写作时间：2019-11-30
Spring Boot: 2.2.1 ,JDK: 1.8, IDE: IntelliJ IDEA

1. 说明

此篇从代码层面对响应式Reactor的实战，包括源代码的解析。

2. 核心概念

先看看Reactor的主要角色图：在这里插入图片描述
Operators - Publisher(生成者) / Subscriber(订阅者) 通过push的方式交互

Nothing Happens Until You subscribe(), 必须调用subscribe()方法才会执行。
Flux [ 0..N ] 执行序列 - onNext(), onComplete(), onError()
Mono [ 0..1 ] - onNext(), onComplete(), onError()

Backpressure 反压力：表示上游的生产者可以生产多个产品，订阅者可以根据自身处理能力慢条斯理按个执行。
4. Subscription订阅
5. onRequest()请求, onCancel()取消, onDispose()处理

线程调度Schedulers
6. 单个线程:
immediate() 在当前的线程数执行 /
可复用的线程single() / 独占，
新起的线程newSingle()
7. 线程池：
缓存的线程池elastic()，线程 60秒后被回收/
parallel()可复用的线程池，会创建跟CPU核数匹配的线程，线程不会被回收 /
newParallel()可以独占，新起线程池

错误处理
8. onError相当于try{} catch{} /
onErrorReturn 遇到异常的时候返回默认值 /
onErrorResume 用一段Lambda闭包处理异常逻辑
9. doOnError 异常处理 / doFinally 表示正常、异常都会处理的逻辑。

3. 工程建立

参照教程【SpringBoot 2.1 | 第一篇：构建第一个SpringBoot工程】新建一个Spring Boot项目，名字叫RPCClient，在目录src/main/java/resources 下找到配置文件application.properties，重命名为application.yml。

Spring Boot 版本选择2.2.1，依赖勾选Developer Tools > Lombok.
在这里插入图片描述

pom.xml添加reactor的依赖

<dependency>
    <groupId>io.projectreactor</groupId>
    <artifactId>reactor-core</artifactId>
</dependency>

4. 在主线程调用

package com.zgpeace.reactor;

import lombok.extern.slf4j.Slf4j;
import org.springframework.boot.ApplicationArguments;
import org.springframework.boot.ApplicationRunner;
import org.springframework.boot.SpringApplication;
import org.springframework.boot.autoconfigure.SpringBootApplication;
import reactor.core.publisher.Flux;
import reactor.core.publisher.Mono;
import reactor.core.scheduler.Schedulers;

@SpringBootApplication
@Slf4j
public class ReactorApplication implements ApplicationRunner {
  public static void main(String[] args) {
    SpringApplication.run(ReactorApplication.class, args);
  }

  @Override
  public void run(ApplicationArguments args) throws Exception {
    fluxInMainThread();
    return;
  }

  private void fluxInMainThread() throws InterruptedException {
    Flux.range(1, 6)
        .doOnRequest(n -> log.info("Request {} number", n)) // 注意顺序造成的区别
        .doOnComplete(() -> log.info("Publisher Complete 1"))
        //.publishOn(Schedulers.elastic())
        .map( i -> {
              log.info("Publish {}, {}", Thread.currentThread(), i);
              //return 10 / (i - 3);
              return i;
            }
        )
        .doOnComplete( () -> log.info("Publisher Complete 2"))
        //.subscribeOn(Schedulers.single())
        //.onErrorResume(e -> {
        //  log.error("Exception {}", e.toString());
        //  return Mono.just(-1);
        //})
        //.onErrorReturn(-1)
        .subscribe( i -> log.info("Subscribe {}: {}", Thread.currentThread(), i),
            e -> log.error("error {}", e.toString()),
            () -> log.info("Subscriber Complete")
            //s -> s.request(4)
        );

