Skip to content

Ability to create custom schedulers with behavior based on composing operators. #4154

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Jul 9, 2016
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
78 changes: 78 additions & 0 deletions src/main/java/rx/Scheduler.java
Original file line number Diff line number Diff line change
Expand Up @@ -17,7 +17,10 @@

import java.util.concurrent.TimeUnit;

import rx.annotations.Experimental;
import rx.functions.Action0;
import rx.functions.Func1;
import rx.internal.schedulers.SchedulerWhen;
import rx.schedulers.Schedulers;
import rx.subscriptions.MultipleAssignmentSubscription;

Expand Down Expand Up @@ -182,4 +185,79 @@ public long now() {
return System.currentTimeMillis();
}

/**
* Allows the use of operators for controlling the timing around when
* actions scheduled on workers are actually done. This makes it possible to
* layer additional behavior on this {@link Scheduler}. The only parameter
* is a function that flattens an {@link Observable} of {@link Observable}
* of {@link Completable}s into just one {@link Completable}. There must be
* a chain of operators connecting the returned value to the source
* {@link Observable} otherwise any work scheduled on the returned
* {@link Scheduler} will not be executed.
* <p>
* When {@link Scheduler#createWorker()} is invoked a {@link Observable} of
* {@link Completable}s is onNext'd to the combinator to be flattened. If
* the inner {@link Observable} is not immediately subscribed to an calls to
* {@link Worker#schedule} are buffered. Once the {@link Observable} is
* subscribed to actions are then onNext'd as {@link Completable}s.
* <p>
* Finally the actions scheduled on the parent {@link Scheduler} when the
* inner most {@link Completable}s are subscribed to.
* <p>
* When the {@link Worker} is unsubscribed the {@link Completable} emits an
* onComplete and triggers any behavior in the flattening operator. The
* {@link Observable} and all {@link Completable}s give to the flattening
* function never onError.
* <p>
* Limit the amount concurrency two at a time without creating a new fix
* size thread pool:
*
* <pre>
* Scheduler limitSched = Schedulers.computation().when(workers -> {
* // use merge max concurrent to limit the number of concurrent
* // callbacks two at a time
* return Completable.merge(Observable.merge(workers), 2);
* });
* </pre>
* <p>
* This is a slightly different way to limit the concurrency but it has some
* interesting benefits and drawbacks to the method above. It works by
* limited the number of concurrent {@link Worker}s rather than individual
* actions. Generally each {@link Observable} uses its own {@link Worker}.
* This means that this will essentially limit the number of concurrent
* subscribes. The danger comes from using operators like
* {@link Observable#zip(Observable, Observable, rx.functions.Func2)} where
* subscribing to the first {@link Observable} could deadlock the
* subscription to the second.
*
* <pre>
* Scheduler limitSched = Schedulers.computation().when(workers -> {
* // use merge max concurrent to limit the number of concurrent
* // Observables two at a time
* return Completable.merge(Observable.merge(workers, 2));
* });
* </pre>
*
* Slowing down the rate to no more than than 1 a second. This suffers from
* the same problem as the one above I could find an {@link Observable}
* operator that limits the rate without dropping the values (aka leaky
* bucket algorithm).
*
* <pre>
* Scheduler slowSched = Schedulers.computation().when(workers -> {
* // use concatenate to make each worker happen one at a time.
* return Completable.concat(workers.map(actions -> {
* // delay the starting of the next worker by 1 second.
* return Completable.merge(actions.delaySubscription(1, TimeUnit.SECONDS));
* }));
* });
* </pre>
*
* @param combine
* @return
*/
@Experimental
public <S extends Scheduler & Subscription> S when(Func1<Observable<Observable<Completable>>, Completable> combine) {
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Experimental?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

do'h

return (S) new SchedulerWhen(combine, this);
}
}
302 changes: 302 additions & 0 deletions src/main/java/rx/internal/schedulers/SchedulerWhen.java
Original file line number Diff line number Diff line change
@@ -0,0 +1,302 @@
/**
* Copyright 2016 Netflix, Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package rx.internal.schedulers;

import java.util.concurrent.TimeUnit;
import java.util.concurrent.atomic.AtomicBoolean;
import java.util.concurrent.atomic.AtomicReference;

import rx.Completable;
import rx.Completable.CompletableOnSubscribe;
import rx.Completable.CompletableSubscriber;
import rx.Scheduler.Worker;
import rx.Observable;
import rx.Observer;
import rx.Scheduler;
import rx.Subscription;
import rx.annotations.Experimental;
import rx.functions.Action0;
import rx.functions.Func1;
import rx.internal.operators.BufferUntilSubscriber;
import rx.observers.SerializedObserver;
import rx.subjects.PublishSubject;
import rx.subscriptions.Subscriptions;

