On Wed, 13 Apr 2022, Raul Miller wrote:
Note also that this can create a situation where jbreak does not regain control of the session.
IMO this is an orthogonal issue. Given any useful set of concurrency primitives, it will be possible to deadlock. It's the responsibility of the environment to let you introspect and recover state in such a situation. IOW: if a given task waiting for a pyx (or a lock/queue/...whatever we come up with), it should be possible to interrupt that waiting (without having any effect on the task associated with that pyx).
---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
