开始接触Rxjava是在16年年底，当时由于自身水平的局限，并没有理解到这个框架的优秀与强大。现在两年多时间过去了，辗转了几个项目之后，从最初只知道与Retrofit配合进行网络请求，到现在已经可以灵活的把Rxjava运用到几乎所有的异步操作中，下面我想结合Rxjava的源码，谈谈如何使用这个框架。
首先，我们来看看Rxjava最基础的用法。
Flowable.create(new FlowableOnSubscribe() {
@Override
public void subscribe(FlowableEmitter emitter) throws Exception {
int i = 0;
while (i < 10) {
SystemClock.sleep(1000);
emitter.onNext(i++);
}
emitter.onComplete();
}
}, BackpressureStrategy.BUFFER).subscribeOn(Schedulers.io())
.observeOn(AndroidSchedulers.mainThread())
.subscribe(new Subscriber() {
@Override
public void onSubscribe(Subscription s) {
s.request(Long.MAX_VALUE);
}
@Override
public void onNext(Integer integer) {
Log.e(TAG, "onNext: " + integer);
}
@Override
public void onError(Throwable t) {
}
@Override
public void onComplete() {
Log.e(TAG, "onComplete: ");
}
});
这段代码非常简单，开启了一个子线程，每隔一秒发射一个int数出来，回调方法触发后把这个int数打印到控制台，那么我们开始查看源码，看看整个过程到底发生了什么。
1.create方法
public static Flowable create(FlowableOnSubscribe source, BackpressureStrategy mode) {
ObjectHelper.requireNonNull(source, "source is null");
ObjectHelper.requireNonNull(mode, "mode is null");
return RxJavaPlugins.onAssembly(new FlowableCreate(source, mode));
}
create方法的源码如下，现在我们把不重要的部分忽略，就可以剥离出关键的代码：
- BackpressureStrategy是Rxjava2的新特性，忽略掉
- ObjectHelper的静态方法，是用来判断是不是null的，忽略掉
- onAssembly方法各位可以点过去看看，它的作用是「如果当前给Rxjava设置过一个Function，则此时把这个Funtion应用给Flowable」，不在我们此次研究的范围之内，忽略掉
这样一来，结论就很清晰了，我们需要关注的就是：
- FlowableOnSubscribe
- FlowableCreate
第一个关键点：
public interface FlowableOnSubscribe {
/**
* Called for each Subscriber that subscribes.
* @param emitter the safe emitter instance, never null
* @throws Exception on error
*/
void subscribe(@NonNull FlowableEmitter emitter) throws Exception;
}
FlowableOnSubscribe是一个接口，包含一个方法，这个方法在subscriber的subscribe方法执行时被调用
第二个关键点：
public final class FlowableCreate extends Flowable {
final FlowableOnSubscribe source;
final BackpressureStrategy backpressure;
public FlowableCreate(FlowableOnSubscribe source, BackpressureStrategy backpressure){
this.source = source;
this.backpressure = backpressure;
}
@Override
public void subscribeActual(Subscriber t) {
BaseEmitter emitter;
switch (backpressure) {
case MISSING: {
emitter = new MissingEmitter(t);
break;
}
case ERROR: {
emitter = new ErrorAsyncEmitter(t);
break;
}
case DROP: {
emitter = new DropAsyncEmitter(t);
break;
}
case LATEST: {
emitter = new LatestAsyncEmitter(t);
break;
}
default: {
emitter = new BufferAsyncEmitter(t, bufferSize());
break;
}
}
t.onSubscribe(emitter);
try {
source.subscribe(emitter);
} catch (Throwable ex) {
Exceptions.throwIfFatal(ex);
emitter.onError(ex);
}
}
}
可以看到，FlowableCreate继承Flowable类，我们先来简单的研究一下Flowable这个类：
Flowable是整个Rxjava中最重要的类，它实现了Publisher接口，这个接口中包含的其实就是Flowable最核心的subscribe方法，我们看看subscribe的源码：
public final void subscribe(FlowableSubscriber s) {
ObjectHelper.requireNonNull(s, "s is null");
try {
Subscriber z = RxJavaPlugins.onSubscribe(this, s);
ObjectHelper.requireNonNull(z, "The RxJavaPlugins.onSubscribe hook returned a null FlowableSubscriber. Please check the handler provided to RxJavaPlugins.setOnFlowableSubscribe for invalid null returns. Further reading: https://github.com/ReactiveX/RxJava/wiki/Plugins");
