設置<Tuple>不能正常增加值

Question

我工作的一個開源Redis的樣內存緩存API：http://bit.ly/XVfpRX

誰能幫我分析一下，爲什麼這個代碼是爲String field和String value變量得到一個錯誤的值：

public Long hset(String key, String field, String value) { 
    System.out.println("HSET key="+key+" field=" + field + " value="+value); 
    try { 
     boolean exist = exists(key); 
     if(exist){ 
      Set<Tuple<String,String>> hash = (Set<Tuple<String,String>>) memget(key); 
      Iterator<Tuple<String, String>> it = hash.iterator(); 
      while (it.hasNext()){ 
       Tuple<String, String> t = it.next(); 
       if (t.getFirst().equals(field)){ 
        System.out.println("HSET Removing field=" + t.getFirst() + " value="+t.getSecond()); 
        hash.remove(t); 
        hash.add(new Tuple<String, String>(field, value)); 
       } 
      } 
      memset(key, hash); 
      return 0L; 
     } else { 
      Set<Tuple<String,String>> hash = new HashSet<Tuple<String,String>>(); 
      System.out.println("Adding new tuple key="+key+" field=" + field + " value="+value); 
      hash.add(new Tuple<String, String>(field, value)); 
      memset(key, hash); 
      return 1L; 
     } 
    } catch(Exception e) { // ClassCastException and NPE 
     // What to do when CCE is encountered? 
     Set<Tuple<String,String>> hash = new HashSet<Tuple<String,String>>(); 
     hash.add(new Tuple<String, String>(field, value)); 
     memset(key, hash);    
    } 
    return 1L; 
} 

下面是測試日誌：

Start hset tests 
HSET key=null field=null value=null 
Adding new tuple key=null field=null value=null 
MEMSET key=null value=[Tuple [first=null, second=null]] 
HSET key=foo field=foo value=bar 
Adding new tuple key=foo field=foo value=bar 
MEMSET key=foo value=[Tuple [first=foo, second=bar]] 
HSET key=foo field=bar value=foobar 
MEMSET key=foo value=[Tuple [first=foo, second=bar]] 
HSET key=user:1 field=uname value=foo 
Adding new tuple key=user:1 field=uname value=foo 
MEMSET key=user:1 value=[Tuple [first=uname, second=foo]] 
HSET key=user:1 field=fname value=Bar 
MEMSET key=user:1 value=[Tuple [first=uname, second=foo]] 
HSET key=user:1 field=uname value=bar 
HSET Removing field=uname value=foo 
MEMSET key=user:1 value=[Tuple [first=uname, second=bar]] 

而且有實際的JUnit測試：

@Test 
public void testHset() { 
    System.out.println("Start hset tests"); 
    Long ret = lingo.hset(key, null, value); 
    assertEquals(1L, ret.longValue()); 
    ret = lingo.hset("foo", "foo", "bar"); 
    assertEquals(1L, ret.longValue()); 
    ret = lingo.hset("foo", "bar", "foobar"); 
    assertEquals(0L, ret.longValue()); 

    ret = lingo.hset("user:1", "uname", "foo"); 
    assertEquals(1L, ret.longValue()); 
    ret = lingo.hset("user:1", "fname", "Bar"); 
    assertEquals(0L, ret.longValue()); 
    ret = lingo.hset("user:1", "uname", "bar"); 
    assertEquals(0L, ret.longValue()); 

    String username = lingo.hget("user:1", "uname"); 
    String firstname = lingo.hget("user:1", "fname"); 
    assertEquals("bar", username); 
    assertEquals("Bar", firstname); 
    System.out.println("End hset tests"); 
} 

問題區域是這樣的：

HSET key=foo field=bar value=foobar 
MEMSET key=foo value=[Tuple [first=foo, second=bar]] 

和此：

HSET key=user:1 field=fname value=Bar 
MEMSET key=user:1 value=[Tuple [first=uname, second=foo]] 

的字段和值傳播到memset的（這是G​​AE內存緩存鍵值包裝方法放）是錯誤的。

The Tuple code can be found here。

更新（新增hget法）：

public String hget(String key, String field) { 
    try { 
     boolean exist = exists(key); 
     if(exist){ 
      Set<Tuple<String,String>> hash = (Set<Tuple<String,String>>) memget(key); 
      Iterator<Tuple<String, String>> it = hash.iterator(); 
      while (it.hasNext()){ 
       Tuple<String, String> t = it.next(); 
       if (t.getFirst().equals(field)){ 
        return t.getSecond(); 
       } 
      } 
     } 
    } catch(Exception e) { // ClassCastException and NPE 
     // What to do when CCE is encountered? 
    } 
    return "nil"; 
} 

Answer 1

你寫Tuple類？

我的第一個想法是，你沒有正確覆蓋equals和hashCode。 Joshua Bloch的Effective Java告訴你如何。