I’ve got a bit further in my tests, and got to check almost every topic related to OpenGL drawing in the forum, but i can’t find a final answer, so i’ve few more questions :
-
Looking at the OpenGLContext source code, it seems actually impossible to do anything with a context without attaching it to a component ? My question is : is it difficult or just impossible ?
-
When using the technique above (using only one context and dispatch the openGLRender in my other classes), when i try to use the image.getPixelAt() on a image created with OpenGLImageType, this mess up the whole rendering : all my components are not rendered anymore, and i can’t draw anything after this call. If i create another Image (non-GL) and draw the GL image into this one, then i can access the getPixelAt function on that later image and everything works fine. But as i have to do that at each renderGL() call, when dealing with 1920x1080 textures, everything gets reaaally slow.
-
When profiling from VS the CPU activity, there is a strong CPU activity on the deactivateCurrentContext() function (almost 100% of one my CPUs), even when i don’t copy the GL image to a CPU image.
Any suggestion is welcome !
Thank you