備戰-Java 容器
玉階生白露，夜久侵羅襪。
簡介：備戰-Java 容器
一、概述
容器主要包括 Collection 和 Map 兩種，Collection 儲存著物件的集合，而 Map 儲存著key-value 鍵值對（兩個物件）的對映表。
Collection
1. Set
TreeSet：基於紅黑樹實現，支援有序性操作，例如根據一個範圍查詢元素的操作。但是查詢效率不如 HashSet，HashSet 查詢的時間複雜度為 O(1)，TreeSet 則為 O(logN)。
HashSet：基於雜湊表實現，支援快速查詢，但不支援有序性操作。並且失去了元素的插入順序資訊，也就是說使用 Iterator 遍歷 HashSet 得到的結果是不確定的。
LinkedHashSet：具有 HashSet 的查詢效率，並且內部使用雙向連結串列維護元素的插入順序。
2. List
ArrayList：基於動態陣列實現，支援隨機訪問。
Vector：和 ArrayList 類似，但它是執行緒安全的。
LinkedList：基於雙向連結串列實現，只能順序訪問，但是可以快速地在連結串列中間插入和刪除元素。不僅如此，LinkedList 還可以用作棧、佇列和雙向佇列。
3. Queue
LinkedList：可以用它來實現雙向佇列。
PriorityQueue：基於堆結構實現，可以用它來實現優先佇列。
Map
TreeMap：基於紅黑樹實現。
HashMap：基於雜湊表實現。
HashTable：和 HashMap 類似，但它是執行緒安全的，這意味著同一時刻多個執行緒同時寫入 HashTable 不會導致資料不一致。它是遺留類，不應該去使用它，而是使用 ConcurrentHashMap 來支援執行緒安全，ConcurrentHashMap 的效率會更高，因為 ConcurrentHashMap 引入了分段鎖。
LinkedHashMap：使用雙向連結串列來維護元素的順序，順序為插入順序或者最近最少使用（LRU）順序。(LRU演算法是Least Recently Used的縮寫，即最近最少使用)
二、容器中的設計模式
迭代器模式
Collection 繼承了 Iterable 介面，其中的 iterator() 方法能夠產生一個 Iterator 物件，通過這個物件就可以迭代遍歷 Collection 中的元素。
從 JDK 1.5 之後可以使用 foreach 方法來遍歷實現了 Iterable 介面的聚合物件。
1 List<String> list = new ArrayList<>();
2 list.add("a");
3 list.add("b");
4 for (String item : list) {
5 System.out.println(item);
6 }
介面卡模式
java.util.Arrays.asList() 可以把陣列型別轉換為 List 型別。
1 @SafeVarargs
2 public static <T> List<T> asList(T... a)
值得注意的是 asList() 的引數為泛型的變長引數，不能使用基本型別陣列作為引數，只能使用相應的包裝型別陣列。
1 Integer[] arr = {1, 2, 3};
2 List list = Arrays.asList(arr);
也可以使用以下方式呼叫 asList()：
List list = Arrays.asList(1, 2, 3);
三、原始碼分析
如果沒有特別說明，以下原始碼分析基於 JDK 1.8。
在 IDEA 中 雙擊 shift 鍵調出 Search EveryWhere，查詢原始碼檔案，找到之後就可以閱讀原始碼。
ArrayList
1. 概覽
因為 ArrayList 是基於陣列實現的，所以支援快速隨機訪問。RandomAccess 介面標識著該類支援快速隨機訪問，其預設陣列大小為10
1 /*
2 * Copyright (c) 1997, 2017, Oracle and/or its affiliates. All rights reserved.
3 * ORACLE PROPRIETARY/CONFIDENTIAL. Use is subject to license terms.
4 *
5 *
6 *
7 *
8 *
9 *
10 *
11 *
12 *
13 *
14 *
15 *
16 *
17 *
18 *
19 *
20 *
21 *
22 *
23 *
24 */
25
26 package java.util;
27
28 import java.util.function.Consumer;
29 import java.util.function.Predicate;
30 import java.util.function.UnaryOperator;
31 import sun.misc.SharedSecrets;
32
33 /**
34 * Resizable-array implementation of the <tt>List</tt> interface. Implements
35 * all optional list operations, and permits all elements, including
36 * <tt>null</tt>. In addition to implementing the <tt>List</tt> interface,
37 * this class provides methods to manipulate the size of the array that is
38 * used internally to store the list. (This class is roughly equivalent to
39 * <tt>Vector</tt>, except that it is unsynchronized.)
40 *
41 * <p>The <tt>size</tt>, <tt>isEmpty</tt>, <tt>get</tt>, <tt>set</tt>,
42 * <tt>iterator</tt>, and <tt>listIterator</tt> operations run in constant
43 * time. The <tt>add</tt> operation runs in <i>amortized constant time</i>,
44 * that is, adding n elements requires O(n) time. All of the other operations
45 * run in linear time (roughly speaking). The constant factor is low compared
46 * to that for the <tt>LinkedList</tt> implementation.
47 *
48 * <p>Each <tt>ArrayList</tt> instance has a <i>capacity</i>. The capacity is
49 * the size of the array used to store the elements in the list. It is always
50 * at least as large as the list size. As elements are added to an ArrayList,
51 * its capacity grows automatically. The details of the growth policy are not
52 * specified beyond the fact that adding an element has constant amortized
53 * time cost.
54 *
55 * <p>An application can increase the capacity of an <tt>ArrayList</tt> instance
56 * before adding a large number of elements using the <tt>ensureCapacity</tt>
57 * operation. This may reduce the amount of incremental reallocation.
58 *
59 * <p><strong>Note that this implementation is not synchronized.</strong>
60 * If multiple threads access an <tt>ArrayList</tt> instance concurrently,
61 * and at least one of the threads modifies the list structurally, it
62 * <i>must</i> be synchronized externally. (A structural modification is
63 * any operation that adds or deletes one or more elements, or explicitly
64 * resizes the backing array; merely setting the value of an element is not
65 * a structural modification.) This is typically accomplished by
66 * synchronizing on some object that naturally encapsulates the list.
67 *
68 * If no such object exists, the list should be "wrapped" using the
69 * {@link Collections#synchronizedList Collections.synchronizedList}
70 * method. This is best done at creation time, to prevent accidental
71 * unsynchronized access to the list:<pre>
72 * List list = Collections.synchronizedList(new ArrayList(...));</pre>
73 *
74 * <p><a name="fail-fast">
75 * The iterators returned by this class's {@link #iterator() iterator} and
76 * {@link #listIterator(int) listIterator} methods are <em>fail-fast</em>:</a>
77 * if the list is structurally modified at any time after the iterator is
78 * created, in any way except through the iterator's own
79 * {@link ListIterator#remove() remove} or
80 * {@link ListIterator#add(Object) add} methods, the iterator will throw a
81 * {@link ConcurrentModificationException}. Thus, in the face of
82 * concurrent modification, the iterator fails quickly and cleanly, rather
83 * than risking arbitrary, non-deterministic behavior at an undetermined
84 * time in the future.
85 *
86 * <p>Note that the fail-fast behavior of an iterator cannot be guaranteed
87 * as it is, generally speaking, impossible to make any hard guarantees in the
88 * presence of unsynchronized concurrent modification. Fail-fast iterators
89 * throw {@code ConcurrentModificationException} on a best-effort basis.
90 * Therefore, it would be wrong to write a program that depended on this
91 * exception for its correctness: <i>the fail-fast behavior of iterators
92 * should be used only to detect bugs.</i>
93 *
94 * <p>This class is a member of the
95 * <a href="{@docRoot}/../technotes/guides/collections/index.html">
96 * Java Collections Framework</a>.
97 *
98 * @author Josh Bloch
99 * @author Neal Gafter
100 * @see Collection
101 * @see List
102 * @see LinkedList
103 * @see Vector
104 * @since 1.2
105 */
106
107 public class ArrayList<E> extends AbstractList<E>
108 implements List<E>, RandomAccess, Cloneable, java.io.Serializable
109 {
110 private static final long serialVersionUID = 8683452581122892189L;
111
112 /**
113 * Default initial capacity.
114 */
115 private static final int DEFAULT_CAPACITY = 10;
116
117 /**
118 * Shared empty array instance used for empty instances.
119 */
120 private static final Object[] EMPTY_ELEMENTDATA = {};
121
122 /**
123 * Shared empty array instance used for default sized empty instances. We
124 * distinguish this from EMPTY_ELEMENTDATA to know how much to inflate when
125 * first element is added.
126 */
127 private static final Object[] DEFAULTCAPACITY_EMPTY_ELEMENTDATA = {};
128
129 /**
130 * The array buffer into which the elements of the ArrayList are stored.
131 * The capacity of the ArrayList is the length of this array buffer. Any
132 * empty ArrayList with elementData == DEFAULTCAPACITY_EMPTY_ELEMENTDATA
133 * will be expanded to DEFAULT_CAPACITY when the first element is added.
134 */
135 transient Object[] elementData; // non-private to simplify nested class access
136
137 /**
138 * The size of the ArrayList (the number of elements it contains).
139 *
140 * @serial
141 */
142 private int size;
143
144 /**
145 * Constructs an empty list with the specified initial capacity.
