前言
网上冲浪时刷到线程池的文章，想想看自己好像还没在实际场景中设置过线程名称，小小研究一下。
研究过程
默认命名
创建的线程都会有自己的名字，如果不设置，程序会给线程默认的名字，如Thread-0
Thread t = new Thread(() -> {System.out.println(Thread.currentThread().getName());
});
t.start();
// Thread-0
设置线程名称，应当有个理由。现在一个项目中有订单相关线程池、有付款相关线程池，给线程命名，可以容易区分线程种类。如Thread-order-0、Thread-fund-0
Thread内部实现原理
首先我们打开Thread类，属性name就是线程的名称。
public
class Thread implements Runnable {/* Make sure registerNatives is the first thing <clinit> does. */private static native void registerNatives();static {registerNatives();}// 线程名字private volatile String name;private int priority;private Thread threadQ;private long eetop;// ...省略}
测试：
抽象实现
首先实现线程工厂构造器，主要构造线程工厂对象
@Getter
public class ThreadFactoryBuilder {private String nameFormat;public ThreadFactoryBuilder setNameFormat(String nameFormat) {this.nameFormat = nameFormat;return this;}public ThreadFactory build(){return buildThreadFactory(this);}public static ThreadFactory buildThreadFactory(ThreadFactoryBuilder builder) {final AtomicInteger counter = new AtomicInteger(0);return r -> {Thread t = new Thread(r);// 这里设置了线程池名称，使用counter区分不同线程t.setName(String.format(builder.getNameFormat(), counter.getAndIncrement()));return t;};}
}
查看源码ExecutorService源码，发现预留了线程工厂的入参
// ExecutorService newFixedThreadPool 预留了ThreadFactory
public static ExecutorService newFixedThreadPool(int nThreads, ThreadFactory threadFactory) {return new ThreadPoolExecutor(nThreads, nThreads,0L, TimeUnit.MILLISECONDS,new LinkedBlockingQueue<Runnable>(),threadFactory);
}
测试代码正确性：
ThreadFactory threadFactory = new ThreadFactoryBuilder().setNameFormat("测试命名线程-%d").build();
ExecutorService executor = Executors.newFixedThreadPool(5, threadFactory);for (int i = 0; i < 10; i++) {executor.execute(() -> {try {sleep(1000);} catch (InterruptedException e) {throw new RuntimeException(e);}System.out.println(Thread.currentThread().getName());});
}// 测试命名线程-3
// 测试命名线程-2
// 测试命名线程-1
// 测试命名线程-0
// 测试命名线程-4
// 测试命名线程-0
// 测试命名线程-2
// 测试命名线程-1
// 测试命名线程-3
// 测试命名线程-4
