On 04-14, Aurélien Leblond wrote:
  I have been debugging an issue in Giada where
sometimes it would get
 disconnected from Jack1 at start (not Jack2).
 During my debugging, I found that the JackSyncCallback method setup in
 Giada's jack_set_sync_callback uses Fl::lock which blocks the main
 process and Jack doesn't like that.
 I believe that's what the problem is (especially as after removing
 them I cannot reproduce it anymore)...
 Does anyone who knows Jack much better than I do (in fact I'm a real
 noob in that area) think I'm right in my investigation's finding? 
From 
http://jackaudio.org/api/group__TransportControl.html
"This realtime function must not wait."
So, this blocking behavior should not exist and can cause all sorts of
issues.
After running stoat on giada there are a few issues in both the process
callback and the sync callback.
These issues are now documented on giada's github issue tracker:
https://github.com/monocasual/giada/issues/49
--Mark