    Thread.sleep(2000);
  }

}

运行结果

[           main] com.zgpeace.reactor.ReactorApplication   : Request 9223372036854775807 number
[           main] com.zgpeace.reactor.ReactorApplication   : Publish Thread[main,5,main], 1
[           main] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[main,5,main]: 1
[           main] com.zgpeace.reactor.ReactorApplication   : Publish Thread[main,5,main], 2
[           main] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[main,5,main]: 2
[           main] com.zgpeace.reactor.ReactorApplication   : Publish Thread[main,5,main], 3
[           main] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[main,5,main]: 3
[           main] com.zgpeace.reactor.ReactorApplication   : Publish Thread[main,5,main], 4
[           main] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[main,5,main]: 4
[           main] com.zgpeace.reactor.ReactorApplication   : Publish Thread[main,5,main], 5
[           main] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[main,5,main]: 5
[           main] com.zgpeace.reactor.ReactorApplication   : Publish Thread[main,5,main], 6
[           main] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[main,5,main]: 6
[           main] com.zgpeace.reactor.ReactorApplication   : Publisher Complete 1
[           main] com.zgpeace.reactor.ReactorApplication   : Publisher Complete 2
[           main] com.zgpeace.reactor.ReactorApplication   : Subscriber Complete

代码解析：

Request 9223372036854775807 number 表示请求数据为long的最大值，表示不管生产多少，订阅者都要消费掉
发现Publish生产者和Subscribe消费者都在主线程执行
执行结束以后Publisher执行Complete逻辑(可以有多个)，Subscriber执行Complete逻辑

5. 生产者、消费者在不同的线程池

生产者在线程池Schedulers.elastic()执行
消费者在复用线程Schedulers.single()执行

@Override
public void run(ApplicationArguments args) throws Exception {
  //fluxInMainThread();
  fluxInDifferentThread();

  return;
}

private void fluxInDifferentThread() throws InterruptedException {
  Flux.range(1, 6)
      .doOnRequest(n -> log.info("Request {} number", n)) // 注意顺序造成的区别
      .doOnComplete(() -> log.info("Publisher Complete 1"))
      .publishOn(Schedulers.elastic())
      .map( i -> {
            log.info("Publish {}, {}", Thread.currentThread(), i);
            //return 10 / (i - 3);
            return i;
          }
      )
      .doOnComplete( () -> log.info("Publisher Complete 2"))
      .subscribeOn(Schedulers.single())
      //.onErrorResume(e -> {
      //  log.error("Exception {}", e.toString());
      //  return Mono.just(-1);
      //})
      //.onErrorReturn(-1)
      .subscribe( i -> log.info("Subscribe {}: {}", Thread.currentThread(), i),
          e -> log.error("error {}", e.toString()),
          () -> log.info("Subscriber Complete")
          //s -> s.request(4)
      );

  Thread.sleep(2000);
}

运行结果：

[       single-1] com.zgpeace.reactor.ReactorApplication   : Request 256 number
[       single-1] com.zgpeace.reactor.ReactorApplication   : Publisher Complete 1
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 1
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: 1
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 2
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: 2
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 3
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: 3
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 4
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: 4
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 5
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: 5
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 6
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: 6
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publisher Complete 2
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscriber Complete

解析：

[ single-1] Request 256 number 在single申请请求数为256(以前是在主线程申请)
[ elastic-2]后面的生产者消费者都从线程池elastic-2中获取线程。

6. 异常处理返回默认值

代码实现

@Override
public void run(ApplicationArguments args) throws Exception {
  //fluxInMainThread();
  //fluxInDifferentThread();
  fluxOnErrorReturn();

  return;
}

private void fluxOnErrorReturn() throws InterruptedException {
  Flux.range(1, 6)
      .doOnRequest(n -> log.info("Request {} number", n)) // 注意顺序造成的区别
      .doOnComplete(() -> log.info("Publisher Complete 1"))
      .publishOn(Schedulers.elastic())
      .map( i -> {
            log.info("Publish {}, {}", Thread.currentThread(), i);
            return 10 / (i - 3);
            //return i;
          }
      )
      .doOnComplete( () -> log.info("Publisher Complete 2"))
      .subscribeOn(Schedulers.single())
      //.onErrorResume(e -> {
      //  log.error("Exception {}", e.toString());
      //  return Mono.just(-1);
      //})
      .onErrorReturn(-1)
      .subscribe( i -> log.info("Subscribe {}: {}", Thread.currentThread(), i),
          e -> log.error("error {}", e.toString()),
          () -> log.info("Subscriber Complete")
          //s -> s.request(4)
      );

