ConcurrentMap线程安全的正确用法

import java.util.HashMap;import java.util.Map;import java.util.concurrent.ConcurrentHashMap;import java.util.concurrent.ConcurrentMap;import java.util.concurrent.ExecutionException;public class ConcurrentMapWithMap {          private static Map<String, Long> mapWordCounts = new HashMap<>();      private static ConcurrentMap<String, Long> concurrentMapWordCounts = new ConcurrentHashMap<>();      public static int count=0;            public static long mapIncrease(String word) {          Long oldValue = mapWordCounts.get(word);          Long newValue = (oldValue == null) ? 1L : oldValue + 1;          mapWordCounts.put(word, newValue);          return newValue;      }                   public static long ConcurrentMapIncrease(String word) {          Long oldValue, newValue;          while (true) {              oldValue = concurrentMapWordCounts.get(word);              if (oldValue == null) {                  newValue = 1L;                  //已经有key了就返回放入的值，否则返回空                if (concurrentMapWordCounts.putIfAbsent(word, newValue) == null) {                      break;                  }              } else {                  newValue = oldValue + 1;                  //值替换，每次替换时都会比较上面拿到oldValue是否就是当前map里面的值，是才替换，否则继续获取                if (concurrentMapWordCounts.replace(word, oldValue, newValue)) {                      break;                  }              }          }          return newValue;    }            public static void mapWordCount() throws InterruptedException, ExecutionException {          new Thread(new Runnable(){                public void run() {                  int count=0;                  while(count++<10000)                     ConcurrentMapWithMap.mapIncrease("work");              }          }).start();          new Thread(new Runnable(){                public void run() {                  int count=0;                  while(count++<10000)   ;                    ConcurrentMapWithMap.mapIncrease("work");              }          }).start();          new Thread(new Runnable(){                public void run() {                  int count=0;                  while(count++<10000)                      ConcurrentMapWithMap.mapIncrease("work");              }          }).start();          new Thread(new Runnable(){                public void run() {                  int count=0;                  while(count++<10000)                       ConcurrentMapWithMap.mapIncrease("work");              }          }).start();      }                               public static void concurrentWordCount() throws InterruptedException, ExecutionException {          new Thread(new Runnable(){                public void run() {                  int count=0;                  while(count++<10000)                       ConcurrentMapWithMap.ConcurrentMapIncrease("work");              }          }).start();          new Thread(new Runnable(){                public void run() {                  int count=0;                  while(count++<10000)                       ConcurrentMapWithMap.ConcurrentMapIncrease("work");              }          }).start();          new Thread(new Runnable(){                public void run() {                  int count=0;                  while(count++<10000)                       ConcurrentMapWithMap.ConcurrentMapIncrease("work");              }          }).start();          new Thread(new Runnable(){                public void run() {                  int count=0;                  while(count++<10000)                       ConcurrentMapWithMap.ConcurrentMapIncrease("work");              }          }).start();      }          public static void main(String[] args) throws InterruptedException, ExecutionException {          ConcurrentMapWithMap.mapWordCount();          Thread.sleep(10000);          System.out.println("final count map"+ConcurrentMapWithMap.mapWordCounts.get("work"));//多线程累加，每次都少于40000，故线程不安全          ConcurrentMapWithMap.concurrentWordCount();          Thread.sleep(10000);          System.out.println("final count concurrentMap"+ConcurrentMapWithMap.concurrentMapWordCounts.get("work"));//多线程累加，每次都是40000      }    }