j.filter~: thread-safety crash
Opened this issue · 1 comments
tap commented
Using the "benchmark" tab of the j.filter~ help patcher I just saw a crash switching from the halfband.9 filter to the halfband.33.
It was a thread-safety problem: the halfband.9 was still being processed in the perform routine at the same time that the halfband.9 allpass filters were being deleted.
Ouch!
lossius commented
Is this something that we'll manage to fix in Jamoma 1, or is it depending on Jamoma 2?
Alternatively I guess that we'll have to schedule change of filter for the future, and do it in between vector processing?