/**
* Allows the use of operators for controlling the timing around when actions
* scheduled on workers are actually done. This makes it possible to layer
* additional behavior on this {@link Scheduler}. The only parameter is a
* function that flattens an {@link Observable} of {@link Observable} of
* {@link Completable}s into just one {@link Completable}. There must be a chain
* of operators connecting the returned value to the source {@link Observable}
* otherwise any work scheduled on the returned {@link Scheduler} will not be
* executed.
* <p>
* When {@link Scheduler#createWorker()} is invoked a {@link Observable} of
* {@link Completable}s is onNext'd to the combinator to be flattened. If the
* inner {@link Observable} is not immediately subscribed to an calls to
* {@link Worker#schedule} are buffered. Once the {@link Observable} is
* subscribed to actions are then onNext'd as {@link Completable}s.
* <p>
* Finally the actions scheduled on the parent {@link Scheduler} when the inner
* most {@link Completable}s are subscribed to.
* <p>
* When the {@link Worker} is unsubscribed the {@link Completable} emits an
* onComplete and triggers any behavior in the flattening operator. The
* {@link Observable} and all {@link Completable}s give to the flattening
* function never onError.
* <p>
* Limit the amount concurrency two at a time without creating a new fix size
* thread pool:
*
* <pre>
* Scheduler limitSched = Schedulers.computation().when(workers -> {
* // use merge max concurrent to limit the number of concurrent
* // callbacks two at a time
* return Completable.merge(Observable.merge(workers), 2);
* });
* </pre>
* <p>
* This is a slightly different way to limit the concurrency but it has some
* interesting benefits and drawbacks to the method above. It works by limited
* the number of concurrent {@link Worker}s rather than individual actions.
* Generally each {@link Observable} uses its own {@link Worker}. This means
* that this will essentially limit the number of concurrent subscribes. The
* danger comes from using operators like
* {@link Observable#zip(Observable, Observable, rx.functions.Func2)} where
* subscribing to the first {@link Observable} could deadlock the subscription
* to the second.
*
* <pre>
* Scheduler limitSched = Schedulers.computation().when(workers -> {
* // use merge max concurrent to limit the number of concurrent
* // Observables two at a time
* return Completable.merge(Observable.merge(workers, 2));
* });
* </pre>
*
* Slowing down the rate to no more than than 1 a second. This suffers from the
* same problem as the one above I could find an {@link Observable} operator
* that limits the rate without dropping the values (aka leaky bucket
* algorithm).
*
* <pre>
* Scheduler slowSched = Schedulers.computation().when(workers -> {
* // use concatenate to make each worker happen one at a time.
* return Completable.concat(workers.map(actions -> {
* // delay the starting of the next worker by 1 second.
* return Completable.merge(actions.delaySubscription(1, TimeUnit.SECONDS));
* }));
* });
* </pre>
*
* @param combine
* @return
*/
@Experimental
public class SchedulerWhen extends Scheduler implements Subscription {
private final Scheduler actualScheduler;
private final Observer<Observable<Completable>> workerObserver;
private final Subscription subscription;

public SchedulerWhen(Func1<Observable<Observable<Completable>>, Completable> combine, Scheduler actualScheduler) {
this.actualScheduler = actualScheduler;
// workers are converted into completables and put in this queue.
PublishSubject<Observable<Completable>> workerSubject = PublishSubject.create();
this.workerObserver = new SerializedObserver<Observable<Completable>>(workerSubject);
// send it to a custom combinator to pick the order and rate at which
// workers are processed.
this.subscription = combine.call(workerSubject.onBackpressureBuffer()).subscribe();
}

@Override
public void unsubscribe() {
subscription.unsubscribe();
}

@Override
public boolean isUnsubscribed() {
return subscription.isUnsubscribed();
}

