Re: [LAU] Bristol B3 and stuck notes

The issue was debuged with the assistance of Andrew and Damon collecting debug information and the issue was due to the B3 manipulating some voice flags to ensure the percussive envelope was played legato style. Extended the semaphore to ensure exclusion of the MIDI and audio threads to include the engine 'postOps' which was previously not the case. Release 0.40.3 has been uploaded with this and other fixes. The problem only occured on HT or multicore systems, my old development system is single core and this affected the race condition. There were about 4 changes that applied to this one issue however the Jack interfaces was cleaned up, providing diagnostics for registration failures when the Jack ports table is full, using jack_client_open() rather than the deprecated jack_client_new(), and more importantly bristol no longer causes subgraph timeouts on exit. This required some changes to the cleanup processes - the audio thread no longer does all its cleanup as that caused the timeouts. The MIDI thread takes care GUI signalling as such like. Nick. _________________________________________________________________ More than messages–check out the rest of the Windows Live™. http://www.microsoft.com/windows/windowslive/