Java多线程系列--“JUC集合”07之 ArrayBlockingQueue
来源：互联网 发布：macbook自带修图软件 编辑：程序博客网 时间：2024/05/22 11:38
概要
本章对Java.util.concurrent包中的ArrayBlockingQueue类进行详细的介绍。内容包括：
ArrayBlockingQueue介绍
ArrayBlockingQueue原理和数据结构
ArrayBlockingQueue函数列表
ArrayBlockingQueue源码分析(JDK1.7.0_40版本)
ArrayBlockingQueue示例
转载请注明出处：http://www.cnblogs.com/skywang12345/p/3498652.html
ArrayBlockingQueue介绍
ArrayBlockingQueue是数组实现的线程安全的有界的阻塞队列。
线程安全是指，ArrayBlockingQueue内部通过“互斥锁”保护竞争资源，实现了多线程对竞争资源的互斥访问。而有界，则是指ArrayBlockingQueue对应的数组是有界限的。 阻塞队列，是指多线程访问竞争资源时，当竞争资源已被某线程获取时，其它要获取该资源的线程需要阻塞等待；而且，ArrayBlockingQueue是按 FIFO（先进先出）原则对元素进行排序，元素都是从尾部插入到队列，从头部开始返回。
注意：ArrayBlockingQueue不同于ConcurrentLinkedQueue，ArrayBlockingQueue是数组实现的，并且是有界限的；而ConcurrentLinkedQueue是链表实现的，是无界限的。
ArrayBlockingQueue原理和数据结构
ArrayBlockingQueue的数据结构，如下图所示：
说明：
1. ArrayBlockingQueue继承于AbstractQueue，并且它实现了BlockingQueue接口。
2. ArrayBlockingQueue内部是通过Object[]数组保存数据的，也就是说ArrayBlockingQueue本质上是通过数组实现的。ArrayBlockingQueue的大小，即数组的容量是创建ArrayBlockingQueue时指定的。
3. ArrayBlockingQueue与ReentrantLock是组合关系，ArrayBlockingQueue中包含一个ReentrantLock对象(lock)。ReentrantLock是可重入的互斥锁，ArrayBlockingQueue就是根据该互斥锁实现“多线程对竞争资源的互斥访问”。而且，ReentrantLock分为公平锁和非公平锁，关于具体使用公平锁还是非公平锁，在创建ArrayBlockingQueue时可以指定；而且，ArrayBlockingQueue默认会使用非公平锁。
4. ArrayBlockingQueue与Condition是组合关系，ArrayBlockingQueue中包含两个Condition对象(notEmpty和notFull)。而且，Condition又依赖于ArrayBlockingQueue而存在，通过Condition可以实现对ArrayBlockingQueue的更精确的访问 -- (01)若某线程(线程A)要取数据时，数组正好为空，则该线程会执行notEmpty.await()进行等待；当其它某个线程(线程B)向数组中插入了数据之后，会调用notEmpty.signal()唤醒“notEmpty上的等待线程”。此时，线程A会被唤醒从而得以继续运行。(02)若某线程(线程H)要插入数据时，数组已满，则该线程会它执行notFull.await()进行等待；当其它某个线程(线程I)取出数据之后，会调用notFull.signal()唤醒“notFull上的等待线程”。此时，线程H就会被唤醒从而得以继续运行。
关于ReentrantLock，公平锁，非公平锁，以及Condition等更多的内容，可以参考：
(01) Java多线程系列--“JUC锁”02之 互斥锁ReentrantLock
(02) Java多线程系列--“JUC锁”03之 公平锁(一)
(03) Java多线程系列--“JUC锁”04之 公平锁(二)
(04) Java多线程系列--“JUC锁”05之 非公平锁
(05) Java多线程系列--“JUC锁”06之 Condition条件
ArrayBlockingQueue函数列表
// 创建一个带有给定的（固定）容量和默认访问策略的 ArrayBlockingQueue。ArrayBlockingQueue(int capacity)// 创建一个具有给定的（固定）容量和指定访问策略的 ArrayBlockingQueue。ArrayBlockingQueue(int capacity, boolean fair)// 创建一个具有给定的（固定）容量和指定访问策略的 ArrayBlockingQueue，它最初包含给定 collection 的元素，并以 collection 迭代器的遍历顺序添加元素。ArrayBlockingQueue(int capacity, boolean fair, Collection<? extends E> c)// 将指定的元素插入到此队列的尾部（如果立即可行且不会超过该队列的容量），在成功时返回 true，如果此队列已满，则抛出 IllegalStateException。boolean add(E e)// 自动移除此队列中的所有元素。void clear()// 如果此队列包含指定的元素，则返回 true。boolean contains(Object o)// 移除此队列中所有可用的元素，并将它们添加到给定 collection 中。int drainTo(Collection<? super E> c)// 最多从此队列中移除给定数量的可用元素，并将这些元素添加到给定 collection 中。int drainTo(Collection<? super E> c, int maxElements)// 返回在此队列中的元素上按适当顺序进行迭代的迭代器。Iterator<E> iterator()// 将指定的元素插入到此队列的尾部（如果立即可行且不会超过该队列的容量），在成功时返回 true，如果此队列已满，则返回 false。boolean offer(E e)// 将指定的元素插入此队列的尾部，如果该队列已满，则在到达指定的等待时间之前等待可用的空间。boolean offer(E e, long timeout, TimeUnit unit)// 获取但不移除此队列的头；如果此队列为空，则返回 null。E peek()// 获取并移除此队列的头，如果此队列为空，则返回 null。E poll()// 获取并移除此队列的头部，在指定的等待时间前等待可用的元素（如果有必要）。E poll(long timeout, TimeUnit unit)// 将指定的元素插入此队列的尾部，如果该队列已满，则等待可用的空间。void put(E e)// 返回在无阻塞的理想情况下（不存在内存或资源约束）此队列能接受的其他元素数量。int remainingCapacity()// 从此队列中移除指定元素的单个实例（如果存在）。boolean remove(Object o)// 返回此队列中元素的数量。int size()// 获取并移除此队列的头部，在元素变得可用之前一直等待（如果有必要）。E take()// 返回一个按适当顺序包含此队列中所有元素的数组。Object[] toArray()// 返回一个按适当顺序包含此队列中所有元素的数组；返回数组的运行时类型是指定数组的运行时类型。<T> T[] toArray(T[] a)// 返回此 collection 的字符串表示形式。String toString()
ArrayBlockingQueue源码分析(JDK1.7.0_40版本)
ArrayBlockingQueue.java的完整源码如下：
1 /* 2 * ORACLE PROPRIETARY/CONFIDENTIAL. Use is subject to license terms. 3 * 4 * 5 * 6 * 7 * 8 * 9 * 10 * 11 * 12 * 13 * 14 * 15 * 16 * 17 * 18 * 19 * 20 * 21 * 22 * 23 */ 24 25 /* 26 * 27 * 28 * 29 * 30 * 31 * Written by Doug Lea with assistance from members of JCP JSR-166 32 * Expert Group and released to the public domain, as explained at 33 * http://creativecommons.org/publicdomain/zero/1.0/ 34 */ 35 36 package java.util.concurrent; 37 import java.util.concurrent.locks.*; 38 import java.util.