摘要：接下來(lái)就是會(huì)把任務(wù)提交到隊(duì)列中給線程池調(diào)度處理因?yàn)橹饕P(guān)心的是這個(gè)線程怎么執(zhí)行，異常的拋出和處理，所以我們暫時(shí)不解析多余的邏輯。

今天遇到了一個(gè)bug，現(xiàn)象是，一個(gè)任務(wù)放入線程池中，似乎“沒(méi)有被執(zhí)行”，日志也沒(méi)有打。

經(jīng)過(guò)本地代碼調(diào)試之后，發(fā)現(xiàn)在任務(wù)邏輯的前半段，拋出了NPE，但是代碼外層沒(méi)有try-catch，導(dǎo)致這個(gè)異常被吃掉。

這個(gè)問(wèn)題解決起來(lái)是很簡(jiǎn)單的，外層加個(gè)try-catch就好了，但是這個(gè)異常如果沒(méi)有被catch，線程池內(nèi)部邏輯是怎么處理這個(gè)異常的呢？這個(gè)異常最后會(huì)跑到哪里呢？

帶著疑問(wèn)和好奇心，我研究了一下線程池那一塊的源碼，并且做了以下的總結(jié)。

項(xiàng)目中出問(wèn)題的代碼差不多就是下面這個(gè)樣子

ExecutorService threadPool = Executors.newFixedThreadPool(3);

threadPool.submit(() -> {
    String pennyStr = null;
    Double penny = Double.valueOf(pennyStr);
    ...
})

先進(jìn)到newFixedThreadPool這個(gè)工廠方法中看生成的具體實(shí)現(xiàn)類，發(fā)現(xiàn)是ThreadPoolExecutor

public static ExecutorService newFixedThreadPool(int nThreads) {
        return new ThreadPoolExecutor(nThreads, nThreads,
                                      0L, TimeUnit.MILLISECONDS,
                                      new LinkedBlockingQueue());
    }

再看這個(gè)類的繼承關(guān)系，

再進(jìn)到submit方法，這個(gè)方法在ExecutorService接口中約定，其實(shí)是在AbstractExectorService中實(shí)現(xiàn)，ThreadPoolExecutor并沒(méi)有override這個(gè)方法。

 public Future submit(Runnable task) {
        if (task == null) throw new NullPointerException();
        RunnableFuture ftask = newTaskFor(task, null);
        execute(ftask);
        return ftask;
    }

protected  RunnableFuture newTaskFor(Runnable runnable, T value) {
        return new FutureTask(runnable, value);
    }

對(duì)應(yīng)的FutureTask對(duì)象的構(gòu)造方法

public FutureTask(Runnable runnable, V result) {
        this.callable = Executors.callable(runnable, result);
        this.state = NEW;       // state由volatile 修飾 保證多線程下的可見(jiàn)性
    }

對(duì)應(yīng)Callable 對(duì)象的構(gòu)造方法

public static  Callable callable(Runnable task, T result) {
        if (task == null)
            throw new NullPointerException();
        return new RunnableAdapter(task, result);
    }

對(duì)應(yīng)RunnableAdapter 對(duì)象的構(gòu)造方法

 /**
     * A callable that runs given task and returns given result
     * 一個(gè)能執(zhí)行所給任務(wù)并且返回結(jié)果的Callable對(duì)象
     */
    static final class RunnableAdapter implements Callable {
        final Runnable task;
        final T result;
        RunnableAdapter(Runnable task, T result) {
            this.task = task;
            this.result = result;
        }
        public T call() {
            task.run();
            return result;
        }
    }

總結(jié)上面的，newTaskFor就是把我們提交的Runnable 對(duì)象包裝成了一個(gè)Future。

接下來(lái)就是會(huì)把任務(wù)提交到隊(duì)列中給線程池調(diào)度處理：

public void execute(Runnable command) {
        if (command == null)
            throw new NullPointerException();
    
        int c = ctl.get();
        if (workerCountOf(c) < corePoolSize) {
            if (addWorker(command, true))
                return;
            c = ctl.get();
        }
        if (isRunning(c) && workQueue.offer(command)) {
            int recheck = ctl.get();
            if (! isRunning(recheck) && remove(command))
                reject(command);
            else if (workerCountOf(recheck) == 0)
                addWorker(null, false);
        }
        else if (!addWorker(command, false))
            reject(command);
    }