  Thread.sleep(2000);
}

解析：
当有异常发生时候，返回-1， .onErrorReturn(-1)

运行结果：

[       single-1] com.zgpeace.reactor.ReactorApplication   : Request 256 number
[       single-1] com.zgpeace.reactor.ReactorApplication   : Publisher Complete 1
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 1
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: -5
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 2
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: -10
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 3
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: -1
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscriber Complete

解析：

当执行到3的时候，除数为0，发生异常，返回-1.
因为是异常退出，所以没有执行完成，Publisher Complete 2 没有执行。
流程已经结束，订阅者执行完成Subscriber Complete

7. 异常发生处理闭包逻辑

@Override
public void run(ApplicationArguments args) throws Exception {
  //fluxInMainThread();
  //fluxInDifferentThread();
  //fluxOnErrorReturn();
  fluxOnErrorResume();

  return;
}

private void fluxOnErrorResume() throws InterruptedException {
  Flux.range(1, 6)
      .doOnRequest(n -> log.info("Request {} number", n)) // 注意顺序造成的区别
      .doOnComplete(() -> log.info("Publisher Complete 1"))
      .publishOn(Schedulers.elastic())
      .map( i -> {
            log.info("Publish {}, {}", Thread.currentThread(), i);
            return 10 / (i - 3);
            //return i;
          }
      )
      .doOnComplete( () -> log.info("Publisher Complete 2"))
      .subscribeOn(Schedulers.single())
      .onErrorResume(e -> {
        log.error("Exception {}", e.toString());
        return Mono.just(-1);
      })
      //.onErrorReturn(-1)
      .subscribe( i -> log.info("Subscribe {}: {}", Thread.currentThread(), i),
          e -> log.error("error {}", e.toString()),
          () -> log.info("Subscriber Complete")
          //s -> s.request(4)
      );

  Thread.sleep(2000);
}

执行结果：

[       single-1] com.zgpeace.reactor.ReactorApplication   : Request 256 number
[       single-1] com.zgpeace.reactor.ReactorApplication   : Publisher Complete 1
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 1
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: -5
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 2
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: -10
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 3
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Exception java.lang.ArithmeticException: / by zero
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: -1
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscriber Complete

解析：
打印异常信息为除数为0， Exception java.lang.ArithmeticException: / by zero

8. 消费者只消费部分生产者产品

这里把消费者设置为4，而生产还是6。
并把.publishOn(Schedulers.elastic())挪到最前面执行。

@Override
public void run(ApplicationArguments args) throws Exception {
  //fluxInMainThread();
  //fluxInDifferentThread();
  //fluxOnErrorReturn();
  //fluxOnErrorResume();
  fluxWithSubscribe();

  return;
}

private void fluxWithSubscribe() throws InterruptedException {
  Flux.range(1, 6)
      .publishOn(Schedulers.elastic())
      .doOnRequest(n -> log.info("Request {} number", n)) // 注意顺序造成的区别
      .doOnComplete(() -> log.info("Publisher Complete 1"))
      .map( i -> {
            log.info("Publish {}, {}", Thread.currentThread(), i);
            //return 10 / (i - 3);
            return i;
          }
      )
      .doOnComplete( () -> log.info("Publisher Complete 2"))
      .subscribeOn(Schedulers.single())
      .onErrorResume(e -> {
        log.error("Exception {}", e.toString());
        return Mono.just(-1);
      })
      //.onErrorReturn(-1)
      .subscribe( i -> log.info("Subscribe {}: {}", Thread.currentThread(), i),
          e -> log.error("error {}", e.toString()),
          () -> log.info("Subscriber Complete"),
          s -> s.request(4)
      );

  Thread.sleep(2000);
}

运行结果

[       single-1] com.zgpeace.reactor.ReactorApplication   : Request 4 number
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 1
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: 1
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 2
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: 2
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 3
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: 3
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 4
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: 4