@Override
public Worker createWorker() {
final Worker actualWorker = actualScheduler.createWorker();
// a queue for the actions submitted while worker is waiting to get to
// the subscribe to off the workerQueue.
BufferUntilSubscriber<ScheduledAction> actionSubject = BufferUntilSubscriber.<ScheduledAction>create();
final Observer<ScheduledAction> actionObserver = new SerializedObserver<ScheduledAction>(actionSubject);
// convert the work of scheduling all the actions into a completable
Observable<Completable> actions = actionSubject.map(new Func1<ScheduledAction, Completable>() {
@Override
public Completable call(final ScheduledAction action) {
return Completable.create(new CompletableOnSubscribe() {
@Override
public void call(CompletableSubscriber actionCompletable) {
actionCompletable.onSubscribe(action);
action.call(actualWorker);
actionCompletable.onCompleted();
}
});
}
});

// a worker that queues the action to the actionQueue subject.
Worker worker = new Worker() {
private final AtomicBoolean unsubscribed = new AtomicBoolean();

@Override
public void unsubscribe() {
// complete the actionQueue when worker is unsubscribed to make
// room for the next worker in the workerQueue.
if (unsubscribed.compareAndSet(false, true)) {
actualWorker.unsubscribe();
actionObserver.onCompleted();
}
}

@Override
public boolean isUnsubscribed() {
return unsubscribed.get();
}

@Override
public Subscription schedule(final Action0 action, final long delayTime, final TimeUnit unit) {
// send a scheduled action to the actionQueue
DelayedAction delayedAction = new DelayedAction(action, delayTime, unit);
actionObserver.onNext(delayedAction);
return delayedAction;
}

@Override
public Subscription schedule(final Action0 action) {
// send a scheduled action to the actionQueue
ImmediateAction immediateAction = new ImmediateAction(action);
actionObserver.onNext(immediateAction);
return immediateAction;
}
};

// enqueue the completable that process actions put in reply subject
workerObserver.onNext(actions);

// return the worker that adds actions to the reply subject
return worker;
}

private static final Subscription SUBSCRIBED = new Subscription() {
@Override
public void unsubscribe() {
}

@Override
public boolean isUnsubscribed() {
return false;
}
};

private static final Subscription UNSUBSCRIBED = Subscriptions.unsubscribed();

@SuppressWarnings("serial")
private static abstract class ScheduledAction extends AtomicReference<Subscription> implements Subscription {
public ScheduledAction() {
super(SUBSCRIBED);
}

private final void call(Worker actualWorker) {
Subscription oldState = get();
// either SUBSCRIBED or UNSUBSCRIBED
if (oldState == UNSUBSCRIBED) {
// no need to schedule return
return;
}
if (oldState != SUBSCRIBED) {
// has already been scheduled return
// should not be able to get here but handle it anyway by not
// rescheduling.
return;
}

Subscription newState = callActual(actualWorker);

if (!compareAndSet(SUBSCRIBED, newState)) {
// set would only fail if the new current state is some other
// subscription from a concurrent call to this method.
// Unsubscribe from the action just scheduled because it lost
// the race.
newState.unsubscribe();
}
}

protected abstract Subscription callActual(Worker actualWorker);

@Override
public boolean isUnsubscribed() {
return get().isUnsubscribed();
}

@Override
public void unsubscribe() {
Subscription oldState;
// no matter what the current state is the new state is going to be
Subscription newState = UNSUBSCRIBED;
do {
oldState = get();
if (oldState == UNSUBSCRIBED) {
// the action has already been unsubscribed
return;
}
} while (!compareAndSet(oldState, newState));

if (oldState != SUBSCRIBED) {
// the action was scheduled. stop it.
oldState.unsubscribe();
}
}
}

@SuppressWarnings("serial")
private static class ImmediateAction extends ScheduledAction {
private final Action0 action;

public ImmediateAction(Action0 action) {
this.action = action;
}

@Override
protected Subscription callActual(Worker actualWorker) {
return actualWorker.schedule(action);
}
}

@SuppressWarnings("serial")
private static class DelayedAction extends ScheduledAction {
private final Action0 action;
private final long delayTime;
private final TimeUnit unit;

public DelayedAction(Action0 action, long delayTime, TimeUnit unit) {
this.action = action;
this.delayTime = delayTime;
this.unit = unit;
}

@Override
protected Subscription callActual(Worker actualWorker) {
return actualWorker.schedule(action, delayTime, unit);
}
}
}
Loading