*; 39 40 /** 41 * A bounded {@linkplain BlockingQueue blocking queue} backed by an 42 * array. This queue orders elements FIFO (first-in-first-out). The 43 * <em>head</em> of the queue is that element that has been on the 44 * queue the longest time. The <em>tail</em> of the queue is that 45 * element that has been on the queue the shortest time. New elements 46 * are inserted at the tail of the queue, and the queue retrieval 47 * operations obtain elements at the head of the queue. 48 * 49 * <p>This is a classic "bounded buffer", in which a 50 * fixed-sized array holds elements inserted by producers and 51 * extracted by consumers. Once created, the capacity cannot be 52 * changed. Attempts to {@code put} an element into a full queue 53 * will result in the operation blocking; attempts to {@code take} an 54 * element from an empty queue will similarly block. 55 * 56 * <p>This class supports an optional fairness policy for ordering 57 * waiting producer and consumer threads. By default, this ordering 58 * is not guaranteed. However, a queue constructed with fairness set 59 * to {@code true} grants threads access in FIFO order. Fairness 60 * generally decreases throughput but reduces variability and avoids 61 * starvation. 62 * 63 * <p>This class and its iterator implement all of the 64 * <em>optional</em> methods of the {@link Collection} and {@link 65 * Iterator} interfaces. 66 * 67 * <p>This class is a member of the 68 * <a href="{@docRoot}/../technotes/guides/collections/index.html"> 69 * Java Collections Framework</a>. 70 * 71 * @since 1.5 72 * @author Doug Lea 73 * @param <E> the type of elements held in this collection 74 */ 75 public class ArrayBlockingQueue<E> extends AbstractQueue<E> 76 implements BlockingQueue<E>, java.io.Serializable { 77 78 /** 79 * Serialization ID. This class relies on default serialization 80 * even for the items array, which is default-serialized, even if 81 * it is empty. Otherwise it could not be declared final, which is 82 * necessary here. 83 */ 84 private static final long serialVersionUID = -817911632652898426L; 85 86 /** The queued items */ 87 final Object[] items; 88 89 /** items index for next take, poll, peek or remove */ 90 int takeIndex; 91 92 /** items index for next put, offer, or add */ 93 int putIndex; 94 95 /** Number of elements in the queue */ 96 int count; 97 98 /* 99 * Concurrency control uses the classic two-condition algorithm100 * found in any textbook.101 */102 103 /** Main lock guarding all access */104 final ReentrantLock lock;105 /** Condition for waiting takes */106 private final Condition notEmpty;107 /** Condition for waiting puts */108 private final Condition notFull;109 110 // Internal helper methods111 112 /**113 * Circularly increment i.114 */115 final int inc(int i) {116 return (++i == items.length) ? 0 : i;117 }118 119 /**120 * Circularly decrement i.121 */122 final int dec(int i) {123 return ((i == 0) ? items.length : i) - 1;124 }125 126 @SuppressWarnings("unchecked")127 static <E> E cast(Object item) {128 return (E) item;129 }130 131 /**132 * Returns item at index i.133 */134 final E itemAt(int i) {135 return this.<E>cast(items[i]);136 }137 138 /**139 * Throws NullPointerException if argument is null.140 *141 * @param v the element142 */143 private static void checkNotNull(Object v) {144 if (v == null)145 throw new NullPointerException();146 }147 148 /**149 * Inserts element at current put position, advances, and signals.150 * Call only when holding lock.151 */152 private void insert(E x) {153 items[putIndex] = x;154 putIndex = inc(putIndex);155 ++count;156 notEmpty.signal();157 }158 159 /**160 * Extracts element at current take position, advances, and signals.161 * Call only when holding lock.162 */163 private E extract() {164 final Object[] items = this.items;165 E x = this.