解析：

Request 4 number 生产者只会请求4个线程。
这里Publisher Complete 1，Publish Complete 2，Subscriber Complete 都没有打印，因为上面逻辑都在线程[elastic-2,5,main]完毕的时候处理，目前都没有结束。

9. 消费者只消费部分生产者产品(并调换线程位置)

public void run(ApplicationArguments args) throws Exception {
  //fluxInMainThread();
  //fluxInDifferentThread();
  //fluxOnErrorReturn();
  //fluxOnErrorResume();
  //fluxWithSubscribe();
  fluxWithSubscribeAndChangePublishOnThreadOrder();

  return;
}

private void fluxWithSubscribeAndChangePublishOnThreadOrder() throws InterruptedException {
  Flux.range(1, 6)
      .doOnRequest(n -> log.info("Request {} number", n)) // 注意顺序造成的区别
      .publishOn(Schedulers.elastic())
      .doOnComplete(() -> log.info("Publisher Complete 1"))
      .map( i -> {
            log.info("Publish {}, {}", Thread.currentThread(), i);
            //return 10 / (i - 3);
            return i;
          }
      )
      .doOnComplete( () -> log.info("Publisher Complete 2"))
      .subscribeOn(Schedulers.single())
      .onErrorResume(e -> {
        log.error("Exception {}", e.toString());
        return Mono.just(-1);
      })
      //.onErrorReturn(-1)
      .subscribe( i -> log.info("Subscribe {}: {}", Thread.currentThread(), i),
          e -> log.error("error {}", e.toString()),
          () -> log.info("Subscriber Complete"),
          s -> s.request(4)
      );

  Thread.sleep(2000);
}

结果打印：

[       single-1] com.zgpeace.reactor.ReactorApplication   : Request 256 number
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 1
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: 1
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 2
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: 2
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 3
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: 3
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 4
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: 4

解析：
Request 256 number 订阅者线程池被生产者线程池覆盖掉了，所以显示还是256.

10. 消费者只消费部分产品，并且发生异常

运行结果

[       single-1] com.zgpeace.reactor.ReactorApplication   : Request 256 number
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 1
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: -5
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 2
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: -10
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 3
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Exception java.lang.ArithmeticException: / by zero
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: -1
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscriber Complete

解析：
打印Subscriber Complete，因为异常导致订阅者提前结束。

11. 源码解析

还记得说Publisher在subscribe()调用之前啥都没有处理么？

上面的结果显示，线程先打印subscirbe()订阅者的线程[ single-1]，后面才是生产者的线程[ elastic-2]

[       single-1] com.zgpeace.reactor.ReactorApplication   : Request 4 number
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Publish Thread[elastic-2,5,main], 1
[      elastic-2] com.zgpeace.reactor.ReactorApplication   : Subscribe Thread[elastic-2,5,main]: 1

`.map`的实现

public final <V> Flux<V> map(Function<? super T, ? extends V> mapper) {
  if (this instanceof Fuseable) {
	return onAssembly(new FluxMapFuseable<>(this, mapper));
  }
  return onAssembly(new FluxMap<>(this, mapper));
}

protected static <T> Flux<T> onAssembly(Flux<T> source) {
	Function<Publisher, Publisher> hook = Hooks.onEachOperatorHook;
	if(hook != null) {
		source = (Flux<T>) hook.apply(source);
	}
	if (Hooks.GLOBAL_TRACE) {
		AssemblySnapshot stacktrace = new AssemblySnapshot(null, Traces.callSiteSupplierFactory.get());
		source = (Flux<T>) Hooks.addAssemblyInfo(source, stacktrace);
	}
	return source;
}

解析：
Publisher的map仅仅是做了数据的封装，没有任何行动。

`.subscribe()`的调用

public final Disposable subscribe(@Nullable Consumer<? super T> consumer, 
@Nullable Consumer<? super Throwable> errorConsumer, 
@Nullable Runnable completeConsumer, 
@Nullable Consumer<? super Subscription> subscriptionConsumer) {
  return (Disposable)this.subscribeWith(new LambdaSubscriber(consumer, errorConsumer, completeConsumer, subscriptionConsumer, (Context)null));
}

