* Special means of indicating that an executable will proceed in the background without consuming a native thread ({@link Executor}).
* May be thrown from {@link hudson.model.Queue.Executable#run} after doing any preparatory work synchronously.
* May be thrown from {@link Executable#run} after doing any preparatory work synchronously.
* <p>{@link Executor#isActive} will remain true (even though {@link Executor#isAlive} is not) until {@link #completed} is called.
* The thrower will need to hold on to a reference to this instance as a handle to call {@link #completed}.
* <p>The execution may not extend into another Jenkins session; if you wish to model a long-running execution, you must schedule a new task after restart.
* This class is not serializable anyway.
* <p>Mainly intended for use with {@link OneOffExecutor} (from a {@link hudson.model.Queue.FlyweightTask}), of which there could be many,
* <p>Mainly intended for use with {@link OneOffExecutor} (from a {@link FlyweightTask}), of which there could be many,
* but could also be used with a heavyweight executor even though the number of executors is bounded by node configuration.
* <p>{@link ResourceController}/{@link ResourceActivity}/{@link ResourceList}/{@link Resource} are not currently supported.
* Nor are {@link hudson.model.Queue.Task#getSubTasks} other than the primary task.
* To be called when the task is actually complete.
* @param error normally null (preferable to handle errors yourself), but may be specified to simulate an exception from {@link hudson.model.Queue.Executable#run}, as per {@link ExecutorListener#taskCompletedWithProblems}
* @param error normally null (preferable to handle errors yourself), but may be specified to simulate an exception from {@link Executable#run}, as per {@link ExecutorListener#taskCompletedWithProblems}