Jobs in a job stream should not be enqueued twice when a job is stopped

Current Situation

Assuming:
in condition is not marked as consumed
a job is stopped due to an error in a job stream

then the task for the job ends. Then job will be stopped due to an error. The event is still there and the in condition will be evaluated to true. The task will not start but enqueued. So far it is the desired behavior

With the next evaluation of the in condition a new task will be enqueued.

Desired Behavior

With the next evaluation the in condition still should be disabled. The reasonm for this issue is that the in condition is enabled already on the task_closed event. But it should only be enabled on the task_end event.