<E>cast(items[takeIndex]);166 items[takeIndex] = null;167 takeIndex = inc(takeIndex);168 --count;169 notFull.signal();170 return x;171 }172 173 /**174 * Deletes item at position i.175 * Utility for remove and iterator.remove.176 * Call only when holding lock.177 */178 void removeAt(int i) {179 final Object[] items = this.items;180 // if removing front item, just advance181 if (i == takeIndex) {182 items[takeIndex] = null;183 takeIndex = inc(takeIndex);184 } else {185 // slide over all others up through putIndex.186 for (;;) {187 int nexti = inc(i);188 if (nexti != putIndex) {189 items[i] = items[nexti];190 i = nexti;191 } else {192 items[i] = null;193 putIndex = i;194 break;195 }196 }197 }198 --count;199 notFull.signal();200 }201 202 /**203 * Creates an {@code ArrayBlockingQueue} with the given (fixed)204 * capacity and default access policy.205 *206 * @param capacity the capacity of this queue207 * @throws IllegalArgumentException if {@code capacity < 1}208 */209 public ArrayBlockingQueue(int capacity) {210 this(capacity, false);211 }212 213 /**214 * Creates an {@code ArrayBlockingQueue} with the given (fixed)215 * capacity and the specified access policy.216 *217 * @param capacity the capacity of this queue218 * @param fair if {@code true} then queue accesses for threads blocked219 * on insertion or removal, are processed in FIFO order;220 * if {@code false} the access order is unspecified.221 * @throws IllegalArgumentException if {@code capacity < 1}222 */223 public ArrayBlockingQueue(int capacity, boolean fair) {224 if (capacity <= 0)225 throw new IllegalArgumentException();226 this.items = new Object[capacity];227 lock = new ReentrantLock(fair);228 notEmpty = lock.newCondition();229 notFull = lock.newCondition();230 }231 232 /**233 * Creates an {@code ArrayBlockingQueue} with the given (fixed)234 * capacity, the specified access policy and initially containing the235 * elements of the given collection,236 * added in traversal order of the collection's iterator.237 *238 * @param capacity the capacity of this queue239 * @param fair if {@code true} then queue accesses for threads blocked240 * on insertion or removal, are processed in FIFO order;241 * if {@code false} the access order is unspecified.242 * @param c the collection of elements to initially contain243 * @throws IllegalArgumentException if {@code capacity} is less than244 * {@code c.size()}, or less than 1.245 * @throws NullPointerException if the specified collection or any246 * of its elements are null247 */248 public ArrayBlockingQueue(int capacity, boolean fair,249 Collection<? extends E> c) {250 this(capacity, fair);251 252 final ReentrantLock lock = this.lock;253 lock.lock(); // Lock only for visibility, not mutual exclusion254 try {255 int i = 0;256 try {257 for (E e : c) {258 checkNotNull(e);259 items[i++] = e;260 }261 } catch (ArrayIndexOutOfBoundsException ex) {262 throw new IllegalArgumentException();263 }264 count = i;265 putIndex = (i == capacity) ? 0 : i;266 } finally {267 lock.unlock();268 }269 }270 271 /**272 * Inserts the specified element at the tail of this queue if it is273 * possible to do so immediately without exceeding the queue's capacity,274 * returning {@code true} upon success and throwing an275 * {@code IllegalStateException} if this queue is full.276 *277 * @param e the element to add278 * @return {@code true} (as specified by {@link Collection#add})279 * @throws IllegalStateException if this queue is full280 * @throws NullPointerException if the specified element is null281 */282 public boolean add(E e) {283 return super.add(e);284 }285 286 /**287 * Inserts the specified element at the tail of this queue if it is288 * possible to do so immediately without exceeding the queue's capacity,289 * returning {@code true} upon success and {@code false} if this queue290 * is full. This method is generally preferable to method {@link #add},291 * which can fail to insert an element only by throwing an exception.292 *293 * @throws NullPointerException if the specified element is null294 */295 public boolean offer(E e) {296 checkNotNull(e);297 final ReentrantLock lock = this.lock;298 lock.lock();299 try {300 if (count == items.length)301 return false;302 else {303 insert(e);304 return true;305 }306 } finally {307 lock.unlock();308 }309 }310 311 /**312 * Inserts the specified element at the tail of this queue, waiting313 * for space to become available if the queue is full.314 *315 * @throws InterruptedException {@inheritDoc}316 * @throws NullPointerException {@inheritDoc}317 */318 public void put(E e) throws InterruptedException {319 checkNotNull(e);320 final ReentrantLock lock = this.lock;321 lock.lockInterruptibly();322 try {323 while (count == items.length)324 notFull.await();325 insert(e);326 } finally {327 lock.unlock();328 }329 }330 331 /**332 * Inserts the specified element at the tail of this queue, waiting333 * up to the specified wait time for space to become available if334 * the queue is full.335 *336 * @throws InterruptedException {@inheritDoc}337 * @throws NullPointerException {@inheritDoc}338 */339 public boolean offer(E e, long timeout, TimeUnit unit)340 throws InterruptedException {341 342 checkNotNull(e);343 long nanos = unit.toNanos(timeout);344 final ReentrantLock lock = this.lock;345 lock.lockInterruptibly();346 try {347 while (count == items.length) {348 if (nanos <= 0)349 return false;350 nanos = notFull.awaitNanos(nanos);351 }352 insert(e);353 return true;354 } finally {355 lock.unlock();356 }357 }358 359 public E poll() {360 final ReentrantLock lock = this.lock;361 lock.lock();362 try {363 return (count == 0) ? null : extract();364 } finally {365 lock.unlock();366 }367 }368 369 public E take() throws InterruptedException {370 final ReentrantLock lock = this.lock;371 lock.lockInterruptibly();372 try {373 while (count == 0)374 notEmpty.await();375 return extract();376 } finally {377 lock.unlock();378 }379 }380 381 public E poll(long timeout, TimeUnit unit) throws InterruptedException {382 long nanos = unit.toNanos(timeout);383 final ReentrantLock lock = this.lock;384 lock.lockInterruptibly();385 try {386 while (count == 0) {387 if (nanos <= 0)388 return null;389 nanos = notEmpty.awaitNanos(nanos);390 }391 return extract();392 } finally {393 lock.unlock();394 }395 }396 397 public E peek() {398 final ReentrantLock lock = this.lock;399 lock.lock();400 try {401 return (count == 0) ? null : itemAt(takeIndex);402 } finally {403 lock.unlock();404 }405 }406 407 // this doc comment is overridden to remove the reference to collections408 // greater in size than Integer.MAX_VALUE409 /**410 * Returns the number of elements in this queue.411 *412 * @return the number of elements in this queue413 */414 public int size() {415 final ReentrantLock lock = this.lock;416 lock.lock();417 try {418 return count;419 } finally {420 lock.unlock();421 }422 }423 424 // this doc comment is a modified copy of the inherited doc comment,425 // without the reference to unlimited queues.426 /**427 * Returns the number of additional elements that this queue can ideally428 * (in the absence of memory or resource constraints) accept without429 * blocking. This is always equal to the initial capacity of this queue430 * less the current {@code size} of this queue.431 *432 * <p>Note that you <em>cannot</em> always tell if an attempt to insert433 * an element will succeed by inspecting {@code remainingCapacity}434 * because it may be the case that another thread is about to435 * insert or remove an element.436 */437 public int remainingCapacity() {438 final ReentrantLock lock = this.lock;439 lock.lock();440 try {441 return items.length - count;442 } finally {443 lock.unlock();444 }445 }446 447 /**448 * Removes a single instance of the specified element from this queue,449 * if it is present. More formally, removes an element {@code e} such450 * that {@code o.equals(e)}, if this queue contains one or more such451 * elements.452 * Returns {@code true} if this queue contained the specified element453 * (or equivalently, if this queue changed as a result of the call).454 *455 * <p>Removal of interior elements in circular array based queues456 * is an intrinsically slow and disruptive operation, so should457 * be undertaken only in exceptional circumstances, ideally458 * only when the queue is known not to be accessible by other459 * threads.460 *461 * @param o element to be removed from this queue, if present462 * @return {@code true} if this queue changed as a result of the call463 */464 public boolean remove(Object o) {465 if (o == null) return false;466 final Object[] items = this.items;467 final ReentrantLock lock = this.lock;468 lock.lock();469 try {470 for (int i = takeIndex, k = count; k > 0; i = inc(i), k--) {471 if (o.equals(items[i])) {472 removeAt(i);473 return true;474 }475 }476 return false;477 } finally {478 lock.unlock();479 }480 }481 482 /**483 * Returns {@code true} if this queue contains the specified element.484 * More formally, returns {@code true} if and only if this queue contains485 * at least one element {@code e} such that {@code o.equals(e)}.486 *487 * @param o object to be checked for containment in this queue488 * @return {@code true} if this queue contains the specified element489 */490 public boolean contains(Object o) {491 if (o == null) return false;492 final Object[] items = this.items;493 final ReentrantLock lock = this.lock;494 lock.lock();495 try {496 for (int i = takeIndex, k = count; k > 0; i = inc(i), k--)497 if (o.equals(items[i]))498 return true;499 return false;500 } finally {501 lock.unlock();502 }503 }504 505 /**506 * Returns an array containing all of the elements in this queue, in507 * proper sequence.508 *509 * <p>The returned array will be "safe" in that no references to it are510 * maintained by this queue. (In other words, this method must allocate511 * a new array). The caller is thus free to modify the returned array.512 *513 * <p>This method acts as bridge between array-based and collection-based514 * APIs.515 *516 * @return an array containing all of the elements in this queue517 */518 public Object[] toArray() {519 final Object[] items = this.items;520 final ReentrantLock lock = this.lock;521 lock.lock();522 try {523 final int count = this.count;524 Object[] a = new Object[count];525 for (int i = takeIndex, k = 0; k < count; i = inc(i), k++)526 a[k] = items[i];527 return a;528 } finally {529 lock.unlock();530 }531 }532 533 /**534 * Returns an array containing all of the elements in this queue, in535 * proper sequence; the runtime type of the returned array is that of536 * the specified array. If the queue fits in the specified array, it537 * is returned therein. Otherwise, a new array is allocated with the538 * runtime type of the specified array and the size of this queue.539 *540 * <p>If this queue fits in the specified array with room to spare541 * (i.e., the array has more elements than this queue), the element in542 * the array immediately following the end of the queue is set to543 * {@code null}.544 *545 * <p>Like the {@link #toArray()} method, this method acts as bridge between546 * array-based and collection-based APIs. Further, this method allows547 * precise control over the runtime type of the output array, and may,548 * under certain circumstances, be used to save allocation costs.549 *550 * <p>Suppose {@code x} is a queue known to contain only strings.551 * The following code can be used to dump the queue into a newly552 * allocated array of {@code String}:553 *554 * <pre>555 * String[] y = x.toArray(new String[0]);</pre>556 *557 * Note that {@code toArray(new Object[0])} is identical in function to558 * {@code toArray()}.559 *560 * @param a the array into which the elements of the queue are to561 * be stored, if it is big enough; otherwise, a new array of the562 * same runtime type is allocated for this purpose563 * @return an array containing all of the elements in this queue564 * @throws ArrayStoreException if the runtime type of the specified array565 * is not a supertype of the runtime type of every element in566 * this queue567 * @throws NullPointerException if the specified array is null568 */569 @SuppressWarnings("unchecked")570 public <T> T[] toArray(T[] a) {571 final Object[] items = this.items;572 final ReentrantLock lock = this.lock;573 lock.lock();574 try {575 final int count = this.count;576 final int len = a.length;577 if (len < count)578 a = (T[])java.lang.reflect.Array.newInstance(579 a.getClass().getComponentType(), count);580 for (int i = takeIndex, k = 0; k < count; i = inc(i), k++)581 a[k] = (T) items[i];582 if (len > count)583 a[count] = null;584 return a;585 } finally {586 lock.unlock();587 }588 }589 590 public String toString() {591 final ReentrantLock lock = this.lock;592 lock.lock();593 try {594 int k = count;595 if (k == 0)596 return "[]";597 598 StringBuilder sb = new StringBuilder();599 sb.append('[');600 for (int i = takeIndex; ; i = inc(i)) {601 Object e = items[i];602 sb.append(e == this ? "(this Collection)" : e);603 if (--k == 0)604 return sb.append(']').toString();605 sb.append(',').append(' ');606 }607 } finally {608 lock.unlock();609 }610 }611 612 /**613 * Atomically removes all of the elements from this queue.614 * The queue will be empty after this call returns.615 */616 public void clear() {617 final Object[] items = this.items;618 final ReentrantLock lock = this.lock;619 lock.lock();620 try {621 for (int i = takeIndex, k = count; k > 0; i = inc(i), k--)622 items[i] = null;623 count = 0;624 putIndex = 0;625 takeIndex = 0;626 notFull.signalAll();627 } finally {628 lock.unlock();629 }630 }631 632 /**633 * @throws UnsupportedOperationException {@inheritDoc}634 * @throws ClassCastException {@inheritDoc}635 * @throws NullPointerException {@inheritDoc}636 * @throws IllegalArgumentException {@inheritDoc}637 */638 public int drainTo(Collection<? super E> c) {639 checkNotNull(c);640 if (c == this)641 throw new IllegalArgumentException();642 final Object[] items = this.items;643 final ReentrantLock lock = this.lock;644 lock.lock();645 try {646 int i = takeIndex;647 int n = 0;648 int max = count;649 while (n < max) {650 c.add(this.<E>cast(items[i]));651 items[i] = null;652 i = inc(i);653 ++n;654 }655 if (n > 0) {656 count = 0;657 putIndex = 0;658 takeIndex = 0;659 notFull.signalAll();660 }661 return n;662 } finally {663 lock.unlock();664 }665 }666 667 /**668 * @throws UnsupportedOperationException {@inheritDoc}669 * @throws ClassCastException {@inheritDoc}670 * @throws NullPointerException {@inheritDoc}671 * @throws IllegalArgumentException {@inheritDoc}672 */673 public int drainTo(Collection<? super E> c, int maxElements) {674 checkNotNull(c);675 if (c == this)676 throw new IllegalArgumentException();677 if (maxElements <= 0)678 return 0;679 final Object[] items = this.items;680 final ReentrantLock lock = this.lock;681 lock.lock();682 try {683 int i = takeIndex;684 int n = 0;685 int max = (maxElements < count) ? maxElements : count;686 while (n < max) {687 c.add(this.<E>cast(items[i]));688 items[i] = null;689 i = inc(i);690 ++n;691 }692 if (n > 0) {693 count -= n;694 takeIndex = i;695 notFull.signalAll();696 }697 return n;698 } finally {699 lock.unlock();700 }701 }702 703 /**704 * Returns an iterator over the elements in this queue in proper sequence.705 * The elements will be returned in order from first (head) to last (tail).706 *707 * <p>The returned {@code Iterator} is a "weakly consistent" iterator that708 * will never throw {@link java.util.ConcurrentModificationException709 * ConcurrentModificationException},710 * and guarantees to traverse elements as they existed upon711 * construction of the iterator, and may (but is not guaranteed to)712 * reflect any modifications subsequent to construction.713 *714 * @return an iterator over the elements in this queue in proper sequence715 */716 public Iterator<E> iterator() {717 return new Itr();718 }719 720 /**721 * Iterator for ArrayBlockingQueue. To maintain weak consistency722 * with respect to puts and takes, we (1) read ahead one slot, so723 * as to not report hasNext true but then not have an element to724 * return -- however we later recheck this slot to use the most725 * current value; (2) ensure that each array slot is traversed at726 * most once (by tracking "remaining" elements); (3) skip over727 * null slots, which can occur if takes race ahead of iterators.728 * However, for circular array-based queues, we cannot rely on any729 * well established definition of what it means to be weakly730 * consistent with respect to interior removes since these may731 * require slot overwrites in the process of sliding elements to732 * cover gaps. So we settle for resiliency, operating on733 * established apparent nexts, which may miss some elements that734 * have moved between calls to next.735 */736 private class Itr implements Iterator<E> {737 private int remaining; // Number of elements yet to be returned738 private int nextIndex; // Index of element to be returned by next739 private E nextItem; // Element to be returned by next call to next740 private E lastItem; // Element returned by last call to next741 private int lastRet; // Index of last element returned, or -1 if none742 743 Itr() {744 final ReentrantLock lock = ArrayBlockingQueue.this.lock;745 lock.lock();746 try {747 lastRet = -1;748 if ((remaining = count) > 0)749 