146 *
147 * @param initialCapacity the initial capacity of the list
148 * @throws IllegalArgumentException if the specified initial capacity
149 * is negative
150 */
151 public ArrayList(int initialCapacity) {
152 if (initialCapacity > 0) {
153 this.elementData = new Object[initialCapacity];
154 } else if (initialCapacity == 0) {
155 this.elementData = EMPTY_ELEMENTDATA;
156 } else {
157 throw new IllegalArgumentException("Illegal Capacity: "+
158 initialCapacity);
159 }
160 }
161
162 /**
163 * Constructs an empty list with an initial capacity of ten.
164 */
165 public ArrayList() {
166 this.elementData = DEFAULTCAPACITY_EMPTY_ELEMENTDATA;
167 }
168
169 /**
170 * Constructs a list containing the elements of the specified
171 * collection, in the order they are returned by the collection's
172 * iterator.
173 *
174 * @param c the collection whose elements are to be placed into this list
175 * @throws NullPointerException if the specified collection is null
176 */
177 public ArrayList(Collection<? extends E> c) {
178 elementData = c.toArray();
179 if ((size = elementData.length) != 0) {
180 // c.toArray might (incorrectly) not return Object[] (see 6260652)
181 if (elementData.getClass() != Object[].class)
182 elementData = Arrays.copyOf(elementData, size, Object[].class);
183 } else {
184 // replace with empty array.
185 this.elementData = EMPTY_ELEMENTDATA;
186 }
187 }
188
189 /**
190 * Trims the capacity of this <tt>ArrayList</tt> instance to be the
191 * list's current size. An application can use this operation to minimize
192 * the storage of an <tt>ArrayList</tt> instance.
193 */
194 public void trimToSize() {
195 modCount++;
196 if (size < elementData.length) {
197 elementData = (size == 0)
198 ? EMPTY_ELEMENTDATA
199 : Arrays.copyOf(elementData, size);
200 }
201 }
202
203 /**
204 * Increases the capacity of this <tt>ArrayList</tt> instance, if
205 * necessary, to ensure that it can hold at least the number of elements
206 * specified by the minimum capacity argument.
207 *
208 * @param minCapacity the desired minimum capacity
209 */
210 public void ensureCapacity(int minCapacity) {
211 int minExpand = (elementData != DEFAULTCAPACITY_EMPTY_ELEMENTDATA)
212 // any size if not default element table
213 ? 0
214 // larger than default for default empty table. It's already
215 // supposed to be at default size.
216 : DEFAULT_CAPACITY;
217
218 if (minCapacity > minExpand) {
219 ensureExplicitCapacity(minCapacity);
220 }
221 }
222
223 private static int calculateCapacity(Object[] elementData, int minCapacity) {
224 if (elementData == DEFAULTCAPACITY_EMPTY_ELEMENTDATA) {
225 return Math.max(DEFAULT_CAPACITY, minCapacity);
226 }
227 return minCapacity;
228 }
229
230 private void ensureCapacityInternal(int minCapacity) {
231 ensureExplicitCapacity(calculateCapacity(elementData, minCapacity));
232 }
233
234 private void ensureExplicitCapacity(int minCapacity) {
235 modCount++;
236
237 // overflow-conscious code
238 if (minCapacity - elementData.length > 0)
239 grow(minCapacity);
240 }
241
242 /**
243 * The maximum size of array to allocate.
244 * Some VMs reserve some header words in an array.
245 * Attempts to allocate larger arrays may result in
246 * OutOfMemoryError: Requested array size exceeds VM limit
247 */
248 private static final int MAX_ARRAY_SIZE = Integer.MAX_VALUE - 8;
249
250 /**
251 * Increases the capacity to ensure that it can hold at least the
252 * number of elements specified by the minimum capacity argument.
253 *
254 * @param minCapacity the desired minimum capacity
255 */
256 private void grow(int minCapacity) {
257 // overflow-conscious code
258 int oldCapacity = elementData.length;
259 int newCapacity = oldCapacity + (oldCapacity >> 1);
260 if (newCapacity - minCapacity < 0)
261 newCapacity = minCapacity;
262 if (newCapacity - MAX_ARRAY_SIZE > 0)
263 newCapacity = hugeCapacity(minCapacity);
264 // minCapacity is usually close to size, so this is a win:
265 elementData = Arrays.copyOf(elementData, newCapacity);
266 }
267
268 private static int hugeCapacity(int minCapacity) {
269 if (minCapacity < 0) // overflow
270 throw new OutOfMemoryError();
271 return (minCapacity > MAX_ARRAY_SIZE) ?
272 Integer.MAX_VALUE :
273 MAX_ARRAY_SIZE;
274 }
275
276 /**
277 * Returns the number of elements in this list.
278 *
279 * @return the number of elements in this list
280 */
281 public int size() {
282 return size;
283 }
284
285 /**
286 * Returns <tt>true</tt> if this list contains no elements.
287 *
288 * @return <tt>true</tt> if this list contains no elements
289 */
290 public boolean isEmpty() {
291 return size == 0;
292 }
293
294 /**
295 * Returns <tt>true</tt> if this list contains the specified element.
296 * More formally, returns <tt>true</tt> if and only if this list contains
297 * at least one element <tt>e</tt> such that
298 * <tt>(o==null ? e==null : o.equals(e))</tt>.
299 *
300 * @param o element whose presence in this list is to be tested
301 * @return <tt>true</tt> if this list contains the specified element
302 */
303 public boolean contains(Object o) {
304 return indexOf(o) >= 0;
305 }
306
307 /**
308 * Returns the index of the first occurrence of the specified element
309 * in this list, or -1 if this list does not contain the element.
310 * More formally, returns the lowest index <tt>i</tt> such that
311 * <tt>(o==null ? get(i)==null : o.equals(get(i)))</tt>,
312 * or -1 if there is no such index.
313 */
314 public int indexOf(Object o) {
315 if (o == null) {
316 for (int i = 0; i < size; i++)
317 if (elementData[i]==null)
318 return i;
319 } else {
320 for (int i = 0; i < size; i++)
321 if (o.equals(elementData[i]))
322 return i;
323 }
324 return -1;
325 }
326
327 /**
328 * Returns the index of the last occurrence of the specified element
329 * in this list, or -1 if this list does not contain the element.
330 * More formally, returns the highest index <tt>i</tt> such that
331 * <tt>(o==null ? get(i)==null : o.equals(get(i)))</tt>,
332 * or -1 if there is no such index.
333 */
334 public int lastIndexOf(Object o) {
335 if (o == null) {
336 for (int i = size-1; i >= 0; i--)
337 if (elementData[i]==null)
338 return i;
339 } else {
340 for (int i = size-1; i >= 0; i--)
341 if (o.equals(elementData[i]))
342 return i;
343 }
344 return -1;
345 }
346
347 /**
348 * Returns a shallow copy of this <tt>ArrayList</tt> instance. (The
349 * elements themselves are not copied.)
350 *
351 * @return a clone of this <tt>ArrayList</tt> instance
352 */
353 public Object clone() {
354 try {
355 ArrayList<?> v = (ArrayList<?>) super.clone();
356 v.elementData = Arrays.copyOf(elementData, size);
357 v.modCount = 0;
358 return v;
359 } catch (CloneNotSupportedException e) {
360 // this shouldn't happen, since we are Cloneable
361 throw new InternalError(e);
362 }
363 }
364
365 /**
366 * Returns an array containing all of the elements in this list
367 * in proper sequence (from first to last element).
368 *
369 * <p>The returned array will be "safe" in that no references to it are
370 * maintained by this list. (In other words, this method must allocate
371 * a new array). The caller is thus free to modify the returned array.
372 *
373 * <p>This method acts as bridge between array-based and collection-based
374 * APIs.
375 *
376 * @return an array containing all of the elements in this list in
377 * proper sequence
378 */
379 public Object[] toArray() {
380 return Arrays.copyOf(elementData, size);
381 }
382
383 /**
384 * Returns an array containing all of the elements in this list in proper
385 * sequence (from first to last element); the runtime type of the returned
386 * array is that of the specified array. If the list fits in the
387 * specified array, it is returned therein. Otherwise, a new array is
388 * allocated with the runtime type of the specified array and the size of
389 * this list.
390 *
391 * <p>If the list fits in the specified array with room to spare
392 * (i.e., the array has more elements than the list), the element in
393 * the array immediately following the end of the collection is set to
394 * <tt>null</tt>. (This is useful in determining the length of the
395 * list <i>only</i> if the caller knows that the list does not contain
396 * any null elements.)
397 *
398 * @param a the array into which the elements of the list are to
399 * be stored, if it is big enough; otherwise, a new array of the
400 * same runtime type is allocated for this purpose.
401 * @return an array containing the elements of the list
402 * @throws ArrayStoreException if the runtime type of the specified array
403 * is not a supertype of the runtime type of every element in
404 * this list
405 * @throws NullPointerException if the specified array is null
406 */
407 @SuppressWarnings("unchecked")
408 public <T> T[] toArray(T[] a) {
409 if (a.length < size)
410 // Make a new array of a's runtime type, but my contents:
411 return (T[]) Arrays.copyOf(elementData, size, a.getClass());
412 System.arraycopy(elementData, 0, a, 0, size);
413 if (a.length > size)
414 a[size] = null;
415 return a;
416 }
417
418 // Positional Access Operations
419
420 @SuppressWarnings("unchecked")
421 E elementData(int index) {
422 return (E) elementData[index];
423 }
424
425 /**
426 * Returns the element at the specified position in this list.
