I scratched out that god forsaken memory management beast from coreloop and have started replacing it with a backwards approach using mutex locks. The idea is the same, the coreloop thread will still run save for inter-cycle Python calls and the Python interpreter will wait until the end of the current render/physics cycle to make changes.
For those who were not part of this discussion a few months back, neither GLUT, nor ODE, nor Python are "thread-safe". We must do that ourselves. Theora - while there are video extensions to opengl they're not common. We should instead render Theora to a texture and let users use it where ever textures can be used - for an image widget, for a mesh, for a sprite, etc. Sending it to projector may seem intuitive but on the backend it'd be a hack. _______________________________________________ PySoy-Dev mailing list [email protected] http://www.pysoy.org/mailman/listinfo/pysoy-dev
