HashMap夺命14问，你能坚持到第几问？

final Node<K,V>[] resize() {
        //oldTab：引用扩容前的哈希表
        Node<K,V>[] oldTab = table;
        //oldCap：表示扩容前的table数组的长度
        int oldCap = (oldTab == null) ? 0 : oldTab.length;
        //获得旧哈希表的扩容阈值
        int oldThr = threshold;
        //newCap:扩容之后table数组大小
        //newThr:扩容之后下次触发扩容的条件
        int newCap, newThr = 0;
        //条件成立说明hashMap中的散列表已经初始化过了，是一次正常扩容
        if (oldCap > 0) {
            //判断旧的容量是否大于等于最大容量，如果是，则无法扩容，并且设置扩容条件为int最大值，
            //这种情况属于非常少数的情况
            if (oldCap >= MAXIMUM_CAPACITY) {
                threshold = Integer.MAX_VALUE;
                return oldTab;
            }//设置newCap新容量为oldCap旧容量的二倍（<<1）,并且<最大容量，而且>=16，则新阈值等于旧阈值的两倍
            else if ((newCap = oldCap << 1) < MAXIMUM_CAPACITY &&
                     oldCap >= DEFAULT_INITIAL_CAPACITY)
                newThr = oldThr << 1; // double threshold
        }
        //如果oldCap=0并且边界值大于0，说明散列表是null，但此时oldThr>0
        //说明此时hashMap的创建是通过指定的构造方法创建的,新容量直接等于阈值
        //1.new HashMap(intitCap,loadFactor)
        //2.new HashMap(initCap)
        //3.new HashMap(map)
        else if (oldThr > 0) // initial capacity was placed in threshold
            newCap = oldThr;
        //这种情况下oldThr=0;oldCap=0，说明没经过初始化，创建hashMap
        //的时候是通过new HashMap()的方式创建的
        else { // zero initial threshold signifies using defaults
            newCap = DEFAULT_INITIAL_CAPACITY;
            newThr = (int)(DEFAULT_LOAD_FACTOR * DEFAULT_INITIAL_CAPACITY);
        }
        //newThr为0时，通过newCap和loadFactor计算出一个newThr
        if (newThr == 0) {
            //容量*0.75
            float ft = (float)newCap * loadFactor;
            newThr = (newCap < MAXIMUM_CAPACITY && ft < (float)MAXIMUM_CAPACITY ?
                      (int)ft : Integer.MAX_VALUE);
        }
        threshold = newThr;
        @SuppressWarnings({"rawtypes","unchecked"})
                //根据上面计算出的结果创建一个更长更大的数组
            Node<K,V>[] newTab = (Node<K,V>[])new Node[newCap];
        //将table指向新创建的数组
        table = newTab;
        //本次扩容之前table不为null
        if (oldTab != null) {
            //对数组中的元素进行遍历
            for (int j = 0; j < oldCap; ++j) {
                //设置e为当前node节点
                Node<K,V> e;
                //当前桶位数据不为空，但不能知道里面是单个元素，还是链表或红黑树，
                //e = oldTab[j]，先用e记录下当前元素
                if ((e = oldTab[j]) != null) {
                    //将老数组j桶位置为空，方便回收
                    oldTab[j] = null;
                    //如果e节点不存在下一个节点，说明e是单个元素，则直接放置在新数组的桶位
                    if (e.next == null)
                        newTab[e.hash & (newCap - 1)] = e;
                    //如果e是树节点，证明该节点处于红黑树中
                    else if (e instanceof TreeNode)
                        ((TreeNode<K,V>)e).split(this, newTab, j, oldCap);
                    //e为链表节点，则对链表进行遍历
                    else { // preserve order
                        //低位链表：存放在扩容之后的数组的下标位置，与当前数组下标位置一致
                        //loHead：低位链表头节点
                        //loTail低位链表尾节点
                        Node<K,V> loHead = null, loTail = null;
                        //高位链表，存放扩容之后的数组的下标位置，=原索引+扩容之前数组容量
                        //hiHead:高位链表头节点
                        //hiTail:高位链表尾节点
                        Node<K,V> hiHead = null, hiTail = null;
                        Node<K,V> next;
                        do {
                            next = e.next;
                            //oldCap为16:10000，与e.hsah做&运算可以得到高位为1还是0
                            //高位为0，放在低位链表
                            if ((e.hash & oldCap) == 0) {
                                if (loTail == null)
                                    //loHead指向e
                                    loHead = e;
                                else
                                    loTail.next = e;
                                loTail = e;
                            }
                            //高位为1，放在高位链表
                            else {
                                if (hiTail == null)
                                    hiHead = e;
                                else
                                    hiTail.next = e;
                                hiTail = e;
                            }
                        } while ((e = next) != null);
                        //低位链表已成，将头节点loHead指向在原位
                        if (loTail != null) {
                            loTail.next = null;
                            newTab[j] = loHead;
                        }
                        //高位链表已成，将头节点指向新索引
                        if (hiTail != null) {
                            hiTail.next = null;
                            newTab[j + oldCap] = hiHead;
                        }
                    }
                }
            }
        }
        return newTab;
    }

 * Because TreeNodes are about twice the size of regular nodes, we
 * use them only when bins contain enough nodes to warrant use
 * (see TREEIFY_THRESHOLD). And when they become too small (due to
 * removal or resizing) they are converted back to plain bins. In
 * usages with well-distributed user hashCodes, tree bins are
 * rarely used. Ideally, under random hashCodes, the frequency of
 * nodes in bins follows a Poisson distribution
 * (http://en.wikipedia.org/wiki/Poisson_distribution) with a
 * parameter of about 0.5 on average for the default resizing
 * threshold of 0.75, although with a large variance because of
 * resizing granularity. Ignoring variance, the expected
 * occurrences of list size k are (exp(-0.5) * pow(0.5, k) /
 * factorial(k)).

* 0: 0.60653066
 * 1: 0.30326533
 * 2: 0.07581633
 * 3: 0.01263606
 * 4: 0.00157952
 * 5: 0.00015795
 * 6: 0.00001316
 * 7: 0.00000094
 * 8: 0.00000006
 * more: less than 1 in ten million