gpt4 book ai didi

java - 为什么 slf4j 记录器有时会打印出父线程,即使代码应该由子线程执行?

转载 作者:行者123 更新时间:2023-12-02 11:07:35 25 4
gpt4 key购买 nike

考虑这个例子

@Test
public void test2() throws InterruptedException {
int MAX_ENTRIES = 2;
Map cache = new LinkedHashMap(MAX_ENTRIES+1, .75F, true) {
public boolean removeEldestEntry(Map.Entry eldest) {
return size() > MAX_ENTRIES;
}
};

MyBufferService testService = new MyBufferService("test2");

for (int i = 0; i < 100000; i++) {
testService.putBufferTask(Integer.toString(i));
}

TimeUnit.SECONDS.sleep(1);
}


public class MyBufferService {

private ThreadPoolExecutor executor;

private final Logger LOGGER = LoggerFactory.getLogger(MyBufferService.class);
private final Map cache;
final int MAX_ENTRIES = 1;

public MyBufferService(String buffName) {
executor = new ThreadPoolExecutor(1, // corePoolSize
1, // maximumPoolSize
60, TimeUnit.SECONDS, // keepAlive
new LinkedBlockingQueue<>(10000), // workQueue
new ThreadFactoryBuilder().setNameFormat(buffName + "-MyBufferService-thread-%d").build(), // factory
new ThreadPoolExecutor.CallerRunsPolicy() // rejected execution handler
);

this.cache = new LinkedHashMap(MAX_ENTRIES+1, .75F, true) {
public boolean removeEldestEntry(Map.Entry eldest) {
return size() > MAX_ENTRIES;
}
};
}


private class BufferTask implements Runnable {

private final String mystring;
private final Map cache;

BufferTask(String mystring, Map cache) throws NullPointerException {
this.mystring = mystring;
this.cache = cache;
}
@Override
public void run() {
try {
synchronized (this.cache) {
this.cache.put(this.mystring, "hi");
if (this.cache.size() > 0) {
LOGGER.info("this is size {}", this.cache.size() );
}
}
} catch (Throwable t) {

}
}
}

public void putBufferTask(
String mystring) throws RejectedExecutionException, NullPointerException {
executor.submit(new BufferTask(mystring, this.cache));
}

}

并考虑该行的以下输出片段

LOGGER.info("this is size {}", this.cache.size() );


2018-06-13 18:19:46,760 [INFO] [main] c.t.u.w.$MyBufferService - this is size 1
2018-06-13 18:19:46,761 [INFO] [main] c.t.u.w.$MyBufferService - this is size 1
2018-06-13 18:19:46,761 [INFO] [main] c.t.u.w.$MyBufferService - this is size 1
2018-06-13 18:19:46,761 [INFO] [main] c.t.u.w.$MyBufferService - this is size 1
2018-06-13 18:19:46,761 [INFO] [main] c.t.u.w.$MyBufferService - this is size 1
2018-06-13 18:19:46,761 [INFO] [test2-MyBufferService-thread-0] c.t.u.w.$MyBufferService - this is size 1
2018-06-13 18:19:46,761 [INFO] [test2-MyBufferService-thread-0] c.t.u.w.$MyBufferService - this is size 1
2018-06-13 18:19:46,761 [INFO] [test2-MyBufferService-thread-0] c.t.u.w.$MyBufferService - this is size 1
2018-06-13 18:19:46,761 [INFO] [test2-MyBufferService-thread-0] c.t.u.w.$MyBufferService - this is size 1

我们看到 BufferTask 内的代码由 2 个线程运行,其中的 main 线程和子线程 MyBufferService-thread-0 线程

我本以为只有子线程正在执行任务,但看起来父线程也在执行任务。

为什么会发生这种情况?我做错了什么吗?

最佳答案

您设置了拒绝excecution policy让调用者运行任务。这意味着,如果任务在 #execute 中被拒绝,则调用线程(即您的情况下的主线程)将运行该任务,这就是您所看到的。更改该策略或处理 #execute 方法上的异常。

关于java - 为什么 slf4j 记录器有时会打印出父线程,即使代码应该由子线程执行?,我们在Stack Overflow上找到一个类似的问题: https://stackoverflow.com/questions/50848363/

25 4 0
Copyright 2021 - 2024 cfsdn All Rights Reserved 蜀ICP备2022000587号
广告合作:1813099741@qq.com 6ren.com