427 *
428 * @param index index of the element to return
429 * @return the element at the specified position in this list
430 * @throws IndexOutOfBoundsException {@inheritDoc}
431 */
432 public E get(int index) {
433 rangeCheck(index);
434
435 return elementData(index);
436 }
437
438 /**
439 * Replaces the element at the specified position in this list with
440 * the specified element.
441 *
442 * @param index index of the element to replace
443 * @param element element to be stored at the specified position
444 * @return the element previously at the specified position
445 * @throws IndexOutOfBoundsException {@inheritDoc}
446 */
447 public E set(int index, E element) {
448 rangeCheck(index);
449
450 E oldValue = elementData(index);
451 elementData[index] = element;
452 return oldValue;
453 }
454
455 /**
456 * Appends the specified element to the end of this list.
457 *
458 * @param e element to be appended to this list
459 * @return <tt>true</tt> (as specified by {@link Collection#add})
460 */
461 public boolean add(E e) {
462 ensureCapacityInternal(size + 1); // Increments modCount!!
463 elementData[size++] = e;
464 return true;
465 }
466
467 /**
468 * Inserts the specified element at the specified position in this
469 * list. Shifts the element currently at that position (if any) and
470 * any subsequent elements to the right (adds one to their indices).
471 *
472 * @param index index at which the specified element is to be inserted
473 * @param element element to be inserted
474 * @throws IndexOutOfBoundsException {@inheritDoc}
475 */
476 public void add(int index, E element) {
477 rangeCheckForAdd(index);
478
479 ensureCapacityInternal(size + 1); // Increments modCount!!
480 System.arraycopy(elementData, index, elementData, index + 1,
481 size - index);
482 elementData[index] = element;
483 size++;
484 }
485
486 /**
487 * Removes the element at the specified position in this list.
488 * Shifts any subsequent elements to the left (subtracts one from their
489 * indices).
490 *
491 * @param index the index of the element to be removed
492 * @return the element that was removed from the list
493 * @throws IndexOutOfBoundsException {@inheritDoc}
494 */
495 public E remove(int index) {
496 rangeCheck(index);
497
498 modCount++;
499 E oldValue = elementData(index);
500
501 int numMoved = size - index - 1;
502 if (numMoved > 0)
503 System.arraycopy(elementData, index+1, elementData, index,
504 numMoved);
505 elementData[--size] = null; // clear to let GC do its work
506
507 return oldValue;
508 }
509
510 /**
511 * Removes the first occurrence of the specified element from this list,
512 * if it is present. If the list does not contain the element, it is
513 * unchanged. More formally, removes the element with the lowest index
514 * <tt>i</tt> such that
515 * <tt>(o==null ? get(i)==null : o.equals(get(i)))</tt>
516 * (if such an element exists). Returns <tt>true</tt> if this list
517 * contained the specified element (or equivalently, if this list
518 * changed as a result of the call).
519 *
520 * @param o element to be removed from this list, if present
521 * @return <tt>true</tt> if this list contained the specified element
522 */
523 public boolean remove(Object o) {
524 if (o == null) {
525 for (int index = 0; index < size; index++)
526 if (elementData[index] == null) {
527 fastRemove(index);
528 return true;
529 }
530 } else {
531 for (int index = 0; index < size; index++)
532 if (o.equals(elementData[index])) {
533 fastRemove(index);
534 return true;
535 }
536 }
537 return false;
538 }
539
540 /*
541 * Private remove method that skips bounds checking and does not
542 * return the value removed.
543 */
544 private void fastRemove(int index) {
545 modCount++;
546 int numMoved = size - index - 1;
547 if (numMoved > 0)
548 System.arraycopy(elementData, index+1, elementData, index,
549 numMoved);
550 elementData[--size] = null; // clear to let GC do its work
551 }
552
553 /**
554 * Removes all of the elements from this list. The list will
555 * be empty after this call returns.
556 */
557 public void clear() {
558 modCount++;
559
560 // clear to let GC do its work
561 for (int i = 0; i < size; i++)
562 elementData[i] = null;
563
564 size = 0;
565 }
566
567 /**
568 * Appends all of the elements in the specified collection to the end of
569 * this list, in the order that they are returned by the
570 * specified collection's Iterator. The behavior of this operation is
571 * undefined if the specified collection is modified while the operation
572 * is in progress. (This implies that the behavior of this call is
573 * undefined if the specified collection is this list, and this
574 * list is nonempty.)
575 *
576 * @param c collection containing elements to be added to this list
577 * @return <tt>true</tt> if this list changed as a result of the call
578 * @throws NullPointerException if the specified collection is null
579 */
580 public boolean addAll(Collection<? extends E> c) {
581 Object[] a = c.toArray();
582 int numNew = a.length;
583 ensureCapacityInternal(size + numNew); // Increments modCount
584 System.arraycopy(a, 0, elementData, size, numNew);
585 size += numNew;
586 return numNew != 0;
587 }
588
589 /**
590 * Inserts all of the elements in the specified collection into this
591 * list, starting at the specified position. Shifts the element
592 * currently at that position (if any) and any subsequent elements to
593 * the right (increases their indices). The new elements will appear
594 * in the list in the order that they are returned by the
595 * specified collection's iterator.
596 *
597 * @param index index at which to insert the first element from the
598 * specified collection
599 * @param c collection containing elements to be added to this list
600 * @return <tt>true</tt> if this list changed as a result of the call
601 * @throws IndexOutOfBoundsException {@inheritDoc}
602 * @throws NullPointerException if the specified collection is null
603 */
604 public boolean addAll(int index, Collection<? extends E> c) {
605 rangeCheckForAdd(index);
606
607 Object[] a = c.toArray();
608 int numNew = a.length;
609 ensureCapacityInternal(size + numNew); // Increments modCount
610
611 int numMoved = size - index;
612 if (numMoved > 0)
613 System.arraycopy(elementData, index, elementData, index + numNew,
614 numMoved);
615
616 System.arraycopy(a, 0, elementData, index, numNew);
617 size += numNew;
618 return numNew != 0;
619 }
620
621 /**
622 * Removes from this list all of the elements whose index is between
623 * {@code fromIndex}, inclusive, and {@code toIndex}, exclusive.
624 * Shifts any succeeding elements to the left (reduces their index).
625 * This call shortens the list by {@code (toIndex - fromIndex)} elements.
626 * (If {@code toIndex==fromIndex}, this operation has no effect.)
627 *
628 * @throws IndexOutOfBoundsException if {@code fromIndex} or
629 * {@code toIndex} is out of range
630 * ({@code fromIndex < 0 ||
631 * fromIndex >= size() ||
632 * toIndex > size() ||
633 * toIndex < fromIndex})
634 */
635 protected void removeRange(int fromIndex, int toIndex) {
636 modCount++;
637 int numMoved = size - toIndex;
638 System.arraycopy(elementData, toIndex, elementData, fromIndex,
639 numMoved);
640
641 // clear to let GC do its work
642 int newSize = size - (toIndex-fromIndex);
643 for (int i = newSize; i < size; i++) {
644 elementData[i] = null;
645 }
646 size = newSize;
647 }
648
649 /**
650 * Checks if the given index is in range. If not, throws an appropriate
651 * runtime exception. This method does *not* check if the index is
652 * negative: It is always used immediately prior to an array access,
653 * which throws an ArrayIndexOutOfBoundsException if index is negative.
654 */
655 private void rangeCheck(int index) {
656 if (index >= size)
657 throw new IndexOutOfBoundsException(outOfBoundsMsg(index));
658 }
659
660 /**
661 * A version of rangeCheck used by add and addAll.
662 */
663 private void rangeCheckForAdd(int index) {
664 if (index > size || index < 0)
665 throw new IndexOutOfBoundsException(outOfBoundsMsg(index));
666 }
667
668 /**
669 * Constructs an IndexOutOfBoundsException detail message.
670 * Of the many possible refactorings of the error handling code,
671 * this "outlining" performs best with both server and client VMs.
672 */
673 private String outOfBoundsMsg(int index) {
674 return "Index: "+index+", Size: "+size;
675 }
676
677 /**
678 * Removes from this list all of its elements that are contained in the
679 * specified collection.
680 *
681 * @param c collection containing elements to be removed from this list
682 * @return {@code true} if this list changed as a result of the call
683 * @throws ClassCastException if the class of an element of this list
684 * is incompatible with the specified collection
685 * (<a href="Collection.html#optional-restrictions">optional</a>)
686 * @throws NullPointerException if this list contains a null element and the
687 * specified collection does not permit null elements
688 * (<a href="Collection.html#optional-restrictions">optional</a>),
689 * or if the specified collection is null
690 * @see Collection#contains(Object)
691 */
692 public boolean removeAll(Collection<?> c) {
693 Objects.requireNonNull(c);
694 return batchRemove(c, false);
695 }
696
697 /**
698 * Retains only the elements in this list that are contained in the
699 * specified collection. In other words, removes from this list all
700 * of its elements that are not contained in the specified collection.
