A task that relies on a scheduler for ongoing execution

Scheduled tasks do not run when their step() methods are called, but instead simply put themselves on the scheduler's schedule for later execution. This means that scheduled tasks do not run during operations that cause callbacks, and so they cannot raise errors in apparently-unrelated code. They also are less likely to cause unintended or unexpected side-effects due to their executing between yield statements in other tasks. And finally, scheduled tasks can reliably signal that they were aborted due to an uncaught exception, via their aborted attribute.

There are some drawbacks, however. First, scheduled tasks require a scheduler, and the scheduler must tick() repeatedly as long as one wishes the tasks to continue running. Second, scheduled tasks can take slightly longer to task switch than unscheduled ones, because multiple callbacks are required.

Last, but far from least, when a scheduled task is resumed, it cannot be guaranteed that another task or callback has not already contravened whatever condition the task was waiting for. This is true even if only one task is waiting for that condition, since non-task callbacks or other code may be executed between the triggering of the event, and the time at which the task's resumption is scheduled.

step ( source=None,  event=NOT_GIVEN )