先展示一下jdk8中ThreadPoolExecutor构造方法的源码
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;
}
corePoolSize //核心线程数量,会一直存在,除非allowCoreThreadTimeOut设置为true
maximumPoolSize //线程池允许的最大线程池数量
keepAliveTime //线程数量超过corePoolSize,空闲线程的最大超时时间
unit //超时时间的单位
workQueue //工作队列,保存未执行的Runnable 任务
threadFactory //创建线程的工厂类
handler //当线程已满,工作队列也满了的时候,会被调用。被用来实现各种拒绝策略。