因?yàn)橹饕P(guān)心的是這個(gè)線程怎么執(zhí)行，異常的拋出和處理，所以我們暫時(shí)不解析多余的邏輯。很容易發(fā)現(xiàn)，如果任務(wù)要被執(zhí)行，肯定是進(jìn)到了addWorker方法當(dāng)中，所以我們?cè)龠M(jìn)去看，鑒于addWorker方法的很長(zhǎng)，不想列太多的代碼，我就摘了關(guān)鍵代碼段：

private boolean addWorker(Runnable firstTask, boolean core) {

   ...
   boolean workerStarted = false;
   boolean workerAdded = false;
   Worker w = null;
   try {
      // 實(shí)例化一個(gè)worker對(duì)象
      w = new Worker(firstTask);
      final Thread t = w.thread;
      if (t != null) {
          final ReentrantLock mainLock = this.mainLock;
          mainLock.lock();
          try {
            
              int rs = runStateOf(ctl.get());

              if (rs < SHUTDOWN ||
                  (rs == SHUTDOWN && firstTask == null)) {
                  if (t.isAlive()) // precheck that t is startable
                      throw new IllegalThreadStateException();
                  workers.add(w);
                  int s = workers.size();
                  if (s > largestPoolSize)
                      largestPoolSize = s;
                  workerAdded = true;
              }
          } finally {
              mainLock.unlock();
          }
          if (workerAdded) {
              // 從Worker對(duì)象的構(gòu)造方法看，當(dāng)這個(gè)thread對(duì)象start之后，
              // 之后實(shí)際上就是調(diào)用Worker對(duì)象的run()
              t.start();
              workerStarted = true;
          }
      }
   } finally {
      if (! workerStarted)
          addWorkerFailed(w);
   }
   return workerStarted;
}

// Worker的構(gòu)造方法
  Worker(Runnable firstTask) {
            setState(-1); // inhibit interrupts until runWorker
            this.firstTask = firstTask;
            this.thread = getThreadFactory().newThread(this);
        }
 

我們?cè)倏催@個(gè)ThreadPoolExecutor的內(nèi)部類Worker對(duì)象：

private final class Worker
        extends AbstractQueuedSynchronizer
        implements Runnable
   {
        ...

        /** Delegates main run loop to outer runWorker  */
        public void run() {
            runWorker(this);
        }

      ...
   }

看來(lái)真正執(zhí)行任務(wù)的是在這個(gè)外部的runWorker當(dāng)中，讓我們?cè)倏纯催@個(gè)方法是怎么消費(fèi)Worker線程的。

final void runWorker(Worker w) {
    Thread wt = Thread.currentThread();
    Runnable task = w.firstTask;
    w.firstTask = null;
    w.unlock(); // allow interrupts
    boolean completedAbruptly = true;
    try {
        while (task != null || (task = getTask()) != null) {
            w.lock();
   
            if ((runStateAtLeast(ctl.get(), STOP) ||
                 (Thread.interrupted() &&
                  runStateAtLeast(ctl.get(), STOP))) &&
                !wt.isInterrupted())
                wt.interrupt();
            try {
                beforeExecute(wt, task);
                Throwable thrown = null;
                // ==== 關(guān)鍵代碼 start ====
                try {
                    // 很簡(jiǎn)潔明了，調(diào)用了任務(wù)的run方法
                    task.run();
                } catch (RuntimeException x) {
                    thrown = x; throw x;
                } catch (Error x) {
                    thrown = x; throw x;
                } catch (Throwable x) {
                    thrown = x; throw new Error(x);
                } finally {
                    afterExecute(task, thrown);
                }
                 // ==== 關(guān)鍵代碼 end ====
            } finally {
                task = null;
                w.completedTasks++;
                w.unlock();
            }
        }
        completedAbruptly = false;
    } finally {
        processWorkerExit(w, completedAbruptly);
    }
}

終于走到底了，可以看到關(guān)鍵代碼中的try-catch block代碼塊中，調(diào)用了本次執(zhí)行任務(wù)的run方法。

// ==== 關(guān)鍵代碼 start ====
try {
  // 很簡(jiǎn)潔明了，調(diào)用了任務(wù)的run方法
  task.run();
} catch (RuntimeException x) {
  thrown = x; throw x;
} catch (Error x) {
  thrown = x; throw x;
} catch (Throwable x) {
  thrown = x; throw new Error(x);
} finally {
  afterExecute(task, thrown);
}
// ==== 關(guān)鍵代碼 end ====

可以看到捕捉了異常之后，會(huì)再向外拋出，只不過(guò)再finally block 中有個(gè)afterExecute()方法，似乎在這里是可以處理這個(gè)異常信息的，進(jìn)去看看

protected void afterExecute(Runnable r, Throwable t) { }

可以看到ThreadPoolExecutor#afterExecute()方法中，是什么都沒(méi)做的，看來(lái)是讓使用者通過(guò)override這個(gè)方法來(lái)定制化任務(wù)執(zhí)行之后的邏輯，其中可以包括異常處理。