701 *
702 * @param c collection containing elements to be retained in this list
703 * @return {@code true} if this list changed as a result of the call
704 * @throws ClassCastException if the class of an element of this list
705 * is incompatible with the specified collection
706 * (<a href="Collection.html#optional-restrictions">optional</a>)
707 * @throws NullPointerException if this list contains a null element and the
708 * specified collection does not permit null elements
709 * (<a href="Collection.html#optional-restrictions">optional</a>),
710 * or if the specified collection is null
711 * @see Collection#contains(Object)
712 */
713 public boolean retainAll(Collection<?> c) {
714 Objects.requireNonNull(c);
715 return batchRemove(c, true);
716 }
717
718 private boolean batchRemove(Collection<?> c, boolean complement) {
719 final Object[] elementData = this.elementData;
720 int r = 0, w = 0;
721 boolean modified = false;
722 try {
723 for (; r < size; r++)
724 if (c.contains(elementData[r]) == complement)
725 elementData[w++] = elementData[r];
726 } finally {
727 // Preserve behavioral compatibility with AbstractCollection,
728 // even if c.contains() throws.
729 if (r != size) {
730 System.arraycopy(elementData, r,
731 elementData, w,
732 size - r);
733 w += size - r;
734 }
735 if (w != size) {
736 // clear to let GC do its work
737 for (int i = w; i < size; i++)
738 elementData[i] = null;
739 modCount += size - w;
740 size = w;
741 modified = true;
742 }
743 }
744 return modified;
745 }
746
747 /**
748 * Save the state of the <tt>ArrayList</tt> instance to a stream (that
749 * is, serialize it).
750 *
751 * @serialData The length of the array backing the <tt>ArrayList</tt>
752 * instance is emitted (int), followed by all of its elements
753 * (each an <tt>Object</tt>) in the proper order.
754 */
755 private void writeObject(java.io.ObjectOutputStream s)
756 throws java.io.IOException{
757 // Write out element count, and any hidden stuff
758 int expectedModCount = modCount;
759 s.defaultWriteObject();
760
761 // Write out size as capacity for behavioural compatibility with clone()
762 s.writeInt(size);
763
764 // Write out all elements in the proper order.
765 for (int i=0; i<size; i++) {
766 s.writeObject(elementData[i]);
767 }
768
769 if (modCount != expectedModCount) {
770 throw new ConcurrentModificationException();
771 }
772 }
773
774 /**
775 * Reconstitute the <tt>ArrayList</tt> instance from a stream (that is,
776 * deserialize it).
777 */
778 private void readObject(java.io.ObjectInputStream s)
779 throws java.io.IOException, ClassNotFoundException {
780 elementData = EMPTY_ELEMENTDATA;
781
782 // Read in size, and any hidden stuff
783 s.defaultReadObject();
784
785 // Read in capacity
786 s.readInt(); // ignored
787
788 if (size > 0) {
789 // be like clone(), allocate array based upon size not capacity
790 int capacity = calculateCapacity(elementData, size);
791 SharedSecrets.getJavaOISAccess().checkArray(s, Object[].class, capacity);
792 ensureCapacityInternal(size);
793
794 Object[] a = elementData;
795 // Read in all elements in the proper order.
796 for (int i=0; i<size; i++) {
797 a[i] = s.readObject();
798 }
799 }
800 }
801
802 /**
803 * Returns a list iterator over the elements in this list (in proper
804 * sequence), starting at the specified position in the list.
805 * The specified index indicates the first element that would be
806 * returned by an initial call to {@link ListIterator#next next}.
807 * An initial call to {@link ListIterator#previous previous} would
808 * return the element with the specified index minus one.
809 *
810 * <p>The returned list iterator is <a href="#fail-fast"><i>fail-fast</i></a>.
811 *
812 * @throws IndexOutOfBoundsException {@inheritDoc}
813 */
814 public ListIterator<E> listIterator(int index) {
815 if (index < 0 || index > size)
816 throw new IndexOutOfBoundsException("Index: "+index);
817 return new ListItr(index);
818 }
819
820 /**
821 * Returns a list iterator over the elements in this list (in proper
822 * sequence).
823 *
824 * <p>The returned list iterator is <a href="#fail-fast"><i>fail-fast</i></a>.
825 *
826 * @see #listIterator(int)
827 */
828 public ListIterator<E> listIterator() {
829 return new ListItr(0);
830 }
831
832 /**
833 * Returns an iterator over the elements in this list in proper sequence.
834 *
835 * <p>The returned iterator is <a href="#fail-fast"><i>fail-fast</i></a>.
836 *
837 * @return an iterator over the elements in this list in proper sequence
838 */
839 public Iterator<E> iterator() {
840 return new Itr();
841 }
842
843 /**
844 * An optimized version of AbstractList.Itr
845 */
846 private class Itr implements Iterator<E> {
847 int cursor; // index of next element to return
848 int lastRet = -1; // index of last element returned; -1 if no such
849 int expectedModCount = modCount;
850
851 Itr() {}
852
853 public boolean hasNext() {
854 return cursor != size;
855 }
856
857 @SuppressWarnings("unchecked")
858 public E next() {
859 checkForComodification();
860 int i = cursor;
861 if (i >= size)
862 throw new NoSuchElementException();
863 Object[] elementData = ArrayList.this.elementData;
864 if (i >= elementData.length)
865 throw new ConcurrentModificationException();
866 cursor = i + 1;
867 return (E) elementData[lastRet = i];
868 }
869
870 public void remove() {
871 if (lastRet < 0)
872 throw new IllegalStateException();
873 checkForComodification();
874
875 try {
876 ArrayList.this.remove(lastRet);
877 cursor = lastRet;
878 lastRet = -1;
879 expectedModCount = modCount;
880 } catch (IndexOutOfBoundsException ex) {
881 throw new ConcurrentModificationException();
882 }
883 }
884
885 @Override
886 @SuppressWarnings("unchecked")
887 public void forEachRemaining(Consumer<? super E> consumer) {
888 Objects.requireNonNull(consumer);
889 final int size = ArrayList.this.size;
890 int i = cursor;
891 if (i >= size) {
892 return;
893 }
894 final Object[] elementData = ArrayList.this.elementData;
895 if (i >= elementData.length) {
896 throw new ConcurrentModificationException();
897 }
898 while (i != size && modCount == expectedModCount) {
899 consumer.accept((E) elementData[i++]);
900 }
901 // update once at end of iteration to reduce heap write traffic
902 cursor = i;
903 lastRet = i - 1;
904 checkForComodification();
905 }
906
907 final void checkForComodification() {
908 if (modCount != expectedModCount)
909 throw new ConcurrentModificationException();
910 }
911 }
912
913 /**
914 * An optimized version of AbstractList.ListItr
915 */
916 private class ListItr extends Itr implements ListIterator<E> {
917 ListItr(int index) {
918 super();
919 cursor = index;
920 }
921
922 public boolean hasPrevious() {
923 return cursor != 0;
924 }
925
926 public int nextIndex() {
927 return cursor;
928 }
929
930 public int previousIndex() {
931 return cursor - 1;
932 }
933
934 @SuppressWarnings("unchecked")
935 public E previous() {
936 checkForComodification();
937 int i = cursor - 1;
938 if (i < 0)
939 throw new NoSuchElementException();
940 Object[] elementData = ArrayList.this.elementData;
941 if (i >= elementData.length)
942 throw new ConcurrentModificationException();
943 cursor = i;
944 return (E) elementData[lastRet = i];
945 }
946
947 public void set(E e) {
948 if (lastRet < 0)
949 throw new IllegalStateException();
950 checkForComodification();
951
952 try {
953 ArrayList.this.set(lastRet, e);
954 } catch (IndexOutOfBoundsException ex) {
955 throw new ConcurrentModificationException();
956 }
957 }
958
959 public void add(E e) {
960 checkForComodification();
961
962 try {
963 int i = cursor;
964 ArrayList.this.add(i, e);
965 cursor = i + 1;
966 lastRet = -1;
967 expectedModCount = modCount;
968 } catch (IndexOutOfBoundsException ex) {
969 throw new ConcurrentModificationException();
970 }
971 }
972 }
973
974 /**
975 * Returns a view of the portion of this list between the specified
976 * {@code fromIndex}, inclusive, and {@code toIndex}, exclusive. (If
977 * {@code fromIndex} and {@code toIndex} are equal, the returned list is
978 * empty.) The returned list is backed by this list, so non-structural
979 * changes in the returned list are reflected in this list, and vice-versa.
980 * The returned list supports all of the optional list operations.
981 *
982 * <p>This method eliminates the need for explicit range operations (of
983 * the sort that commonly exist for arrays). Any operation that expects
984 * a list can be used as a range operation by passing a subList view
985 * instead of a whole list. For example, the following idiom
986 * removes a range of elements from a list:
987 * <pre>
988 * list.subList(from, to).clear();
989 * </pre>
990 * Similar idioms may be constructed for {@link #indexOf(Object)} and
991 * {@link #lastIndexOf(Object)}, and all of the algorithms in the
992 * {@link Collections} class can be applied to a subList.
993 *
994 * <p>The semantics of the list returned by this method become undefined if
995 * the backing list (i.e., this list) is <i>structurally modified</i> in
996 * any way other than via the returned list. (Structural modifications are
997 * those that change the size of this list, or otherwise perturb it in such
998 * a fashion that iterations in progress may yield incorrect results.)
