I'm getting several reports of users hitting this VM bug: http://bugs.sun.com/view_bug.do?bug_id=6539705
Unfortunately the recommended workaround - explicitly calling loadLibrary("awt"), or Toolkit.getDefaultToolkit() - doesn't work as someone has already noted. Would one of you guys be able to put the jni workaround (in the last comment) directly into LWJGL?
Thanks.
hmm, afaik we dont need this because we're explicitelt using the /delayload option when linking
Well that's very odd then, because I'm definitely getting reports which look *exactly* like that bug. :(
can you possibly test on older versions ? - thinking that the CI may be interfering?
Quote from: Matzon on December 06, 2010, 11:23:02
can you possibly test on older versions ? - thinking that the CI may be interfering?
I can do a special build for one of my affected users, but obviously I can't ask them to try loads of versions. Have you got a specific lwjgl version you'd like me to try?
Also, what's the CI?
CI = Hudson
I was just thinking about trying an old verson 2.2 for instance?
Quote[12:24:35] <Diablo-D3> that thread that just popped up
[12:24:50] <Diablo-D3> isnt that the same one that happened on my app, and all I had to do was explicitly set DISPLAY=:0 to fix it?
[12:24:57] <Diablo-D3> (happens on headless opencl clients)
I may be being stupid, but what does DISPLAY:=0 mean in this context?
http://lwjgl.org/forum/index.php/topic,3614.0.html