As a JUCE user, I really want to take advantage of native OpenGL and its performance.
To do that, I need the flexibility to create a custom rendering workflow for my needs. (eg: Complete control over what/how/when things happen).
JUCE’s frame rendering mechanisms aren’t helping me do strict OpenGL rendering - OpenGL is difficult enough as it is. The present implementation in JUCE has 3 disadvantages:
- It forces me to cooperate with the JUCE
Componentpainting logic, even when I don’t need it.
- Continuous repainting takes a long time (eg: dealing with thread jobs, image caching)
- The present implementation takes up a lot more memory. (Forced to allocate a variety of things for
Componentpainting logic, and image caching)
I really would prefer a custom method that allows me to directly render to the window/surface on the main thread.
The framework provides many classes and mechanisms that can almost get me there… lots of hacking required to get it sorted out as something separate from the current implementation.