subscribeActual(z);
} catch (NullPointerException e) { // NOPMD
throw e;
} catch (Throwable e) {
Exceptions.throwIfFatal(e);
// can't call onError because no way to know if a Subscription has been set or not
// can't call onSubscribe because the call might have set a Subscription already
RxJavaPlugins.onError(e);
NullPointerException npe = new NullPointerException("Actually not, but can't throw other exceptions due to RS");
npe.initCause(e);
throw npe;
}
}
所以，subscribe方法基本等于subscribeActual方法，而subscribeActual方法是抽象的，也是整个Flowable中唯一一个抽象方法。
我们再回头看FlowableCreate，它继承了Flowable方法，并且实现了subscribeActual方法，那么具体是怎样实现的呢，在上面的源码中可以看到，其实这个方法做了三件事：
- 用Subscriber对象创建出了一个emitter对象
- 执行了t.onSubscribe(emitter)这句代码。这个就比较有意思了，根据这句代码执行的位置，我们不难得出结论：subscriber的onSubscibe这个回调，和其他的三个回调有本质区别，首先，这个回调是在当前线程执行的，不受线程切换的影响，其次，这个方法是在subscrbeActural中直接调用了，也就是说不受子线程中代码执行的影响。
- 执行了source.subscribe(emitter)，也就是create时实现的那个接口的方法，换句话说，如果不调用Flowable的subscribe方法，最开始FlowableOnSubscribe接口中的那个方法中的代码是不会执行的
分析到了这里，我们可以给create方法下一个结论:
create方法，传入了一个FlowableOnSubscribe接口的子类对象，通过这个对象实现了Flowable的subscribeActual方法，从而得到了一个Flowable的对象。
2.线程是如何切换的
a.如何切换订阅的线程
下面是第一段源码
public final Flowable subscribeOn(@NonNull Scheduler scheduler) {
//第一句判断scheduler是否为空
ObjectHelper.requireNonNull(scheduler, "scheduler is null");
//然后调用另一个重载方法，第二个参数是「当前Flowable对象是否属于刚刚创建的Flowable」
return subscribeOn(scheduler, !(this instanceof FlowableCreate));
}
调用的重载方法里面如何处理的呢
public final Flowable subscribeOn(@NonNull Scheduler scheduler, boolean requestOn) {
ObjectHelper.requireNonNull(scheduler, "scheduler is null");
//用传入的scheduler对象将当前Flowable对象转换成FlowableSubscribeOn对象
return RxJavaPlugins.onAssembly(new FlowableSubscribeOn(this, scheduler, requestOn));
}
那么FlowableSubscribeOn是怎么一回事呢，首先，这个类继承AbstractFlowableWithUpstream
abstract class AbstractFlowableWithUpstream extends Flowable implements HasUpstreamPublisher {
/**
* The upstream source Publisher.
*/
protected final Flowable source;
/**
* Constructs a FlowableSource wrapping the given non-null (verified)
* source Publisher.
* @param source the source (upstream) Publisher instance, not null (verified)
*/
AbstractFlowableWithUpstream(Flowable source) {
this.source = ObjectHelper.requireNonNull(source, "source is null");
}
@Override
public final Publisher source() {
return source;
}
}
我们分析一下这个类，AbstractFlowableWithUpstream这个类有两个泛型，第二个泛型是给父类使用的，第一个泛型则是给自己的一个成员变量使用的，这个成员变量也是一个Flowable。
总结一下：AbstractFlowableWithUpstream继承Flowable，但是它额外持有一个Flowable对象作为成员变量，两个Flowable分别对应两个泛型，AbstractFlowableWithUpstream自身的泛型是当前要输出的数据类型，持有的成员变量Flowable的代表的是上游的数据类型，所以AbstractFlowableWithUpstream应该理解为一个带有上游Flowable的Flowable。
所以回头看subscribeOn方法，其实是创建了一个新的Flowable对象，原先的Flowable对象作为成员变量保留在新的Flowable中。
这里最后一个需要注意的点，是FlowableSubscribeOn类，它的subscribeActual方法代码如下
@Override
public void subscribeActual(final Subscriber s) {
Scheduler.Worker w = scheduler.createWorker();
final SubscribeOnSubscriber sos = new SubscribeOnSubscriber(s, w, source, nonScheduledRequests);
s.onSubscribe(sos);
