"In most cases" is exactly the problem. So ChucK isn't going to fail and ruin my live performance.. in most cases? Great! :)
Good point - except that is not what I meant by "in most cases". The callback is designed in a way such that it always favors the synthesis thread to compute apart from the callback thread. But even if the callback thread has to wait (a reasonable amount), which it doesn't "in most cases", it does not interfere with real-time audio. When the processor is maxed, bad things will happen to real-time audio no matter what system you are running. The guarantee in ChucK is that the computations and audio synthesis will still be correct, though potentially late. I am going to save the other points for chuck-dev. We are totally open to changes if it makes ChucK more reliable - but there are reasons for doing things as they are. (I know that may seem shocking!) Best, Ge!