEA PLUGIN: Instance::SendMessageToAppletViewer return
SETTING PEER!!!: 0xa00a3c8
DONE SETTING PEER!!!: 0xa00a3c8
ICEDTEA PLUGIN: Instance::Initialize return
NOT IMPLEMENTED: virtual nsresult IcedTeaPluginInstance::Start()
ICEDTEA PLUGIN: Instance::GetValue
ICEDTEA PLUGIN: Instance::GetValue return
GETTING PEER!!!: 0xa00a3c8
DONE GETTING PEER!!!: 0xa00a3c8, 0xa00a3c8
ICEDTEA PLUGIN: Factory::GetValue
ICEDTEA PLUGIN: Factory::GetValue return
ICEDTEA PLUGIN: Instance::GetValue
ICEDTEA PLUGIN: Instance::GetValue return
ICEDTEA PLUGIN: Instance::SetWindow
ICEDTEA PLUGIN: Instance::SetWindow: setting window.
ICEDTEA PLUGIN: Instance::SendMessageToAppletViewer
Writing to JVM: instance 1 handle 31463479
Wrote 27 bytes to pipe
ICEDTEA PLUGIN: Instance::SendMessageToAppletViewer return
ICEDTEA PLUGIN: Instance::SetWindow return
ICEDTEA PLUGIN: Instance::SetWindow
IcedTeaPluginInstance::SetWindow: Instance 0x9e68ba0 waiting for initialization...
Initialization for instance 1 has timed out. Marking it void
ICEDTEA PLUGIN: Instance::SetWindow return

I tried other site as well, like http://www.java.com/en/download/help/testvm.xml, and it seems to end in time-out all the time....

imrehg · 2009-03-05 11:10:30

I tried to rebuild the previous, 1.4-2 version of the openjdk6 package (r27021 of the svn-packages repo), but in a later stage of the compilation it fails, complaining that libjvm does not exist....
I have 1.4.1-1 installed at the moment, and it has libjvm in the appropriate directories...

Then I tried to recompile the 1.4.1-1 version, and that fails the same way.

Do I need to have the sun jre to compile openjdk? or is there any other issue, that I might have messed up? ....

AndyRTR · 2009-03-05 11:32:41

It should be buildabel with gcj, ecplipse-ecj or any former openjdk6 pkg. I'll have a look at it. But I doubt a simple recompile will fix it.

Both architecture are affected?

Yagi858 · 2009-03-05 17:51:52

AndyRTR wrote:

It should be buildabel with gcj, ecplipse-ecj or any former openjdk6 pkg. I'll have a look at it. But I doubt a simple recompile will fix it.
Both architecture are affected?

I confirm for x86_64. Same problem as http://bugs.archlinux.org/task/13667 moreover when I try to lunch applets (trusted and not) by firefox cpu grows to 100%

Runiq · 2009-03-05 20:50:04

AndyRTR wrote:

Runiq wrote:
AndyRTR wrote:
any upstream bugreport we/I can follow?
http://bugs.sun.com/bugdatabase/view_bu … id=6429775
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=504524
It's supposed to be closed as far as I see, though the grey blobs still exist here (dwm+jabref). The second link doesn't mention a solution either.
I doubt these are the right reports.

True. I overlooked the CPU load problem, sorry.

imrehg · 2009-03-05 22:49:18

Tried to build openjdk6 (i686) with gcj, same oops: complain about missing libjvm.

imrehg · 2009-03-06 16:36:01

duh, user error, openjdk compiles, I just silently run out of space (I certainly wouldn't have thought that 2.5Gb is not enough....)

this, however, does not change the original problem...

Gen2ly · 2009-03-12 22:30:03

OpenJDK has been updated to openjdk6-1.5-0.1 and fixes this for me.

Arch Linux

#1 2009-03-01 03:12:26

openjdk problem

#2 2009-03-01 19:09:39

Re: openjdk problem

#3 2009-03-01 20:36:01

Re: openjdk problem

#4 2009-03-01 20:54:38

Re: openjdk problem

#5 2009-03-01 23:36:39

Re: openjdk problem

#6 2009-03-03 04:24:23

Re: openjdk problem

#7 2009-03-04 08:05:24

Re: openjdk problem

#8 2009-03-04 13:57:18

Re: openjdk problem

#9 2009-03-04 17:34:33

Re: openjdk problem

#10 2009-03-04 18:28:33

Re: openjdk problem

#11 2009-03-04 18:52:49

Re: openjdk problem

#12 2009-03-04 20:17:42

Re: openjdk problem

#13 2009-03-05 00:45:49

Re: openjdk problem

#14 2009-03-05 02:29:33

Re: openjdk problem

#15 2009-03-05 11:10:30

Re: openjdk problem

#16 2009-03-05 11:32:41

Re: openjdk problem

#17 2009-03-05 17:51:52

Re: openjdk problem

#18 2009-03-05 20:50:04

Re: openjdk problem

#19 2009-03-05 22:49:18

Re: openjdk problem

#20 2009-03-06 16:36:01

Re: openjdk problem

#21 2009-03-12 22:30:03

Re: openjdk problem

Board footer