there is a real bug with either dj* python or the audio C++ source. When music and audio are disabled, i get significant cpu usage from the music thread doing nothing. Just looping around trying to enable audio and failing etc. When music and audio are enabled, i get almost no cpu usage from the audio thread.
enabling and disabling sound and music were done through vssetup. You should be able to recreate this and it's readily noticable on dual core systems where you can monitor individual core cpu usage.
Bascially, i was seeing 40% cpu usage with audio and music disabled, 20% on one core 20% on the other. The one core's usage dropped down to below 10% when music and audio were enabled.
I thought I saw this error too a long while... but I forgot about it when that infinite loading process didn't include sucking up all the available RAM within minutes.
I don't see why the Music class should do anything at all if sound is disabled.
I guess this needs to be fixed as well as the other loading thread bug.