S?bastien Bourdeauducq wrote: > This is not true - there is a timeout (at the beginning of > handle_midi_event) that went back to the first byte of the message > after nothing had been received in a while.
Ah, I see. I didn't pay much attention to it, just had a quick look and mentally mis-filed it under "timestamping". Interestingly, when a MIDI clock [1] was present, it may have defeated that timeout mechanism with its ~20-80 Hz message rate. As you observed on #milkymist, your MIDI setup doesn't use the MIDI clock, which may explain why you didn't run into all these problems. [1] http://home.roadrunner.com/~jgglatt/tech/midispec/clock.htm > I'd swear I read somewhere there was no other way to sync MIDI > streams... MIDI is an ancient but - believe it or not - still for-pay standard. It's hardly surprising that all sorts of myths and half-truths are floating around :-( > Anyway, the timeout should be removed (which your patch > doesn't do) and replaced with your better solution. I see that you already did this. Thanks ! - Werner _______________________________________________ http://lists.milkymist.org/listinfo.cgi/devel-milkymist.org IRC: #milkymist@Freenode
