Plz help noob with demo crashes

Can you try using lwjgl .92 jars and dlls and comparing to .93? Seems like we suddenly got a rash of these complaints after the .93 release.

I just did a pull today from CVS and when running the water simulation demo it blue screened by machine and then rebooted! From the time I pressed ESC to the time the computer showed the BIOS screen was about 5 seconds. :frowning:



Dr. A>

Hmm, well the water demo is not an officially sanctioned demo but I don’t see why it would cause a blue screen. Have any other demos had odd behavior from cvs?

Can you try using lwjgl .92 jars and dlls and comparing to .93? Seems like we suddenly got a rash of these complaints after the .93 release.

I downloaded .92 and did a full re-build, but got the same result [although I patched around the SetLocation call in LWGLDisplaySystem to go fullscreen I made no other mods]. Same TID, although NID [not surprisingly] different [0x7d8].
As its a HotSpot error, I cleaned out my system as I had 3 JREs installed [1.4.5, 1.4.6 and 1.5], J2EE 1.4, and a SDK [1.4.6] - phew! - just in case the system was clashing somewhere even though my setup was pointing everything to the SDK. I then reverted back to SDK 1.4.1 being the one to least likely to have a HotSpot error since the JIT compilation was minimal in this release [although my recollection here could well be flawed].

HOWEVER - USING 1.4.1 AND LWJGL0.93 THE HOTSPOT PROBLEM HAS NOW GONE AWAY, hurrah!

I just did a pull today from CVS and when running the water simulation demo it blue screened by machine and then rebooted! From the time I pressed ESC to the time the computer showed the BIOS screen was about 5 seconds.


Also, regarding the other post on TestWaterSurface, my CVS checkout was run on Sunday 7th and with J2SE1.4.1_07 and LWJGL0.93 it runs fine - no BSOD. I will try a CVS pull and a mix of J2SEs to see if I can get a BSOD and post to advise outcome.

Matt

Wheew! Thanks for going through this work. Let us know the outcome.

I will try a CVS pull and a mix of J2SEs to see if I can get a BSOD and post to advise outcome.


Tried the water simulation demo [amongst others] using lwjgl .93 and java 1.4.1, 1.4.2_6 and 1.5. No problem on any item - although 1.5 compilation now complains about unsafe type conversion in jme and in TestWidgetAnimatedImage there was no runtime evidence of a problem.

A CVS check showed my system is up-to-date. Hope this is of use to someone, somehow....

Wonder what the problem was… probably a mix of code, dlls and/or jdks/jvms. Glad to see your test results came back positive though.

I tried doing a pull from CVS on another computer.



I was able to run the water simulation, but it didn’t do anything on this computer either. It just shows a small sliver of blue or so, but nothing larger than say 20 pixels high. It did not crash on exit though, which was nice!



The other computer was using an ATI card in a laptop. This one is a GeForce card.



Any ideas on why the JWS versions on the site don’t work? I can run them fine locally, but not via JWS.



Dr. A>

ah, well the water demo starts you looking at a quad mesh from the side, making it almost invisible… so you need to use AWSD controls and the mouse to orient yourself above that mesh. Then hit a key (Q I think) to simulate a drop hitting the mesh.



As for the JWS demos not working, the only thing I can think of is that we added a small fix since .7 that would allow certain cards to function better. (see the thread http://mojomonkeycoding.com/jmeforum/viewtopic.php?t=927)

Tried the water simulation demo [amongst others] using lwjgl .93 and java 1.4.1, 1.4.2_6 and 1.5. No problem on any item


Err, spoke too soon ://

After tidying up the system and reverting back to 1.4.2_6, the Hot Spot error has returned. Removing 1.4.2_6 and re-installing 1.5 and, hey presto, the Hot Spot error goes.

Conclusion? There's a semi-repeatable discrepancy using 1.4.2_6, but I can't nail it down. If anyone else has a similar problem then (preferably) upgrade to a newer JVM, or otherwise revert to an earlier release.

I had the same Error message.



I’m a newbye of Java and JME but I’ve made several tests.



Here wath I found.



During several experiments in past months, I’ve started to think that J2SE 1.4.2_05 and 1.4.2_06 are unstable versions. Both have given me many problems (not only running JME) on diferent systems, Win 98 SE, XP PRO and XP PRO SP2. But yesterday I’ve become absolutely sure that the 1.4.2_06 version is bugged, when I got the same Error you got. After desperate attempts to run various version mixs of JME and LWJGL, none of them working, I’ve tryed downgrading to version 1.4.2_04 and everything works well.



I’ve never tried to install J2SE 1.5 but I’m almost sure that 1.4.2_04 shuld be the right.



Consider that the system I’ve used are those:


  • for Win 98 SE, a Pentium III 600 MHz with an NVIDIA RIVA TNT 2 with 16 MB VRAM


  • for both XP PRO and XP PRO SP2, a Pentium 4 2.8 GHz with an ATI RADEON 9200 SE with 128 MB VRAM.



    Both systems OpenGL drivers are version 1.3 or above.



    Bye, bye



    Ender

we have had lots of problems with the 1.4.2_06 version in our own java-opengl engine…I’ll get back to you if I find out how we got past it…

http://www.mojomonkeycoding.com/jmeforum/viewtopic.php?t=1195



I posted about this some days ago. Please refer to the above post for more information.



DP

what values are selected in the dropdowns when you click OK? The only place I could see for an NPE would be if one of the dropdowns was returning NULL as a selected value somehow.

if the drop downs are empty it was unable to find a valid setting for an OpenGL context… have you tried the LWJGL demos? www.lwjgl.org



Are you able to run any OpenGL (non-java) apps?

we should probably throw up an error dialog if we can’t populate those boxes.

Thanks for your help. I was unclear about how this system works. :// I

have developed X-Windows programs and they can run on any display

which gives them permission. Apparently, this kind of works with jME.

I get the setup screen, but it does not populate with setup data because it

can not access the hardware.



OnceI ran the jME demo (compiled on the Linux system) on the Linux system, it worked. Once I ran the jME demo on the Windows computer after setting up jME on the Windows computer it worked. I’m not sure if you even want to address this issue, given the limited number of Unix geeks likely to make this mistake, but you might want to at least document that you can not run on a remote display. Although this is apparently obvious to everyone but me.



Again, thanks.