Is glDrawPixels (Drawing pixels to the screen) pretty much the defacto way of displaying images in Blender’s viewport? If so, I appear to be getting horrible results (Black and white/no colors except tiny rainbow pixels all over the place) with the code below. Any chance you’d know how to do this better? (For getting Arnold’s IPR working within Blender’s “rendered” viewport)
Considering Blender 2.8 to be moving towards OpenGL core though, it might make more sense to start developing this IPR for that though. As this will most likely be removed?
If I were to ever build an engine for Blender again, the time would be far better spent (in my extremely biased opinion) is to stick to something like USD’s Hydra - separating rendering logic agnostic of the underlying scene graph - and opening up more opportunity to do more with it.
In this case, a cool idea would be to create a “Hydra Delegate” for eevee.