999 *
1000 * @throws IndexOutOfBoundsException {@inheritDoc}
1001 * @throws IllegalArgumentException {@inheritDoc}
1002 */
1003 public List<E> subList(int fromIndex, int toIndex) {
1004 subListRangeCheck(fromIndex, toIndex, size);
1005 return new SubList(this, 0, fromIndex, toIndex);
1006 }
1007
1008 static void subListRangeCheck(int fromIndex, int toIndex, int size) {
1009 if (fromIndex < 0)
1010 throw new IndexOutOfBoundsException("fromIndex = " + fromIndex);
1011 if (toIndex > size)
1012 throw new IndexOutOfBoundsException("toIndex = " + toIndex);
1013 if (fromIndex > toIndex)
1014 throw new IllegalArgumentException("fromIndex(" + fromIndex +
1015 ") > toIndex(" + toIndex + ")");
1016 }
1017
1018 private class SubList extends AbstractList<E> implements RandomAccess {
1019 private final AbstractList<E> parent;
1020 private final int parentOffset;
1021 private final int offset;
1022 int size;
1023
1024 SubList(AbstractList<E> parent,
1025 int offset, int fromIndex, int toIndex) {
1026 this.parent = parent;
1027 this.parentOffset = fromIndex;
1028 this.offset = offset + fromIndex;
1029 this.size = toIndex - fromIndex;
1030 this.modCount = ArrayList.this.modCount;
1031 }
1032
1033 public E set(int index, E e) {
1034 rangeCheck(index);
1035 checkForComodification();
1036 E oldValue = ArrayList.this.elementData(offset + index);
1037 ArrayList.this.elementData[offset + index] = e;
1038 return oldValue;
1039 }
1040
1041 public E get(int index) {
1042 rangeCheck(index);
1043 checkForComodification();
1044 return ArrayList.this.elementData(offset + index);
1045 }
1046
1047 public int size() {
1048 checkForComodification();
1049 return this.size;
1050 }
1051
1052 public void add(int index, E e) {
1053 rangeCheckForAdd(index);
1054 checkForComodification();
1055 parent.add(parentOffset + index, e);
1056 this.modCount = parent.modCount;
1057 this.size++;
1058 }
1059
1060 public E remove(int index) {
1061 rangeCheck(index);
1062 checkForComodification();
1063 E result = parent.remove(parentOffset + index);
1064 this.modCount = parent.modCount;
1065 this.size--;
1066 return result;
1067 }
1068
1069 protected void removeRange(int fromIndex, int toIndex) {
1070 checkForComodification();
1071 parent.removeRange(parentOffset + fromIndex,
1072 parentOffset + toIndex);
1073 this.modCount = parent.modCount;
1074 this.size -= toIndex - fromIndex;
1075 }
1076
1077 public boolean addAll(Collection<? extends E> c) {
1078 return addAll(this.size, c);
1079 }
1080
1081 public boolean addAll(int index, Collection<? extends E> c) {
1082 rangeCheckForAdd(index);
1083 int cSize = c.size();
1084 if (cSize==0)
1085 return false;
1086
1087 checkForComodification();
1088 parent.addAll(parentOffset + index, c);
1089 this.modCount = parent.modCount;
1090 this.size += cSize;
1091 return true;
1092 }
1093
1094 public Iterator<E> iterator() {
1095 return listIterator();
1096 }
1097
1098 public ListIterator<E> listIterator(final int index) {
1099 checkForComodification();
1100 rangeCheckForAdd(index);
1101 final int offset = this.offset;
1102
1103 return new ListIterator<E>() {
1104 int cursor = index;
1105 int lastRet = -1;
1106 int expectedModCount = ArrayList.this.modCount;
1107
1108 public boolean hasNext() {
1109 return cursor != SubList.this.size;
1110 }
1111
1112 @SuppressWarnings("unchecked")
1113 public E next() {
1114 checkForComodification();
1115 int i = cursor;
1116 if (i >= SubList.this.size)
1117 throw new NoSuchElementException();
1118 Object[] elementData = ArrayList.this.elementData;
1119 if (offset + i >= elementData.length)
1120 throw new ConcurrentModificationException();
1121 cursor = i + 1;
1122 return (E) elementData[offset + (lastRet = i)];
1123 }
1124
1125 public boolean hasPrevious() {
1126 return cursor != 0;
1127 }
1128
1129 @SuppressWarnings("unchecked")
1130 public E previous() {
1131 checkForComodification();
1132 int i = cursor - 1;
1133 if (i < 0)
1134 throw new NoSuchElementException();
1135 Object[] elementData = ArrayList.this.elementData;
1136 if (offset + i >= elementData.length)
1137 throw new ConcurrentModificationException();
1138 cursor = i;
1139 return (E) elementData[offset + (lastRet = i)];
1140 }
1141
1142 @SuppressWarnings("unchecked")
1143 public void forEachRemaining(Consumer<? super E> consumer) {
1144 Objects.requireNonNull(consumer);
1145 final int size = SubList.this.size;
1146 int i = cursor;
1147 if (i >= size) {
1148 return;
1149 }
1150 final Object[] elementData = ArrayList.this.elementData;
1151 if (offset + i >= elementData.length) {
1152 throw new ConcurrentModificationException();
1153 }
1154 while (i != size && modCount == expectedModCount) {
1155 consumer.accept((E) elementData[offset + (i++)]);
1156 }
1157 // update once at end of iteration to reduce heap write traffic
1158 lastRet = cursor = i;
1159 checkForComodification();
1160 }
1161
1162 public int nextIndex() {
1163 return cursor;
1164 }
1165
1166 public int previousIndex() {
1167 return cursor - 1;
1168 }
1169
1170 public void remove() {
1171 if (lastRet < 0)
1172 throw new IllegalStateException();
1173 checkForComodification();
1174
1175 try {
1176 SubList.this.remove(lastRet);
1177 cursor = lastRet;
1178 lastRet = -1;
1179 expectedModCount = ArrayList.this.modCount;
1180 } catch (IndexOutOfBoundsException ex) {
1181 throw new ConcurrentModificationException();
1182 }
1183 }
1184
1185 public void set(E e) {
1186 if (lastRet < 0)
1187 throw new IllegalStateException();
1188 checkForComodification();
1189
1190 try {
1191 ArrayList.this.set(offset + lastRet, e);
1192 } catch (IndexOutOfBoundsException ex) {
1193 throw new ConcurrentModificationException();
1194 }
1195 }
1196
1197 public void add(E e) {
1198 checkForComodification();
1199
1200 try {
1201 int i = cursor;
1202 SubList.this.add(i, e);
1203 cursor = i + 1;
1204 lastRet = -1;
1205 expectedModCount = ArrayList.this.modCount;
1206 } catch (IndexOutOfBoundsException ex) {
1207 throw new ConcurrentModificationException();
1208 }
1209 }
1210
1211 final void checkForComodification() {
1212 if (expectedModCount != ArrayList.this.modCount)
1213 throw new ConcurrentModificationException();
1214 }
1215 };
1216 }
1217
1218 public List<E> subList(int fromIndex, int toIndex) {
1219 subListRangeCheck(fromIndex, toIndex, size);
1220 return new SubList(this, offset, fromIndex, toIndex);
1221 }
1222
1223 private void rangeCheck(int index) {
1224 if (index < 0 || index >= this.size)
1225 throw new IndexOutOfBoundsException(outOfBoundsMsg(index));
1226 }
1227
1228 private void rangeCheckForAdd(int index) {
1229 if (index < 0 || index > this.size)
1230 throw new IndexOutOfBoundsException(outOfBoundsMsg(index));
1231 }
1232
1233 private String outOfBoundsMsg(int index) {
1234 return "Index: "+index+", Size: "+this.size;
1235 }
1236
1237 private void checkForComodification() {
1238 if (ArrayList.this.modCount != this.modCount)
1239 throw new ConcurrentModificationException();
1240 }
1241
1242 public Spliterator<E> spliterator() {
1243 checkForComodification();
1244 return new ArrayListSpliterator<E>(ArrayList.this, offset,
1245 offset + this.size, this.modCount);
1246 }
1247 }
1248
1249 @Override
1250 public void forEach(Consumer<? super E> action) {
1251 Objects.requireNonNull(action);
1252 final int expectedModCount = modCount;
1253 @SuppressWarnings("unchecked")
1254 final E[] elementData = (E[]) this.elementData;
1255 final int size = this.size;
1256 for (int i=0; modCount == expectedModCount && i < size; i++) {
1257 action.accept(elementData[i]);
1258 }
1259 if (modCount != expectedModCount) {
1260 throw new ConcurrentModificationException();
1261 }
1262 }
1263
1264 /**
1265 * Creates a <em><a href="Spliterator.html#binding">late-binding</a></em>
1266 * and <em>fail-fast</em> {@link Spliterator} over the elements in this
1267 * list.
1268 *
1269 * <p>The {@code Spliterator} reports {@link Spliterator#SIZED},
1270 * {@link Spliterator#SUBSIZED}, and {@link Spliterator#ORDERED}.
1271 * Overriding implementations should document the reporting of additional
1272 * characteristic values.
