在《阿里巴巴java开发手册》中指出了线程资源必须通过线程池提供,不允许在应用中自行显示的创建线程,这样一方面是线程的创建更加规范,可以合理控制开辟线程的数量;另一方面线程的细节管理交给线程池处理,优化了资源的开销。而线程池不允许使用Executors去创建,而要通过ThreadPoolExecutor方式,这一方面是由于jdk中Executor框架虽然提供了如newFixedThreadPool()、newSingleThreadExecutor()、newCachedThreadPool()等创建线程池的方法,但都有其局限性,不够灵活;另外由于前面几种方法内部也是通过ThreadPoolExecutor方式实现,使用ThreadPoolExecutor有助于大家明确线程池的运行规则,创建符合自己的业务场景需要的线程池,避免资源耗尽的风险。
下面我们就对ThreadPoolExecutor的使用方法进行一个详细的概述
public ThreadPoolExecutor(int corePoolSize,
int maximumPoolSize,
long keepAliveTime,
TimeUnit unit,
BlockingQueue<Runnable> workQueue,
ThreadFactory threadFactory,
RejectedExecutionHandler handler) {
if (corePoolSize < 0 ||
maximumPoolSize <= 0 ||
maximumPoolSize < corePoolSize ||
keepAliveTime < 0)
throw new IllegalArgumentException();
if (workQueue == null || threadFactory == null || handler == null)
throw new NullPointerException();
this.acc = System.getSecurityManager() == null ?
null :
AccessController.getContext();
this.corePoolSize = corePoolSize;
this.maximumPoolSize = maximumPoolSize;
this.workQueue = workQueue;
this.keepAliveTime = unit.toNanos(keepAliveTime);
this.threadFactory = threadFactory;
this.handler = handler;
}