Added a (hacky) fix for a Java Swing/AWT + GLFW interaction issue… - #968
Conversation
|
@empirephoenix any thought ? |
|
I had the BadRenderer issue a lot of times in smaller Tests, but never in my full Application, this would explain this. While this is extremly ugly, if it makes stuff work 99% of the time it is still a great improvement. So from my point of view this might be ok. Might another approch be, to drain the XWindow error queue on JME-LWJGL3 init? |
|
@pspeed42 |
|
Thanks for investigating this |
|
@empirephoenix thanks for merging it. Regarding
I took a look at it but was unable to understand what is going on there. Sorry about that :( I just cloned this PR from what @Renanse did here : Renanse/Ardor3D@3984929 |
|
Is this fix suitable for JME v3.2.2 ? |
|
Don't see any strong reason why it should not be :) |
) * Added a (hacky) fix for a Java Swing/AWT + GLFW interaction issue… … on Linux. * Added a proper comment * Added an extra call to System.gc() as recommendation from lwjgl’s @Spasi "Run System.gc(), at least two times. Not sure how AWT handles native resources, maybe there’s a finalizable/PhantomRef-ed resource that is not released immediately, but a subsequent GC triggers a (too late) free. "
… on Linux.
Read more on https://hub.jmonkeyengine.org/t/experimenting-lwjgl3/37275