1273 *
1274 * @return a {@code Spliterator} over the elements in this list
1275 * @since 1.8
1276 */
1277 @Override
1278 public Spliterator<E> spliterator() {
1279 return new ArrayListSpliterator<>(this, 0, -1, 0);
1280 }
1281
1282 /** Index-based split-by-two, lazily initialized Spliterator */
1283 static final class ArrayListSpliterator<E> implements Spliterator<E> {
1284
1285 /*
1286 * If ArrayLists were immutable, or structurally immutable (no
1287 * adds, removes, etc), we could implement their spliterators
1288 * with Arrays.spliterator. Instead we detect as much
1289 * interference during traversal as practical without
1290 * sacrificing much performance. We rely primarily on
1291 * modCounts. These are not guaranteed to detect concurrency
1292 * violations, and are sometimes overly conservative about
1293 * within-thread interference, but detect enough problems to
1294 * be worthwhile in practice. To carry this out, we (1) lazily
1295 * initialize fence and expectedModCount until the latest
1296 * point that we need to commit to the state we are checking
1297 * against; thus improving precision. (This doesn't apply to
1298 * SubLists, that create spliterators with current non-lazy
1299 * values). (2) We perform only a single
1300 * ConcurrentModificationException check at the end of forEach
1301 * (the most performance-sensitive method). When using forEach
1302 * (as opposed to iterators), we can normally only detect
1303 * interference after actions, not before. Further
1304 * CME-triggering checks apply to all other possible
1305 * violations of assumptions for example null or too-small
1306 * elementData array given its size(), that could only have
1307 * occurred due to interference. This allows the inner loop
1308 * of forEach to run without any further checks, and
1309 * simplifies lambda-resolution. While this does entail a
1310 * number of checks, note that in the common case of
1311 * list.stream().forEach(a), no checks or other computation
1312 * occur anywhere other than inside forEach itself. The other
1313 * less-often-used methods cannot take advantage of most of
1314 * these streamlinings.
1315 */
1316
1317 private final ArrayList<E> list;
1318 private int index; // current index, modified on advance/split
1319 private int fence; // -1 until used; then one past last index
1320 private int expectedModCount; // initialized when fence set
1321
1322 /** Create new spliterator covering the given range */
1323 ArrayListSpliterator(ArrayList<E> list, int origin, int fence,
1324 int expectedModCount) {
1325 this.list = list; // OK if null unless traversed
1326 this.index = origin;
1327 this.fence = fence;
1328 this.expectedModCount = expectedModCount;
1329 }
1330
1331 private int getFence() { // initialize fence to size on first use
1332 int hi; // (a specialized variant appears in method forEach)
1333 ArrayList<E> lst;
1334 if ((hi = fence) < 0) {
1335 if ((lst = list) == null)
1336 hi = fence = 0;
1337 else {
1338 expectedModCount = lst.modCount;
1339 hi = fence = lst.size;
1340 }
1341 }
1342 return hi;
1343 }
1344
1345 public ArrayListSpliterator<E> trySplit() {
1346 int hi = getFence(), lo = index, mid = (lo + hi) >>> 1;
1347 return (lo >= mid) ? null : // divide range in half unless too small
1348 new ArrayListSpliterator<E>(list, lo, index = mid,
1349 expectedModCount);
1350 }
1351
1352 public boolean tryAdvance(Consumer<? super E> action) {
1353 if (action == null)
1354 throw new NullPointerException();
1355 int hi = getFence(), i = index;
1356 if (i < hi) {
1357 index = i + 1;
1358 @SuppressWarnings("unchecked") E e = (E)list.elementData[i];
1359 action.accept(e);
1360 if (list.modCount != expectedModCount)
1361 throw new ConcurrentModificationException();
1362 return true;
1363 }
1364 return false;
1365 }
1366
1367 public void forEachRemaining(Consumer<? super E> action) {
1368 int i, hi, mc; // hoist accesses and checks from loop
1369 ArrayList<E> lst; Object[] a;
1370 if (action == null)
1371 throw new NullPointerException();
1372 if ((lst = list) != null && (a = lst.elementData) != null) {
1373 if ((hi = fence) < 0) {
1374 mc = lst.modCount;
1375 hi = lst.size;
1376 }
1377 else
1378 mc = expectedModCount;
1379 if ((i = index) >= 0 && (index = hi) <= a.length) {
1380 for (; i < hi; ++i) {
1381 @SuppressWarnings("unchecked") E e = (E) a[i];
1382 action.accept(e);
1383 }
1384 if (lst.modCount == mc)
1385 return;
1386 }
1387 }
1388 throw new ConcurrentModificationException();
1389 }
1390
1391 public long estimateSize() {
1392 return (long) (getFence() - index);
1393 }
1394
1395 public int characteristics() {
1396 return Spliterator.ORDERED | Spliterator.SIZED | Spliterator.SUBSIZED;
1397 }
1398 }
1399
1400 @Override
1401 public boolean removeIf(Predicate<? super E> filter) {
1402 Objects.requireNonNull(filter);
1403 // figure out which elements are to be removed
1404 // any exception thrown from the filter predicate at this stage
1405 // will leave the collection unmodified
1406 int removeCount = 0;
1407 final BitSet removeSet = new BitSet(size);
1408 final int expectedModCount = modCount;
1409 final int size = this.size;
1410 for (int i=0; modCount == expectedModCount && i < size; i++) {
1411 @SuppressWarnings("unchecked")
1412 final E element = (E) elementData[i];
1413 if (filter.test(element)) {
1414 removeSet.set(i);
1415 removeCount++;
1416 }
1417 }
1418 if (modCount != expectedModCount) {
1419 throw new ConcurrentModificationException();
1420 }
1421
1422 // shift surviving elements left over the spaces left by removed elements
1423 final boolean anyToRemove = removeCount > 0;
1424 if (anyToRemove) {
1425 final int newSize = size - removeCount;
1426 for (int i=0, j=0; (i < size) && (j < newSize); i++, j++) {
1427 i = removeSet.nextClearBit(i);
1428 elementData[j] = elementData[i];
1429 }
1430 for (int k=newSize; k < size; k++) {
1431 elementData[k] = null; // Let gc do its work
1432 }
1433 this.size = newSize;
1434 if (modCount != expectedModCount) {
1435 throw new ConcurrentModificationException();
1436 }
1437 modCount++;
1438 }
1439
1440 return anyToRemove;
1441 }
1442
1443 @Override
1444 @SuppressWarnings("unchecked")
1445 public void replaceAll(UnaryOperator<E> operator) {
1446 Objects.requireNonNull(operator);
1447 final int expectedModCount = modCount;
1448 final int size = this.size;
1449 for (int i=0; modCount == expectedModCount && i < size; i++) {
1450 elementData[i] = operator.apply((E) elementData[i]);
1451 }
1452 if (modCount != expectedModCount) {
1453 throw new ConcurrentModificationException();
1454 }
1455 modCount++;
1456 }
1457
1458 @Override
1459 @SuppressWarnings("unchecked")
1460 public void sort(Comparator<? super E> c) {
1461 final int expectedModCount = modCount;
1462 Arrays.sort((E[]) elementData, 0, size, c);
1463 if (modCount != expectedModCount) {
1464 throw new ConcurrentModificationException();
1465 }
1466 modCount++;
1467 }
1468 }
2. 擴容
新增元素時使用 ensureCapacityInternal() 方法來保證容量足夠，如果不夠時，需要使用 grow() 方法進行擴容，新容量的大小為 oldCapacity + (oldCapacity >> 1)，即 oldCapacity+oldCapacity/2。其中 oldCapacity >> 1 需要取整，所以新容量大約是舊容量的 1.5 倍左右。（oldCapacity 為偶數就是 1.5 倍，為奇數就是 1.5 倍-0.5）
擴容操作需要呼叫 Arrays.copyOf() 把原陣列整個複製到新陣列中，這個操作代價很高，因此最好在建立 ArrayList 物件時就指定大概的容量大小，減少擴容操作的次數。
1 public boolean add(E e) {
2 ensureCapacityInternal(size + 1); // Increments modCount!!