nextItem = itemAt(nextIndex = takeIndex);750 } finally {751 lock.unlock();752 }753 }754 755 public boolean hasNext() {756 return remaining > 0;757 }758 759 public E next() {760 final ReentrantLock lock = ArrayBlockingQueue.this.lock;761 lock.lock();762 try {763 if (remaining <= 0)764 throw new NoSuchElementException();765 lastRet = nextIndex;766 E x = itemAt(nextIndex); // check for fresher value767 if (x == null) {768 x = nextItem; // we are forced to report old value769 lastItem = null; // but ensure remove fails770 }771 else772 lastItem = x;773 while (--remaining > 0 && // skip over nulls774 (nextItem = itemAt(nextIndex = inc(nextIndex))) == null)775 ;776 return x;777 } finally {778 lock.unlock();779 }780 }781 782 public void remove() {783 final ReentrantLock lock = ArrayBlockingQueue.this.lock;784 lock.lock();785 try {786 int i = lastRet;787 if (i == -1)788 throw new IllegalStateException();789 lastRet = -1;790 E x = lastItem;791 lastItem = null;792 // only remove if item still at index793 if (x != null && x == items[i]) {794 boolean removingHead = (i == takeIndex);795 removeAt(i);796 if (!removingHead)797 nextIndex = dec(nextIndex);798 }799 } finally {800 lock.unlock();801 }802 }803 }804 805 }
下面从ArrayBlockingQueue的创建，添加，取出，遍历这几个方面对ArrayBlockingQueue进行分析。
1. 创建
下面以ArrayBlockingQueue(int capacity, boolean fair)来进行说明。
public ArrayBlockingQueue(int capacity, boolean fair) { if (capacity <= 0) throw new IllegalArgumentException(); this.items = new Object[capacity]; lock = new ReentrantLock(fair); notEmpty = lock.newCondition(); notFull = lock.newCondition();}
说明：
(01) items是保存“阻塞队列”数据的数组。它的定义如下：
final Object[] items;(02) fair是“可重入的独占锁(ReentrantLock)”的类型。fair为true，表示是公平锁；fair为false，表示是非公平锁。
notEmpty和notFull是锁的两个Condition条件。它们的定义如下：
final ReentrantLock lock;private final Condition notEmpty;private final Condition notFull;
简单对Condition和Lock的用法进行说明，更多内容请参考“Java多线程系列--“JUC锁”06之 Condition条件”。
Lock的作用是提供独占锁机制，来保护竞争资源；而Condition是为了更加精细的对锁进行控制，它依赖于Lock，通过某个条件对多线程进行控制。
notEmpty表示“锁的非空条件”。当某线程想从队列中取数据时，而此时又没有数据，则该线程通过notEmpty.await()进行等待；当其它线程向队列中插入了元素之后，就调用notEmpty.signal()唤醒“之前通过notEmpty.await()进入等待状态的线程”。
同理，notFull表示“锁的满条件”。当某线程想向队列中插入元素，而此时队列已满时，该线程等待；当其它线程从队列中取出元素之后，就唤醒该等待的线程。
2. 添加
下面以offer(E e)为例，对ArrayBlockingQueue的添加方法进行说明。
public boolean offer(E e) { // 创建插入的元素是否为null，是的话抛出NullPointerException异常 checkNotNull(e); // 获取“该阻塞队列的独占锁” final ReentrantLock lock = this.lock; lock.lock(); try { // 如果队列已满，则返回false。 if (count == items.length) return false; else { // 如果队列未满，则插入e，并返回true。 insert(e); return true; } } finally { // 释放锁 lock.unlock(); }}
说明：offer(E e)的作用是将e插入阻塞队列的尾部。如果队列已满，则返回false，表示插入失败；否则，插入元素，并返回true。
(01) count表示”队列中的元素个数“。除此之外，队列中还有另外两个遍历takeIndex和putIndex。takeIndex表示下一个被取出元素的索引，putIndex表示下一个被添加元素的索引。它们的定义如下：
// 队列中的元素个数int takeIndex;// 下一个被取出元素的索引int putIndex;// 下一个被添加元素的索引int count;
(02) insert()的源码如下：
private void insert(E x) { // 将x添加到”队列“中 items[putIndex] = x; // 设置”下一个被取出元素的索引“ putIndex = inc(putIndex); // 将”队列中的元素个数”+1 ++count; // 唤醒notEmpty上的等待线程 notEmpty.signal();}
insert()在插入元素之后，会唤醒notEmpty上面的等待线程。
inc()的源码如下：
final int inc(int i) { return (++i == items.length) ? 0 : i;}
若i+1的值等于“队列的长度”，即添加元素之后，队列满；则设置“下一个被添加元素的索引”为0。
3. 取出
下面以take()为例，对ArrayBlockingQueue的取出方法进行说明。
public E take() throws InterruptedException { // 获取“队列的独占锁” final ReentrantLock lock = this.lock; // 获取“锁”，若当前线程是中断状态，则抛出InterruptedException异常 lock.lockInterruptibly(); try { // 若“队列为空”，则一直等待。 while (count == 0) notEmpty.await(); // 取出元素 return extract(); } finally { // 释放“锁” lock.unlock(); }}
说明：take()的作用是取出并返回队列的头。若队列为空，则一直等待。
extract()的源码如下：
private E extract() { final Object[] items = this.items; // 强制将元素转换为“泛型E” E x = this.<E>cast(items[takeIndex]); // 将第takeIndex元素设为null，即删除。同时，帮助GC回收。 items[takeIndex] = null; // 设置“下一个被取出元素的索引” takeIndex = inc(takeIndex); // 将“队列中元素数量”-1 --count; // 唤醒notFull上的等待线程。 notFull.signal(); return x;}
说明：extract()在删除元素之后，会唤醒notFull上的等待线程。
4. 遍历
下面对ArrayBlockingQueue的遍历方法进行说明。
public Iterator<E> iterator() { return new Itr();}
Itr是实现了Iterator接口的类，它的源码如下：
