Java CompletableFuture supply Async有时在主线程上运行

我一直在试验Java8 CompletableFutures。我的理解是，调用CompletableFuture.supplyAsync（供应商-供应商，执行者-执行者）将始终在传入执行者提供的线程中运行作业，但我注意到，当我传递给它的供应商相当“简单”时，它有时会在主线程上运行。我的测试代码是：

import java.util.concurrent.CompletableFuture;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.atomic.AtomicInteger;

public class SyncTest {

    public static void main(String[] args) {
        ExecutorService pool = Executors.newFixedThreadPool(5);

        CompletableFuture<?>[] cfs = new CompletableFuture<?>[10];
        AtomicInteger mainCount = new AtomicInteger();
        for (int i = 0; i < cfs.length; i++) {
            int index = i;
            CompletableFuture<Integer> cf = 
CompletableFuture.supplyAsync(() -> {
                return index;
            }, pool).thenApply(j -> {
                if (Thread.currentThread().getName().equals("main")) {    
                    mainCount.incrementAndGet();
                }

                System.out.println(Thread.currentThread().getName() + ": " + index + ": doing a heavy computation");
                try {
                    Thread.sleep(10);
                } catch (InterruptedException e) {
                    e.printStackTrace();
                }
                return j;
            });
            cfs[i] = cf;
        }

        System.out.println(Thread.currentThread().getName() + " doing other stuff");
        try {
            Thread.sleep(10);
        } catch (InterruptedException e) {
            e.printStackTrace();
        }

        CompletableFuture.allOf(cfs).join();
        pool.shutdown();
        System.out.println("Jobs on main: " + mainCount.get());
    }
}

我知道这是一个相当简单的例子，还有其他一些方法可以实现CompletableFutures，比如SupplySync和completedFuture。我更好奇的是，其中一些任务是如何在主线程上执行的。

打印“正在进行大量计算”的任务通过

然后apply调用，这意味着您没有指定运行此任务的执行器，并且系统可以自由使用任何执行器，包括当前线程

如果要在预定义的执行器上执行此作业，请使用
然后使用applyasync
，无论是否使用第二个参数-executor。
检查此问题的答案：可能重复
main: 0: doing a heavy computation
main: 1: doing a heavy computation
pool-1-thread-3: 2: doing a heavy computation
pool-1-thread-4: 3: doing a heavy computation
main: 4: doing a heavy computation
main doing other stuff
pool-1-thread-4: 9: doing a heavy computation
pool-1-thread-5: 8: doing a heavy computation
pool-1-thread-2: 7: doing a heavy computation
pool-1-thread-1: 6: doing a heavy computation
pool-1-thread-3: 5: doing a heavy computation
Jobs on main: 3