3 elementData[size++] = e;
4 return true;
5 }
6
7 private void ensureCapacityInternal(int minCapacity) {
8 if (elementData == DEFAULTCAPACITY_EMPTY_ELEMENTDATA) {
9 minCapacity = Math.max(DEFAULT_CAPACITY, minCapacity);
10 }
11 ensureExplicitCapacity(minCapacity);
12 }
13
14 private void ensureExplicitCapacity(int minCapacity) {
15 modCount++;
16 // overflow-conscious code
17 if (minCapacity - elementData.length > 0)
18 grow(minCapacity);
19 }
20
21 private void grow(int minCapacity) {
22 // overflow-conscious code
23 int oldCapacity = elementData.length;
24 int newCapacity = oldCapacity + (oldCapacity >> 1);
25 if (newCapacity - minCapacity < 0)
26 newCapacity = minCapacity;
27 if (newCapacity - MAX_ARRAY_SIZE > 0)
28 newCapacity = hugeCapacity(minCapacity);
29 // minCapacity is usually close to size, so this is a win:
30 elementData = Arrays.copyOf(elementData, newCapacity);
31 }
3. 刪除元素
需要呼叫 System.arraycopy() 將 index+1 後面的元素都複製到 index 位置上，該操作的時間複雜度為 O(N)，可以看到 ArrayList 刪除元素的代價是非常高的。
1 public E remove(int index) {
2 rangeCheck(index);
3 modCount++;
4 E oldValue = elementData(index);
5 int numMoved = size - index - 1;
6 if (numMoved > 0)
7 System.arraycopy(elementData, index+1, elementData, index, numMoved);
8 elementData[--size] = null; // clear to let GC do its work
9 return oldValue;
10 }
4. 序列化
ArrayList 基於陣列實現，並且具有動態擴容特性，因此儲存元素的陣列不一定都會被使用，那麼就沒必要全部進行序列化。
儲存元素的陣列 elementData 使用 transient 修飾，該關鍵字宣告陣列預設不會被序列化。
transient Object[] elementData; // non-private to simplify nested class access
ArrayList 實現了 writeObject() 和 readObject() 來控制只序列化陣列中有元素填充那部分內容。
1 private void readObject(java.io.ObjectInputStream s)
2 throws java.io.IOException, ClassNotFoundException {
3 elementData = EMPTY_ELEMENTDATA;
4
5 // Read in size, and any hidden stuff
6 s.defaultReadObject();
7
8 // Read in capacity
9 s.readInt(); // ignored
10
11 if (size > 0) {
12 // be like clone(), allocate array based upon size not capacity
13 ensureCapacityInternal(size);
14
15 Object[] a = elementData;
16 // Read in all elements in the proper order.
17 for (int i=0; i<size; i++) {
18 a[i] = s.readObject();
19 }
20 }
21 }
1 private void writeObject(java.io.ObjectOutputStream s)
2 throws java.io.IOException{
3 // Write out element count, and any hidden stuff
4 int expectedModCount = modCount;
5 s.defaultWriteObject();
6
7 // Write out size as capacity for behavioural compatibility with clone()
8 s.writeInt(size);
9
10 // Write out all elements in the proper order.
11 for (int i=0; i<size; i++) {
12 s.writeObject(elementData[i]);
13 }
14
15 if (modCount != expectedModCount) {
16 throw new ConcurrentModificationException();
17 }
18 }
序列化時需要使用 ObjectOutputStream 的 writeObject() 將物件轉換為位元組流並輸出。而 writeObject() 方法在傳入的物件存在 writeObject() 的時候會去反射呼叫該物件的 writeObject() 來實現序列化。反序列化使用的是 ObjectInputStream 的 readObject() 方法，原理類似。
1 ArrayList list = new ArrayList();
2 ObjectOutputStream oos = new ObjectOutputStream(new FileOutputStream(file));
3 oos.writeObject(list);
5. modCount
modCount 用來記錄 ArrayList 結構發生變化的次數。結構發生變化是指新增或者刪除至少一個元素的所有操作，或者是調整內部陣列的大小，僅僅只是設定元素的值不算結構發生變化。
在進行序列化或者迭代等操作時，需要比較操作前後 modCount 是否改變，如果改變了需要丟擲 ConcurrentModificationException。程式碼參以上序列化中的 writeObject() 方法。
Vector
1. 同步
它的實現與 ArrayList 類似，但是使用了 synchronized 進行同步。
1 public synchronized boolean add(E e) {
2 modCount++;
3 ensureCapacityHelper(elementCount + 1);
4 elementData[elementCount++] = e;
5 return true;
6 }
7
8 public synchronized E get(int index) {
9 if (index >= elementCount)
10 throw new ArrayIndexOutOfBoundsException(index);
11
12 return elementData(index);
13 }
2. 擴容
Vector 的建構函式可以傳入 capacityIncrement 引數，它的作用是在擴容時使容量 capacity 增長 capacityIncrement。如果這個引數的值小於等於 0，擴容時每次都令 capacity 為原來的兩倍。
1 public Vector(int initialCapacity, int capacityIncrement) {
2 super();
3 if (initialCapacity < 0)
4 throw new IllegalArgumentException("Illegal Capacity: "+
5 initialCapacity);
6 this.elementData = new Object[initialCapacity];
7 this.capacityIncrement = capacityIncrement;
8 }
1 private void grow(int minCapacity) {
2 // overflow-conscious code
3 int oldCapacity = elementData.length;
4 int newCapacity = oldCapacity + ((capacityIncrement > 0) ?
5 capacityIncrement : oldCapacity);
6 if (newCapacity - minCapacity < 0)
7 newCapacity = minCapacity;
8 if (newCapacity - MAX_ARRAY_SIZE > 0)
9 newCapacity = hugeCapacity(minCapacity);
10 elementData = Arrays.copyOf(elementData, newCapacity);
11 }
呼叫沒有 capacityIncrement 的建構函式時，capacityIncrement 值被設定為 0，也就是說預設情況下 Vector 每次擴容時容量都會翻倍。
1 public Vector(int initialCapacity) {
2 this(initialCapacity, 0);
3 }
4
5 public Vector() {
6 this(10);
7 }
3. 與 ArrayList 的比較
- Vector 是同步的，因此開銷就比 ArrayList 要大，訪問速度更慢。最好使用 ArrayList 而不是 Vector，因為同步操作完全可以由程式設計師自己來控制；
- Vector 每次擴容請求其大小的 2 倍（也可以通過建構函式設定增長的容量），而 ArrayList 是 1.5 倍。
4. 替代方案
可以使用 Collections.synchronizedList(); 得到一個執行緒安全的 ArrayList。
1 List<String> list = new ArrayList<>();
2 List<String> synList = Collections.synchronizedList(list);
也可以使用 concurrent 併發包下的 CopyOnWriteArrayList 類。
List<String> list = new CopyOnWriteArrayList<>();
CopyOnWriteArrayList
1. 讀寫分離
寫操作在一個複製的陣列上進行，讀操作還是在原始陣列中進行，讀寫分離，互不影響。
寫操作需要加鎖，防止併發寫入時導致寫入資料丟失。
寫操作結束之後需要把原始陣列指向新的複製陣列。
1 public boolean add(E e) {
2 final ReentrantLock lock = this.lock;
3 lock.lock();
4 try {
5 Object[] elements = getArray();
6 int len = elements.length;
7 Object[] newElements = Arrays.copyOf(elements, len + 1);
8 newElements[len] = e;
9 setArray(newElements);
10 return true;
11 } finally {
12 lock.unlock();
13 }
14 }
15
16 final void setArray(Object[] a) {
17 array = a;
18 }
1 @SuppressWarnings("unchecked")
2 private E get(Object[] a, int index) {
3 return (E) a[index];
4 }
2. 適用場景
CopyOnWriteArrayList 在寫操作的同時允許讀操作，大大提高了讀操作的效能，因此很適合讀多寫少的應用場景。
但是 CopyOnWriteArrayList 有其缺陷：
- 記憶體佔用：在寫操作時需要複製一個新的陣列，使得記憶體佔用為原來的兩倍左右；
- 資料不一致：讀操作不能讀取實時性的資料，因為部分寫操作的資料還未同步到讀陣列中。
所以 CopyOnWriteArrayList 不適合記憶體敏感以及對實時性要求很高的場景。
LinkedList
1. 概覽
基於雙向連結串列實現，使用 Node 儲存連結串列節點資訊。
1 private static class Node<E> {
2 E item;
3 Node<E> next;
4 Node<E> prev;
5 }
每個連結串列儲存了 first 和 last 指標：
1 transient Node<E> first;
2 transient Node<E> last;
2. 與 ArrayList 的比較
ArrayList 基於動態陣列實現，LinkedList 基於雙向連結串列實現。ArrayList 和 LinkedList 的區別可以歸結為陣列和連結串列的區別：
- 陣列支援隨機訪問，但插入刪除的代價很高，需要移動大量元素；
- 連結串列不支援隨機訪問，但插入刪除只需要改變指標。
HashMap
1、HashMap 底層原理備戰連結
https://www.cnblogs.com/taojietaoge/p/11359542.html
2、與 Hashtable 的比較
- Hashtable 使用 synchronized 來進行同步。
- HashMap 可以插入鍵為 null 的 Entry。
- HashMap 的迭代器是 fail-fast 迭代器。
- HashMap 不能保證隨著時間的推移 Map 中的元素次序是不變的。
ConcurrentHashMap
參考連結：https://www.cnblogs.com/taojietaoge/p/10301711.html
1. 儲存結構
1 static final class HashEntry<K,V> {
2 final int hash;
3 final K key;
4 volatile V value;
5 volatile HashEntry<K,V> next;
6 }
ConcurrentHashMap 和 HashMap 實現上類似，最主要的差別是 ConcurrentHashMap 採用了分段鎖（Segment），每個分段鎖維護著幾個桶（HashEntry），多個執行緒可以同時訪問不同分段鎖上的桶，從而使其併發度更高（併發度就是 Segment 的個數）。
Segment 繼承自 ReentrantLock。
1 static final class Segment<K,V> extends ReentrantLock implements Serializable {
2
3 private static final long serialVersionUID = 2249069246763182397L;
4
5 static final int MAX_SCAN_RETRIES =
6 Runtime.getRuntime().availableProcessors() > 1 ? 64 : 1;
7
8 transient volatile HashEntry<K,V>[] table;
9
10 transient int count;
11
12 transient int modCount;
13
14 transient int threshold;
15
16 final float loadFactor;
17 }
final Segment<K,V>[] segments;
預設的併發級別為 16，也就是說預設建立 16 個 Segment。
static final int DEFAULT_CONCURRENCY_LEVEL = 16;
2. size 操作
每個 Segment 維護了一個 count 變數來統計該 Segment 中的鍵值對個數。
1 /**
2 * The number of elements. Accessed only either within locks
3 * or among other volatile reads that maintain visibility.
