Reply to thread

That would probably work but it sounds like it would require some changes that might introduce a bit of risk. If this is truly a threading issue, we could just wrap the remove logic and the render logic in lock statements based on the texture object. That way it would be impossible to unload a texture while MP is trying to use it to render in another thread. I think this would both solve the issue and keep the risk of new bugs being introduced to a minimum.

I'll try to put together a patch and a test core.dll tonight.