rdb

I took a step towards this in 115a716a7dd30a82ed3c3691210388c3f11d4a8c, which (among other things) adds a way to register a custom wrapping function for a TypeHandle, which interrogate invokes instead of constructing...

I implemented this in #1640, going down the `tp_traverse` route. I wanted to go down the callback route initially, but I realised this won't work (for reasons I unfortunately no...

My thought at the moment is to turn `sync-video` into an enum, with additional mode "adaptive", falling back to "full" if this is not supported. This can be expanded further...

CollisionPolygon objects are great for levels, which have large flat surfaces. However, I think it performs poorly for shapes with many small triangles, because there are then many separate solids...

I am 90% sure this is a bug, but we need to compare against the behaviour of the FFP to be sure. If any lights are present we are supposed...

@serkkz you are not understanding the problem. `p3d_LightSource[0].diffuse` has a defined behaviour when no lights are present, namely that the first light will be white and all other lights will...

This is not correct. The ambient value is *added*, not *multiplied* with the light summation. The alpha issue is interesting, it may be that the alpha value needs to be...

This is unfortunate, but unless there is some other API, I don't know that it's possible to fix this. We set a hint and it's up to the driver to...

Maybe we should be setting __GL_SYNC_TO_VBLANK in Panda? I'm a little unsure because SDL doesn't do it either (it evidently suffers from the same problem).

This gives some more info about this behaviour of the NVIDIA driver: https://download.nvidia.com/XFree86/Linux-x86_64/352.79/README/openglenvvariables.html Maybe deploy-ng should set this based on the value of sync-video in the embedded configuration, if it...