LambdaSubscriber(@Nullable Consumer<? super T> consumer, @Nullable Consumer<? super Throwable> errorConsumer, @Nullable Runnable completeConsumer, @Nullable Consumer<? super Subscription> subscriptionConsumer, @Nullable Context initialContext) {
  this.consumer = consumer;
  this.errorConsumer = errorConsumer;
  this.completeConsumer = completeConsumer;
  this.subscriptionConsumer = subscriptionConsumer;
  this.initialContext = initialContext == null ? Context.empty() : initialContext;
}

public final <E extends Subscriber<? super T>> E subscribeWith(E subscriber) {
  this.subscribe(subscriber);
  return subscriber;
}

解析：上面都是数据组装。
consumer ：过程处理消费者
errorConsumer: 错误处理消费者
completeConsumer：完成处理消费者
subscriptionConsumer：订阅处理消费者
处理逻辑在this.subscribe(subscriber)

public final void subscribe(Subscriber<? super T> actual) {
  CorePublisher publisher = Operators.onLastAssembly(this);
  CoreSubscriber subscriber = Operators.toCoreSubscriber(actual);
  if (publisher instanceof OptimizableOperator) {
    OptimizableOperator operator = (OptimizableOperator)publisher;

    while(true) {
      subscriber = operator.subscribeOrReturn(subscriber);
      if (subscriber == null) {
        return;
      }

      OptimizableOperator newSource = operator.nextOptimizableSource();
      if (newSource == null) {
        publisher = operator.source();
        break;
      }

      operator = newSource;
    }
  }

  publisher.subscribe(subscriber);
}

代码解析：
调用publisher.subscribe(subscriber), 真正触发publisher的执行

继续点击publisher.subscribe(subscriber)为接口

void subscribe(CoreSubscriber<? super T> subscriber);

继续看一个实现接口的实例FluxRepeat.java
在这里插入图片描述
FluxRepeat.java 的实现

void resubscribe() {
	if (WIP.getAndIncrement(this) == 0) {
		do {
			if (isCancelled()) {
				return;
			}

			long c = produced;
			if (c != 0L) {
				produced = 0L;
				produced(c);
			}

			source.subscribe(this);

		} while (WIP.decrementAndGet(this) != 0);
	}
}

才会真正调用source.subscribe(this);

所以在没有subscribe()之前，Publisher的所有逻辑都没有发生。

12. 代码下载

https://github.com/zgpeace/Spring-Boot2.1/tree/master/reactor/ReactorSimple

13. 参考

https://github.com/geektime-geekbang/geektime-spring-family/tree/master/Chapter%205/simple-reactor-demo

程序员易筋

发布了127 篇原创文章 · 获赞 12 · 访问量 2万+

私信关注

易筋SpringBoot 2.1 | 第三十一篇：SpringBoot Reactor响应式编程实战一

1. 说明

2. 核心概念

3. 工程建立

4. 在主线程调用

5. 生产者、消费者在不同的线程池

6. 异常处理返回默认值

7. 异常发生处理闭包逻辑

8. 消费者只消费部分生产者产品

9. 消费者只消费部分生产者产品(并调换线程位置)

10. 消费者只消费部分产品，并且发生异常

11. 源码解析

`.map`的实现

`.subscribe()`的调用

12. 代码下载

13. 参考

猜你喜欢

易筋SpringBoot 2.1 | 第三十一篇：SpringBoot Reactor响应式编程实战一

1. 说明

2. 核心概念

3. 工程建立

4. 在主线程调用

5. 生产者、消费者在不同的线程池

6. 异常处理返回默认值

7. 异常发生处理闭包逻辑

8. 消费者只消费部分生产者产品

9. 消费者只消费部分生产者产品(并调换线程位置)

10. 消费者只消费部分产品，并且发生异常

11. 源码解析

.map的实现

.subscribe()的调用

12. 代码下载

13. 参考

猜你喜欢

`.map`的实现

`.subscribe()`的调用