freenet.support

Class DummyJobRunner

java.lang.Object

freenet.support.DummyJobRunner

All Implemented Interfaces:

PersistentJobRunner

public class DummyJobRunner
extends java.lang.Object
implements PersistentJobRunner

A PersistentJobRunner that isn't persistent. Convenient for transient requests, code doesn't need messy if(persistent) everywhere.

Nested Class Summary

Nested classes/interfaces inherited from interface freenet.client.async.PersistentJobRunner

PersistentJobRunner.CheckpointLock

Constructor Summary

Constructors

Constructor and Description
DummyJobRunner(Executor executor, ClientContext context)

Method Summary

Modifier and Type	Method and Description
boolean	hasLoaded() Has the queue started yet?
PersistentJobRunner.CheckpointLock	lock() Obtain a lock which will prevent checkpointing until it is unlocked.
boolean	newSalt() For persistent requests, return true if the bloom filter salt has changed when loading the requests.
void	queue(PersistentJob job, int priority) Start a job immediately unless we are about to write a checkpoint.
void	queueInternal(PersistentJob job) Start an "internal" job.
void	queueInternal(PersistentJob job, int threadPriority) Start an "internal" job.
void	queueNormalOrDrop(PersistentJob job) Queue the job at low thread priority or drop it if persistence is disabled.
void	setCheckpointASAP() Commit ASAP.
boolean	shuttingDown() If true, the node is shutting down

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

DummyJobRunner

public DummyJobRunner(Executor executor,
                      ClientContext context)

Method Detail

queue

public void queue(PersistentJob job,
                  int priority)

Description copied from interface: PersistentJobRunner

Start a job immediately unless we are about to write a checkpoint. If we are, queue the job until after the checkpoint has completed. We do not store the jobs, so this should be used for tasks which either are "external" and so will be retried after an unclean shutdown (for example, fetching a block from the datastore or the network), or where we will check for the situation on restart in onResume() (for example, the callback that starts a FEC decode). This should also be used for e.g. freeing on-disk files after a checkpoint.

Specified by:

queue in interface PersistentJobRunner

Parameters:

job - The job to run now or after the checkpoint.

priority - The priority of the job.

queueNormalOrDrop

public void queueNormalOrDrop(PersistentJob job)

Description copied from interface: PersistentJobRunner

Queue the job at low thread priority or drop it if persistence is disabled.

Specified by:

queueNormalOrDrop in interface PersistentJobRunner

setCheckpointASAP

public void setCheckpointASAP()

Description copied from interface: PersistentJobRunner

Commit ASAP. Can also be set via returning true from a PersistentJob, but it's useful to be able to do it "inline".

Specified by:

setCheckpointASAP in interface PersistentJobRunner

hasLoaded

public boolean hasLoaded()

Description copied from interface: PersistentJobRunner

Has the queue started yet?

Specified by:

hasLoaded in interface PersistentJobRunner

queueInternal

public void queueInternal(PersistentJob job,
                          int threadPriority)

Description copied from interface: PersistentJobRunner

Start an "internal" job. We will not checkpoint until all the internal jobs have finished; we do not queue them at all. Hence a series of internal jobs is atomic. This should be used for stuff like creating the next stage of a request, where storing the half-way state would lead to it potentially not restarting properly after shutdown. It MUST NOT be used for events from outside the client layer, including finding blocks in the datastore, on the network etc. FIXME this doesn't queue at all. Come up with a better name! :)

Specified by:

queueInternal in interface PersistentJobRunner

queueInternal

public void queueInternal(PersistentJob job)

Description copied from interface: PersistentJobRunner

Start an "internal" job. We will not checkpoint until all the internal jobs have finished; we do not queue them at all. Hence a series of internal jobs is atomic. This should be used for stuff like creating the next stage of a request, where storing the half-way state would lead to it potentially not restarting properly after shutdown. It MUST NOT be used for events from outside the client layer, including finding blocks in the datastore, on the network etc. Often when we call this we could have continued the job on the same thread. That's not always the best thing to do however, we often move stuff to another job to minimise locking or increase throughput. FIXME this doesn't queue at all. Come up with a better name! :)

Specified by:

queueInternal in interface PersistentJobRunner

lock

public PersistentJobRunner.CheckpointLock lock()

Description copied from interface: PersistentJobRunner

Obtain a lock which will prevent checkpointing until it is unlocked. This counts as a thread, and can be used for e.g. MemoryLimitedJob's that can change persistent data. Like any lock, you MUST use a try/finally block. lock() may block if a checkpoint is in progress or is scheduled.

Specified by:

lock in interface PersistentJobRunner

newSalt

public boolean newSalt()

Description copied from interface: PersistentJobRunner

For persistent requests, return true if the bloom filter salt has changed when loading the requests. In which case all the Bloom filters will be invalid and will need to be recomputed.

Specified by:

newSalt in interface PersistentJobRunner

shuttingDown

public boolean shuttingDown()

Description copied from interface: PersistentJobRunner

If true, the node is shutting down

Specified by:

shuttingDown in interface PersistentJobRunner