4 */
5 transient int count;
在執行 size 操作時，需要遍歷所有 Segment 然後把 count 累計起來。
ConcurrentHashMap 在執行 size 操作時先嚐試不加鎖，如果連續兩次不加鎖操作得到的結果一致，那麼可以認為這個結果是正確的。
嘗試次數使用 RETRIES_BEFORE_LOCK 定義，該值為 2，retries 初始值為 -1，因此嘗試次數為 3。
如果嘗試的次數超過 3 次，就需要對每個 Segment 加鎖。
1 /**
2 * Number of unsynchronized retries in size and containsValue
3 * methods before resorting to locking. This is used to avoid
4 * unbounded retries if tables undergo continuous modification
5 * which would make it impossible to obtain an accurate result.
6 */
7 static final int RETRIES_BEFORE_LOCK = 2;
8
9 public int size() {
10 // Try a few times to get accurate count. On failure due to
11 // continuous async changes in table, resort to locking.
12 final Segment<K,V>[] segments = this.segments;
13 int size;
14 boolean overflow; // true if size overflows 32 bits
15 long sum; // sum of modCounts
16 long last = 0L; // previous sum
17 int retries = -1; // first iteration isn't retry
18 try {
19 for (;;) {
20 // 超過嘗試次數，則對每個 Segment 加鎖
21 if (retries++ == RETRIES_BEFORE_LOCK) {
22 for (int j = 0; j < segments.length; ++j)
23 ensureSegment(j).lock(); // force creation
24 }
25 sum = 0L;
26 size = 0;
27 overflow = false;
28 for (int j = 0; j < segments.length; ++j) {
29 Segment<K,V> seg = segmentAt(segments, j);
30 if (seg != null) {
31 sum += seg.modCount;
32 int c = seg.count;
33 if (c < 0 || (size += c) < 0)
34 overflow = true;
35 }
36 }
37 // 連續兩次得到的結果一致，則認為這個結果是正確的
38 if (sum == last)
39 break;
40 last = sum;
41 }
42 } finally {
43 if (retries > RETRIES_BEFORE_LOCK) {
44 for (int j = 0; j < segments.length; ++j)
45 segmentAt(segments, j).unlock();
46 }
47 }
48 return overflow ? Integer.MAX_VALUE : size;
49 }
3. JDK 1.8 的改動
JDK 1.7 使用分段鎖機制來實現併發更新操作，核心類為 Segment，它繼承自重入鎖 ReentrantLock，併發度與 Segment 數量相等。
JDK 1.8 使用了 CAS 操作來支援更高的併發度，在 CAS 操作失敗時使用內建鎖 synchronized。
並且 JDK 1.8 的實現也在連結串列過長時會轉換為紅黑樹。
LinkedHashMap
1、儲存結構
繼承自 HashMap，因此具有和 HashMap 一樣的快速查詢特性。
public class LinkedHashMap<K,V> extends HashMap<K,V> implements Map<K,V>
內部維護了一個雙向連結串列，用來維護插入順序或者 LRU 順序。
1 /**
2 * The head (eldest) of the doubly linked list.
3 */
4 transient LinkedHashMap.Entry<K,V> head;
5
6 /**
7 * The tail (youngest) of the doubly linked list.
8 */
9 transient LinkedHashMap.Entry<K,V> tail;
accessOrder 決定了順序，預設為 false，此時維護的是插入順序。
final boolean accessOrder;
LinkedHashMap 最重要的是以下用於維護順序的函式，它們會在 put、get 等方法中呼叫。
1 void afterNodeAccess(Node<K,V> p) { }
2 void afterNodeInsertion(boolean evict) { }
2、afterNodeAccess()
當一個節點被訪問時，如果 accessOrder 為 true，則會將該節點移到連結串列尾部。也就是說指定為 LRU 順序之後，在每次訪問一個節點時，會將這個節點移到連結串列尾部，保證連結串列尾部是最近訪問的節點，那麼連結串列首部就是最近最久未使用的節點。
1 void afterNodeAccess(Node<K,V> e) { // move node to last
2 LinkedHashMap.Entry<K,V> last;
3 if (accessOrder && (last = tail) != e) {
4 LinkedHashMap.Entry<K,V> p =
5 (LinkedHashMap.Entry<K,V>)e, b = p.before, a = p.after;
6 p.after = null;
7 if (b == null)
8 head = a;
9 else
10 b.after = a;
11 if (a != null)
12 a.before = b;
13 else
14 last = b;
15 if (last == null)
16 head = p;
17 else {
18 p.before = last;
19 last.after = p;
20 }
21 tail = p;
22 ++modCount;
23 }
24 }
3、afterNodeInsertion()
在 put 等操作之後執行，當 removeEldestEntry() 方法返回 true 時會移除最晚的節點，也就是連結串列首部節點 first。
evict 只有在構建 Map 的時候才為 false，在這裡為 true。
1 void afterNodeInsertion(boolean evict) { // possibly remove eldest
2 LinkedHashMap.Entry<K,V> first;
3 if (evict && (first = head) != null && removeEldestEntry(first)) {
4 K key = first.key;
5 removeNode(hash(key), key, null, false, true);
6 }
7 }
removeEldestEntry() 預設為 false，如果需要讓它為 true，需要繼承 LinkedHashMap 並且覆蓋這個方法的實現，這在實現 LRU 的快取中特別有用，通過移除最近最久未使用的節點，從而保證快取空間足夠，並且快取的資料都是熱點資料。
1 protected boolean removeEldestEntry(Map.Entry<K,V> eldest) {
2 return false;
3 }
4、LRU 快取
以下是使用 LinkedHashMap 實現的一個 LRU 快取：
- 設定最大快取空間 MAX_ENTRIES 為 3；
- 使用 LinkedHashMap 的建構函式將 accessOrder 設定為 true，開啟 LRU 順序；
- 覆蓋 removeEldestEntry() 方法實現，在節點多於 MAX_ENTRIES 就會將最近最久未使用的資料移除。
1 class LRUCache<K, V> extends LinkedHashMap<K, V> {
2 private static final int MAX_ENTRIES = 3;
3
4 protected boolean removeEldestEntry(Map.Entry eldest) {
5 return size() > MAX_ENTRIES;
6 }
7
8 LRUCache() {
9 super(MAX_ENTRIES, 0.75f, true);
10 }
11 }
1 public static void main(String[] args) {
2 LRUCache<Integer, String> cache = new LRUCache<>();
3 cache.put(1, "a");
4 cache.put(2, "b");
5 cache.put(3, "c");
6 cache.get(1);
7 cache.put(4, "d");
8 System.out.println(cache.keySet()); // [3, 1, 4]
9 }
WeakHashMap
1、儲存結構
WeakHashMap 的 Entry 繼承自 WeakReference，被 WeakReference 關聯的物件在下一次垃圾回收時會被回收。
WeakHashMap 主要用來實現快取，通過使用 WeakHashMap 來引用快取物件，由 JVM 對這部分快取進行回收。
private static class Entry<K,V> extends WeakReference<Object> implements Map.Entry<K,V>
2、ConcurrentCache
Tomcat 中的 ConcurrentCache 使用了 WeakHashMap 來實現快取功能。
ConcurrentCache 採取的是分代快取：
- 經常使用的物件放入 eden 中，eden 使用 ConcurrentHashMap 實現，不用擔心會被回收（伊甸園）；
- 不常用的物件放入 longterm，longterm 使用 WeakHashMap 實現，這些老物件會被垃圾收集器回收。
- 當呼叫 get() 方法時，會先從 eden 區獲取，如果沒有找到的話再到 longterm 獲取，當從 longterm 獲取到就把物件放入 eden 中，從而保證經常被訪問的節點不容易被回收。
- 當呼叫 put() 方法時，如果 eden 的大小超過了 size，那麼就將 eden 中的所有物件都放入 longterm 中，利用虛擬機器回收掉一部分不經常使用的物件。
1 public final class ConcurrentCache<K, V> {
2
3 private final int size;
4
5 private final Map<K, V> eden;
6
7 private final Map<K, V> longterm;
8
9 public ConcurrentCache(int size) {
10 this.size = size;
11 this.eden = new ConcurrentHashMap<>(size);
12 this.longterm = new WeakHashMap<>(size);
13 }
14
15 public V get(K k) {
16 V v = this.eden.get(k);
17 if (v == null) {
18 v = this.longterm.get(k);
19 if (v != null)
20 this.eden.put(k, v);
21 }
22 return v;
23 }
24
25 public void put(K k, V v) {
26 if (this.eden.size() >= size) {
27 this.longterm.putAll(this.eden);
28 this.eden.clear();
29 }
30 this.eden.put(k, v);
31 }
32 }
玉階生白露
夜久侵羅襪