w.schedule(sos);
}
这里有一个SubscribeOnSubscriber，它是FlowableSubscribeOn的一个静态内部类，源码如下
static final class SubscribeOnSubscriber extends AtomicReference
implements FlowableSubscriber, Subscription, Runnable {
private static final long serialVersionUID = 8094547886072529208L;
final Subscriber downstream;
final Scheduler.Worker worker;
final AtomicReference upstream;
final AtomicLong requested;
final boolean nonScheduledRequests;
Publisher source;
SubscribeOnSubscriber(Subscriber actual, Scheduler.Worker worker, Publisher source, boolean requestOn) {
this.downstream = actual;
this.worker = worker;
this.source = source;
this.upstream = new AtomicReference();
this.requested = new AtomicLong();
this.nonScheduledRequests = !requestOn;
}
@Override
public void run() {
lazySet(Thread.currentThread());
Publisher src = source;
source = null;
src.subscribe(this);
}
@Override
public void onSubscribe(Subscription s) {
if (SubscriptionHelper.setOnce(this.upstream, s)) {
long r = requested.getAndSet(0L);
if (r != 0L) {
requestUpstream(r, s);
}
}
}
@Override
public void onNext(T t) {
downstream.onNext(t);
}
@Override
public void onError(Throwable t) {
downstream.onError(t);
worker.dispose();
}
@Override
public void onComplete() {
downstream.onComplete();
worker.dispose();
}
@Override
public void request(final long n) {
if (SubscriptionHelper.validate(n)) {
Subscription s = this.upstream.get();
if (s != null) {
requestUpstream(n, s);
} else {
BackpressureHelper.add(requested, n);
s = this.upstream.get();
if (s != null) {
long r = requested.getAndSet(0L);
if (r != 0L) {
requestUpstream(r, s);
}
}
}
}
}
void requestUpstream(final long n, final Subscription s) {
if (nonScheduledRequests || Thread.currentThread() == get()) {
s.request(n);
} else {
worker.schedule(new Request(s, n));
}
}
@Override
public void cancel() {
SubscriptionHelper.cancel(upstream);
worker.dispose();
}
static final class Request implements Runnable {
final Subscription upstream;
final long n;
Request(Subscription s, long n) {
this.upstream = s;
this.n = n;
}
@Override
public void run() {
upstream.request(n);
}
}
}
首先，SubscribeOnSubscriber实现了三个接口，分别是FlowableSubscriber，Subscription和Runnable，实现Subscription接口，是为了作为s.onSubscribe(sos)的参数，在subscribe方法执行时被调用，实现Runnable，是为了作为w.schedule(sos)的参数，也就是把run方法中的代码放到指定线程中执行，那么我们看看run方法中是什么代码，没错，是上游的Flowable的subscribe方法，也就是FlowableCreate的subscribe方法，也就是我们最开始分析的FlowableOnSubscribe中的subscribe方法中的代码。
这里还有一个细节：FlowableSubscribeOn的subscribeActual方法中我们调用了s.onSubscribe(sos)，经过前面的分析，我们也知道上游的Flowable的subscribeActual方法中也有调用Subscriber的onSubscribe方法，那么这个方法会被调用两次吗，其实是不会的，因为在FlowableSubscribeOn的subscribeActual方法中，SubscribeOnSubscriber其实是一个特殊的Subscriber它实现的接口如下：
public interface FlowableSubscriber extends Subscriber {
/**
* Implementors of this method should make sure everything that needs
* to be visible in {@link #onNext(Object)} is established before
* calling {@link Subscription#request(long)}. In practice this means
* no initialization should happen after the {@code request()} call and
* additional behavior is thread safe in respect to {@code onNext}.