private class Itr implements Iterator<E> { // 队列中剩余元素的个数 private int remaining; // Number of elements yet to be returned // 下一次调用next()返回的元素的索引 private int nextIndex; // Index of element to be returned by next // 下一次调用next()返回的元素 private E nextItem; // Element to be returned by next call to next // 上一次调用next()返回的元素 private E lastItem; // Element returned by last call to next // 上一次调用next()返回的元素的索引 private int lastRet; // Index of last element returned, or -1 if none Itr() { // 获取“阻塞队列”的锁 final ReentrantLock lock = ArrayBlockingQueue.this.lock; lock.lock(); try { lastRet = -1; if ((remaining = count) > 0) nextItem = itemAt(nextIndex = takeIndex); } finally { // 释放“锁” lock.unlock(); } } public boolean hasNext() { return remaining > 0; } public E next() { // 获取“阻塞队列”的锁 final ReentrantLock lock = ArrayBlockingQueue.this.lock; lock.lock(); try { // 若“剩余元素<=0”，则抛出异常。 if (remaining <= 0) throw new NoSuchElementException(); lastRet = nextIndex; // 获取第nextIndex位置的元素 E x = itemAt(nextIndex); // check for fresher value if (x == null) { x = nextItem; // we are forced to report old value lastItem = null; // but ensure remove fails } else lastItem = x; while (--remaining > 0 && // skip over nulls (nextItem = itemAt(nextIndex = inc(nextIndex))) == null) ; return x; } finally { lock.unlock(); } } public void remove() { final ReentrantLock lock = ArrayBlockingQueue.this.lock; lock.lock(); try { int i = lastRet; if (i == -1) throw new IllegalStateException(); lastRet = -1; E x = lastItem; lastItem = null; // only remove if item still at index if (x != null && x == items[i]) { boolean removingHead = (i == takeIndex); removeAt(i); if (!removingHead) nextIndex = dec(nextIndex); } } finally { lock.unlock(); } }}
ArrayBlockingQueue示例
import java.util.*;import java.util.concurrent.*;/* * ArrayBlockingQueue是“线程安全”的队列，而LinkedList是非线程安全的。 * * 下面是“多个线程同时操作并且遍历queue”的示例 * (01) 当queue是ArrayBlockingQueue对象时，程序能正常运行。 * (02) 当queue是LinkedList对象时，程序会产生ConcurrentModificationException异常。 * * @author skywang */public class ArrayBlockingQueueDemo1{ // TODO: queue是LinkedList对象时，程序会出错。 //private static Queue<String> queue = new LinkedList<String>(); private static Queue<String> queue = new ArrayBlockingQueue<String>(20); public static void main(String[] args) { // 同时启动两个线程对queue进行操作！ new MyThread("ta").start(); new MyThread("tb").start(); } private static void printAll() { String value; Iterator iter = queue.iterator(); while(iter.hasNext()) { value = (String)iter.next(); System.out.print(value+", "); } System.out.println(); } private static class MyThread extends Thread { MyThread(String name) { super(name); } @Override public void run() { int i = 0; while (i++ < 6) { // “线程名” + "-" + "序号" String val = Thread.currentThread().getName()+i; queue.add(val); // 通过“Iterator”遍历queue。 printAll(); } } }}
(某一次)运行结果：
ta1, ta1, tb1, ta1, tb1, ta1, ta2, tb1, ta1, ta2, tb1, tb2, ta2, ta1, tb2, tb1, ta3, ta2, ta1, tb2, tb1, ta3, ta2, tb3, tb2, ta1, ta3, tb1, tb3, ta2, ta4, tb2, ta1, ta3, tb1, tb3, ta2, ta4, tb2, tb4, ta3, ta1, tb3, tb1, ta4, ta2, tb4, tb2, ta5, ta3, ta1, tb3, tb1, ta4, ta2, tb4, tb2, ta5, ta3, tb5, tb3, ta1, ta4, tb1, tb4, ta2, ta5, tb2, tb5, ta3, ta6, tb3, ta4, tb4, ta5, tb5, ta6, tb6, 结果说明：如果将源码中的queue改成LinkedList对象时，程序会产生ConcurrentModificationException异常。
更多内容
1. Java多线程系列--“JUC集合”01之 框架
2. Java多线程系列目录(共xx篇)
0 0
- Java多线程系列--“JUC集合”07之 ArrayBlockingQueue
- Java多线程系列--“JUC集合”07之 ArrayBlockingQueue
- Java多线程系列--“JUC集合”07之 ArrayBlockingQueue
- JUC集合-07之 ArrayBlockingQueue
- Java多线程系列--“JUC集合”
- Java多线程系列--“JUC集合”04之 ConcurrentHashMap
- Java多线程系列--“JUC集合”01之 框架
- Java多线程系列--“JUC集合”01之 框架
- Java多线程系列--“JUC集合”02之 CopyOnWriteArrayList
- Java多线程系列--“JUC集合”03之 CopyOnWriteArraySet
- Java多线程系列--“JUC集合”04之 ConcurrentHashMap
- Java多线程系列--“JUC集合”01之 框架
- Java多线程系列--“JUC集合”02之 CopyOnWriteArrayList
- Java多线程系列--“JUC集合”03之 CopyOnWriteArraySet
- Java多线程系列--“JUC集合”04之 ConcurrentHashMap
- Java多线程系列--“JUC集合”05之 ConcurrentSkipListMap
- Java多线程系列--“JUC集合”06之 ConcurrentSkipListSet
- Java多线程系列--“JUC集合”08之 LinkedBlockingQueue
- 数组排序:快速排序,选择排序,冒泡排序,插入排序
- Python——图片转字符小工具
- 线程安全和可重入
- js函数
- Cpp_Primer--Arrays
- Java多线程系列--“JUC集合”07之 ArrayBlockingQueue
- poj 3468
- 【机器学习】神经网络（二）——反向传播算法
- [读书笔记] 深入探索C++对象模型-第二章《构造函数语义学》（上）
- Ubuntu下安装nginx与PHP-FPM两种方法
- 用线段树维护树的直径
- C语言 链表
- java中采用集合使用pull解析xml文件格式的方法
- Java源码阅读-HashMap