那么這個(gè)異常到底是拋到哪里去了呢。我在一個(gè)大佬的文章找到了hotSpot JVM處理線程異常的邏輯，

if (!destroy_vm || JDK_Version::is_jdk12x_version()) {
    // JSR-166: change call from from ThreadGroup.uncaughtException to
    // java.lang.Thread.dispatchUncaughtException
    if (uncaught_exception.not_null()) {
      //如果有未捕獲的異常
      Handle group(this, java_lang_Thread::threadGroup(threadObj()));
      {
        KlassHandle recvrKlass(THREAD, threadObj->klass());
        CallInfo callinfo;
        KlassHandle thread_klass(THREAD, SystemDictionary::Thread_klass());
        /*  
         這里類似一個(gè)方法表，實(shí)際就會(huì)去調(diào)用Thread#dispatchUncaughtException方法
         template(dispatchUncaughtException_name,            "dispatchUncaughtException")                
        */
        LinkResolver::resolve_virtual_call(callinfo, threadObj, recvrKlass, thread_klass,
                                           vmSymbols::dispatchUncaughtException_name(),
                                           vmSymbols::throwable_void_signature(),
                                           KlassHandle(), false, false, THREAD);
        CLEAR_PENDING_EXCEPTION;
        methodHandle method = callinfo.selected_method();
        if (method.not_null()) {
          JavaValue result(T_VOID);
          JavaCalls::call_virtual(&result,
                                  threadObj, thread_klass,
                                  vmSymbols::dispatchUncaughtException_name(),
                                  vmSymbols::throwable_void_signature(),
                                  uncaught_exception,
                                  THREAD);
        } else {
          KlassHandle thread_group(THREAD, SystemDictionary::ThreadGroup_klass());
          JavaValue result(T_VOID);
          JavaCalls::call_virtual(&result,
                                  group, thread_group,
                                  vmSymbols::uncaughtException_name(),
                                  vmSymbols::thread_throwable_void_signature(),
                                  threadObj,           // Arg 1
                                  uncaught_exception,  // Arg 2
                                  THREAD);
        }
        if (HAS_PENDING_EXCEPTION) {
          ResourceMark rm(this);
          jio_fprintf(defaultStream::error_stream(),
                "
Exception: %s thrown from the UncaughtExceptionHandler"
                " in thread "%s"
",
                pending_exception()->klass()->external_name(),
                get_thread_name());
          CLEAR_PENDING_EXCEPTION;
        }
      }
    }

代碼是C寫(xiě)的，有興趣可以去全文，根據(jù)英文注釋能稍微看懂一點(diǎn)

http://hg.openjdk.java.net/jd...

可以看到這里最終會(huì)去調(diào)用Thread#dispatchUncaughtException方法:

/**
     * Dispatch an uncaught exception to the handler. This method is
     * intended to be called only by the JVM.
     */
    private void dispatchUncaughtException(Throwable e) {
        getUncaughtExceptionHandler().uncaughtException(this, e);
    }

/**
 * Called by the Java Virtual Machine when a thread in this
 * thread group stops because of an uncaught exception, and the thread
 * does not have a specific {@link Thread.UncaughtExceptionHandler}
 * installed.
 *
 */
public void uncaughtException(Thread t, Throwable e) {
        if (parent != null) {
            parent.uncaughtException(t, e);
        } else {
            Thread.UncaughtExceptionHandler ueh =
                Thread.getDefaultUncaughtExceptionHandler();
            if (ueh != null) {
                ueh.uncaughtException(t, e);
            } else if (!(e instanceof ThreadDeath)) {
               //可以看到會(huì)打到System.err里面
                System.err.print("Exception in thread ""
                                 + t.getName() + "" ");
                e.printStackTrace(System.err);
            }
        }
    }

jdk的注釋也說(shuō)明的很清楚了，當(dāng)一個(gè)線程拋出了一個(gè)未捕獲的異常，JVM會(huì)去調(diào)用這個(gè)方法。如果當(dāng)前線程沒(méi)有聲明UncaughtExceptionHandler成員變量并且重寫(xiě)uncaughtException方法的時(shí)候，就會(huì)看線程所屬的線程組（如果有線程組的話）有沒(méi)有這個(gè)類，沒(méi)有就會(huì)打到System.err里面。

IBM這篇文章也提倡我們使用ThreadGroup 提供的 uncaughtException 處理程序來(lái)在線程異常終止時(shí)進(jìn)行檢測(cè)。

https://www.ibm.com/developer...

從上述源碼分析中可以看到，對(duì)于本篇的異?！氨怀缘簟钡膯?wèn)題，有以下幾種方法

用try-catch 捕捉，一般都是用這種

線程或者線程組對(duì)象設(shè)置UncaughtExceptionHandler成員變量

  Thread t = new Thread(r);
            t.setUncaughtExceptionHandler(
                (t1, e) -> LOGGER.error(t1 + " throws exception: " + e));
            return t;

override 線程池的afterExecute方法。

本篇雖然是提出問(wèn)題的解決方法，但主旨還是分析源碼，了解了整個(gè)過(guò)程中異常的經(jīng)過(guò)的流程，希望能對(duì)您產(chǎn)生幫助。

https://www.jcp.org/en/jsr/de...

https://www.ibm.com/developer...

http://ifeve.com/%E6%B7%B1%E5...