*
* {@inheritDoc}
*/
@Override
void onSubscribe(@NonNull Subscription s);
}
FlowableSubscriber继承了Subscriber，但是做了一点小小的改动，onSubscribe方法由public变成了默认的，也就意味着，上游的Flowable再去调用的时候，是不成功的。
b.如何切换观察的线程
这里我们追踪源码，会发现与上面有相似之处，也是新创建一个Flowable，当前flowable作为上游的flowable，当作参数保存在新创建的flowable中，那么这个新的flowable是如何改变观察线程的呢？
我们来看看它的subscribeActual方法
@Override
public void subscribeActual(Subscriber s) {
//这里scheduler生成了一个worker对象
Worker worker = scheduler.createWorker();
//这里source(也就是上游的flowable)subscribe了新创建的ObserveOnSubscriber对象
if (s instanceof ConditionalSubscriber) {
source.subscribe(new ObserveOnConditionalSubscriber(
(ConditionalSubscriber) s, worker, delayError, prefetch));
} else {
source.subscribe(new ObserveOnSubscriber(s, worker, delayError, prefetch));
}
}
这里有一个非常重要的类，ObserveOnSubscriber，首先我们看看它的父类
abstract static class BaseObserveOnSubscriber
extends BasicIntQueueSubscription
implements FlowableSubscriber, Runnable {
private static final long serialVersionUID = -8241002408341274697L;
final Worker worker;
final boolean delayError;
final int prefetch;
final int limit;
final AtomicLong requested;
Subscription upstream;
SimpleQueue queue;
volatile boolean cancelled;
volatile boolean done;
Throwable error;
int sourceMode;
long produced;
boolean outputFused;
BaseObserveOnSubscriber(
Worker worker,
boolean delayError,
int prefetch) {
this.worker = worker;
this.delayError = delayError;
this.prefetch = prefetch;
this.requested = new AtomicLong();
this.limit = prefetch - (prefetch >> 2);
}
@Override
public final void onNext(T t) {
if (done) {
return;
}
if (sourceMode == ASYNC) {
trySchedule();
return;
}
if (!queue.offer(t)) {
upstream.cancel();
error = new MissingBackpressureException("Queue is full?!");
done = true;
}
trySchedule();
}
@Override
public final void onError(Throwable t) {
if (done) {
RxJavaPlugins.onError(t);
return;
}
error = t;
done = true;
trySchedule();
}
@Override
public final void onComplete() {
if (!done) {
done = true;
trySchedule();
}
}
@Override
public final void request(long n) {
if (SubscriptionHelper.validate(n)) {
BackpressureHelper.add(requested, n);
trySchedule();
}
}
@Override
public final void cancel() {
if (cancelled) {
return;
}
cancelled = true;
upstream.cancel();
worker.dispose();
if (getAndIncrement() == 0) {
queue.clear();
}
}
final void trySchedule() {
if (getAndIncrement() != 0) {
return;
}
worker.schedule(this);
}
@Override
public final void run() {
if (outputFused) {
runBackfused();
} else if (sourceMode == SYNC) {
runSync();
} else {
runAsync();
}
}
abstract void runBackfused();
abstract void runSync();
abstract void runAsync();
final boolean checkTerminated(boolean d, boolean empty, Subscriber a) {
if (cancelled) {
clear();
return true;
}
if (d) {
if (delayError) {
if (empty) {
cancelled = true;
Throwable e = error;
if (e != null) {
a.onError(e);
} else {
a.onComplete();
}
worker.dispose();
return true;
}
} else {
Throwable e = error;
if (e != null) {
cancelled = true;
clear();
a.onError(e);
worker.dispose();
return true;
} else
if (empty) {
cancelled = true;
a.onComplete();
worker.dispose();
return true;
}
}
}
return false;
}
@Override
public final int requestFusion(int requestedMode) {
if ((requestedMode & ASYNC) != 0) {
outputFused = true;
return ASYNC;
}
return NONE;
}
@Override
public final void clear() {
queue.clear();
}
@Override
public final boolean isEmpty() {
return queue.isEmpty();
}
}
我们首先注意，这个Observer类的几个重要方法(onNext,onError,onComplete)中都没有做什么实际的逻辑，但是他们都调用了一个trySchedule方法。继续追踪，我们可以定位到run方法中，最后是在指定的线程调用了run方法，run方法中的几个方法都是抽象的，我们来看子类的实现,比如下面这个runSync方法
@Override
void runSync() {
int missed = 1;
final Subscriber a = downstream;
final SimpleQueue q = queue;
long e = produced;
for (;;) {
long r = requested.get();
while (e != r) {
T v;
try {
v = q.poll();
} catch (Throwable ex) {
Exceptions.throwIfFatal(ex);
cancelled = true;
upstream.cancel();
//这里调用了下游subscriber的onError
a.onError(ex);
worker.dispose();
return;
}
if (cancelled) {
return;
}
if (v == null) {
cancelled = true;
//这里调用了下游subscriber的onComplete
a.onComplete();
worker.dispose();
return;
}
//这里调用了下游subscriber的onNext
a.onNext(v);
e++;
}
if (cancelled) {
return;
}
if (q.isEmpty()) {
cancelled = true;
a.onComplete();
worker.dispose();
return;
}
int w = get();
if (missed == w) {
produced = e;
missed = addAndGet(-missed);
if (missed == 0) {
break;
}
} else {
missed = w;
}
}
}
看到这里我们可以下结论了：这里通过封装了一个ObserveOnSubscriber类，每次调用这个类的onNext等等回调时，就会尝试在指定线程中触发下游的Subscriber的回调。
3.subscribe方法
前面已经分析过了，这里就不再赘述。