我正在使用条件变量实现Producer Consumer并遇到了消费者阻塞的问题，基本上消费者线程没有接收生产者生产的最后一批产品，因此永远不会结束。
import java.util.ArrayList;
import java.util.List;
import java.util.concurrent.Callable;
import java.util.concurrent.ExecutionException;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.Future;
import java.util.concurrent.locks.Condition;
import java.util.concurrent.locks.Lock;
import java.util.concurrent.locks.ReentrantLock;
public class ProducerConsumerImpl {
protected List<Integer> buffer = new ArrayList<Integer>();
//the lock on which condition variables are taken
protected volatile Lock lock = new ReentrantLock(true);
// the consumer will signal using this condition variable to the producer to start it's production
protected volatile Condition producerStartProducing = lock.newCondition();
// the producer will use this condition variable to start it's production
protected volatile Condition consumerStartConsuming = lock.newCondition();
class Consumer implements Callable<String> {
int num;
public Consumer(int i) {
this.num = i;
}
/**
* In a loop, take a lock each time, check if there is an item to consume from shared buffer
* If yes, then consume it and loop back
* If no, then wait for the producer to signal you.
* Signals the producer each time it consumes an item.
*/
@Override
public String call() throws Exception {
try {
int i = 0;
while (i < 10) {
while (buffer.isEmpty()) {
consumerStartConsuming.await();
}
lock.lock();
System.out.println("Consumer - " + i);
buffer.remove(buffer.size() - 1);
producerStartProducing.signalAll();
i++;
}
} finally {
lock.unlock();
}
System.out.println("Consumed All");
return "Consumed All";
}
}
class Producer implements Callable<String> {
int num;
public Producer(int i) {
this.num = i;
}
/**
* In a loop, take a lock, produce items in a batch of 3, and then wait till the
* consumer signals you to produce more.
* Signals the consumer each time it produces an item
*/
@Override
public String call() throws Exception {
try {
int i = 0;
while (i < 10) {
lock.lock();
while (buffer.size() > 2) {
producerStartProducing.await();
}
System.out.println("Producer - " + i);
buffer.add(1);
consumerStartConsuming.signalAll();
i++;
}
} finally {
lock.unlock();
}
System.out.println("Produced All");
return "Produced All";
}
}
public static void main(String[] args) throws InterruptedException, ExecutionException {
ExecutorService newFixedThreadPool = Executors.newFixedThreadPool(4);
try {
ProducerConsumerImpl producerConsumerImpl = new ProducerConsumerImpl();
Future<String> p1 = newFixedThreadPool.submit(producerConsumerImpl.new Producer(1));
Future<String> c1 = newFixedThreadPool.submit(producerConsumerImpl.new Consumer(1));
final String string1 = p1.get();
final String string2 = c1.get();
System.out.println(string1 + " --- " + string2);
} finally {
newFixedThreadPool.shutdown();
}
}
}
输出：
Producer - 0
Producer - 1
Producer - 2
Consumer - 0
Consumer - 1
Consumer - 2
Producer - 3
Producer - 4
Producer - 5
Consumer - 3
Consumer - 4
Consumer - 5
Producer - 6
Producer - 7
Producer - 8
Consumer - 6
Consumer - 7
Consumer - 8
Producer - 9
Produced All
生产者代码批量生产三种产品，然后等待消费者使用条件变量信号来提取它们。
类似地，消费者线程获取项目并通知生产者每次从缓冲区获取项目时产生更多。
这只是任意的实现，但是这个代码中的这些操作的任何其他组合都存在问题，如果有人可以指出这段代码出了什么问题并指出了什么，那将会很有用。这里出错了。
答案 0 :(得分：1)
锁定范围不正确所以我已在代码下面修复它。
import java.util.ArrayList;
import java.util.List;
import java.util.concurrent.Callable;
import java.util.concurrent.ExecutionException;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.Future;
import java.util.concurrent.locks.Condition;
import java.util.concurrent.locks.Lock;
import java.util.concurrent.locks.ReentrantLock;
public class ProducerConsumerImpl {
protected List<Integer> buffer = new ArrayList<Integer>();
// the lock on which condition variables are taken
protected volatile Lock lock = new ReentrantLock(true);
// the consumer will signal using this condition variable to the producer to
// start it's production
protected volatile Condition producerStartProducing = lock.newCondition();
// the producer will use this condition variable to start it's production
protected volatile Condition consumerStartConsuming = lock.newCondition();
class Consumer implements Callable<String> {
int num;
public Consumer(int i) {
this.num = i;
}
/**
* In a loop, take a lock each time, check if there is an item to
* consume from shared buffer If yes, then consume it and loop back If
* no, then wait for the producer to signal you. Signals the producer
* each time it consumes an item.
*/
@Override
public String call() throws Exception {
int i = 0;
while (i < 10) {
try {
lock.lock();
while (buffer.isEmpty()) {
consumerStartConsuming.await();
}
System.out.println("Consumer - " + i);
buffer.remove(buffer.size() - 1);
producerStartProducing.signalAll();
i++;
} finally {
lock.unlock();
}
}
System.out.println("Consumed All");
return "Consumed All";
}
}
class Producer implements Callable<String> {
int num;
public Producer(int i) {
this.num = i;
}
/**
* In a loop, take a lock, produce items in a batch of 3, and then wait
* till the consumer signals you to produce more. Signals the consumer
* each time it produces an item
*/
@Override
public String call() throws Exception {
int i = 0;
while (i < 10) {
try {
lock.lock();
while (buffer.size() > 2) {
producerStartProducing.await();
}
System.out.println("Producer - " + i);
buffer.add(1);
consumerStartConsuming.signalAll();
i++;
} finally {
lock.unlock();
}
}
System.out.println("Produced All");
return "Produced All";
}
}
public static void main(String[] args) throws InterruptedException, ExecutionException {
ExecutorService newFixedThreadPool = Executors.newFixedThreadPool(4);
try {
ProducerConsumerImpl producerConsumerImpl = new ProducerConsumerImpl();
Future<String> p1 = newFixedThreadPool.submit(producerConsumerImpl.new Producer(1));
Future<String> c1 = newFixedThreadPool.submit(producerConsumerImpl.new Consumer(1));
final String string1 = p1.get();
final String string2 = c1.get();
System.out.println(string1 + " --- " + string2);
} finally {
newFixedThreadPool.shutdown();
}
}
}