com.ebay.bascomtask.main

Interface ITask

public interface ITask

Shadow object for POJO tasks added to an orchestrator.

Author:

brendanmccarthy

Method Summary

Methods

Modifier and Type	Method and Description
ITask	after(Object pojoTask) The inverse of before(Object), achieving the same effect while allowing the arguments to be reversed.
ITask	before(Object pojoTask) Ensures this task completes before the given task.
ITask	fork() Prevents this task from being run by the calling thread, spawning a new thread if needed.
String	getName() The name of this pojoCalled instance in an orchestrator, guaranteed to always be unique that orchestrator.
Orchestrator	getOrchestrator() Returns the orchestrator which created and owns this task.
boolean	isFork() Should this task not be run in the calling thread?
boolean	isMultiMethodOk() Should an exception be thrown when a task has more than one method each of which has all parameters added to the graph?
boolean	isWait() Should the orchestrator wait for this task to finish before completing an invocation of the Orchestrator.execute() method?
ITask	multiMethodOk() Indicates that it is ok if the task has more than one method which can fire.
ITask	name(String name) Each task has a name that defaults to its type name followed by a unique integer for that type, unless overridden here.
ITask	noWait() Convenience method for setWait(false);
ITask	provides(Class<?> pojoTaskClass) Indicates that a task will add to the current orchestrator a pojo task instance of the given class.
ITask	wait(boolean wait) Specifies whether the orchestrator should wait for this instance to finish.

Method Detail

getOrchestrator

Orchestrator getOrchestrator()

Returns the orchestrator which created and owns this task. An ITask is created by an orchestrator e.g. when POJO tasks are added to it.

Returns:

owning orchestrator

See Also:

Orchestrator.addWork(Object), Orchestrator.addPassThru(Object), Orchestrator.addConditionally(Object, boolean)

getName

String getName()

The name of this pojoCalled instance in an orchestrator, guaranteed to always be unique that orchestrator.

Returns:

the user-supplied or auto-generated name, will never be null

name

ITask name(String name)

Each task has a name that defaults to its type name followed by a unique integer for that type, unless overridden here.

Parameters:

name - new name for task, any previous name

Returns:

task name

Throws:

InvalidTask.NameConflict - if a task with this name already exists

isWait

boolean isWait()

Should the orchestrator wait for this task to finish before completing an invocation of the Orchestrator.execute() method?

Returns:

true iff this task should be treated as a 'nowait' task

wait

ITask wait(boolean wait)

Specifies whether the orchestrator should wait for this instance to finish. The default is true. Even if this value is set to false, the orchestrator may still wait for another task that is dependent on this task, and that task has isWait()==true.

Parameters:

wait - behavior on orchestrator execute() exit

Returns:

this task

See Also:

Orchestrator.execute()

noWait

ITask noWait()

Convenience method for setWait(false);

Returns:

this task

multiMethodOk

ITask multiMethodOk()

Indicates that it is ok if the task has more than one method which can fire. By default, this would be flagged as an error.

Returns:

this task

isMultiMethodOk

boolean isMultiMethodOk()

Should an exception be thrown when a task has more than one method each of which has all parameters added to the graph?

Returns:

true iff this task should not reject multiple matching methods

fork

ITask fork()

Prevents this task from being run by the calling thread, spawning a new thread if needed. Usually only needed if there is work needed to be done by the calling thread and it is desirable to get to that work without waiting for this task to complete.

Returns:

this task

isFork

boolean isFork()

Should this task not be run in the calling thread?

Returns:

true iff this task should never be run in the main orchestrator thread

before

ITask before(Object pojoTask)

Ensures this task completes before the given task. Useful when BascomTask's autowiring needs fine tuning, for example when there is a hidden dependency between tasks that is not otherwise exposed through the task dependency tree. If applied between two tasks that already have an autowired dependency, will limit the cardinality of the dependency to only those exposed explicitly in this way. Consider this case: if B depends on A and there are two instances of A, then by default autowiring B would depend on both instances of A. If on the other hand an explicitly wired dependency was set up along the lines of a1.before(b), then the B instance b would only depend on that singular A instances a1.

This method can safely be called before pojoTask is added to the graph, but will result in an error on the next Orchestrator.execute() if pojoTask has not been added by that time.

Parameters:

pojoTask - or ITask which should run before this one is started

Returns:

this task

after

ITask after(Object pojoTask)

The inverse of before(Object), achieving the same effect while allowing the arguments to be reversed.

Parameters:

pojoTask - or ITask which should only run after this one

Returns:

this task

provides

ITask provides(Class<?> pojoTaskClass)

Indicates that a task will add to the current orchestrator a pojo task instance of the given class. This is necessary only when (a) a task is dynamically adding to the orchestrator and (b) that task is adding an instance upon which tasks at the outer level depend. Without invoking this method, Orchestrator.execute() will fail at the outer level, asserting that the graph is incomplete -- the orchestrator has no way to peer inside the inner task code to determine when/if an instance will be added.

If the nested task does not in fact add the promised instance, an exception will be throw when it terminates.

Parameters:

pojoTaskClass - for task that will be provided when this task is invoked

Returns:

this task