Java并发包--ArrayBlockingQueue
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。
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 algorithm
100 * 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 methods
111
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 element
142 */
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 advance
181 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 queue
207 * @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 queue
218 * @param fair if {@code true} then queue accesses for threads blocked
219 * 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 the
235 * elements of the given collection,
236 * added in traversal order of the collection's iterator.
237 *
238 * @param capacity the capacity of this queue
239 * @param fair if {@code true} then queue accesses for threads blocked
240 * 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 contain
243 * @throws IllegalArgumentException if {@code capacity} is less than
244 * {@code c.size()}, or less than 1.
245 * @throws NullPointerException if the specified collection or any
246 * of its elements are null
247 */
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 exclusion
254 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 is
273 * possible to do so immediately without exceeding the queue's capacity,
274 * returning {@code true} upon success and throwing an
275 * {@code IllegalStateException} if this queue is full.
276 *
277 * @param e the element to add
278 * @return {@code true} (as specified by {@link Collection#add})
279 * @throws IllegalStateException if this queue is full
280 * @throws NullPointerException if the specified element is null
281 */
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 is
288 * possible to do so immediately without exceeding the queue's capacity,
289 * returning {@code true} upon success and {@code false} if this queue
290 * 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 null
294 */
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, waiting
313 * 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, waiting
333 * up to the specified wait time for space to become available if
334 * 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 collections
408 // greater in size than Integer.MAX_VALUE
409 /**
410 * Returns the number of elements in this queue.
411 *
412 * @return the number of elements in this queue
413 */
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 ideally
428 * (in the absence of memory or resource constraints) accept without
429 * blocking. This is always equal to the initial capacity of this queue
430 * less the current {@code size} of this queue.
431 *
432 * <p>Note that you <em>cannot</em> always tell if an attempt to insert
433 * an element will succeed by inspecting {@code remainingCapacity}
434 * because it may be the case that another thread is about to
435 * 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} such
450 * that {@code o.equals(e)}, if this queue contains one or more such
451 * elements.
452 * Returns {@code true} if this queue contained the specified element
453 * (or equivalently, if this queue changed as a result of the call).
454 *
455 * <p>Removal of interior elements in circular array based queues
456 * is an intrinsically slow and disruptive operation, so should
457 * be undertaken only in exceptional circumstances, ideally
458 * only when the queue is known not to be accessible by other
459 * threads.
460 *
461 * @param o element to be removed from this queue, if present
462 * @return {@code true} if this queue changed as a result of the call
463 */
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 contains
485 * at least one element {@code e} such that {@code o.equals(e)}.
486 *
487 * @param o object to be checked for containment in this queue
488 * @return {@code true} if this queue contains the specified element
489 */
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, in
507 * proper sequence.
508 *
509 * <p>The returned array will be "safe" in that no references to it are
510 * maintained by this queue. (In other words, this method must allocate
511 * 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-based
514 * APIs.
515 *
516 * @return an array containing all of the elements in this queue
517 */
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, in
535 * proper sequence; the runtime type of the returned array is that of
536 * the specified array. If the queue fits in the specified array, it
537 * is returned therein. Otherwise, a new array is allocated with the
538 * 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 spare
541 * (i.e., the array has more elements than this queue), the element in
542 * the array immediately following the end of the queue is set to
543 * {@code null}.
544 *
545 * <p>Like the {@link #toArray()} method, this method acts as bridge between
546 * array-based and collection-based APIs. Further, this method allows
547 * 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 newly
552 * 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 to
558 * {@code toArray()}.
559 *
560 * @param a the array into which the elements of the queue are to
561 * be stored, if it is big enough; otherwise, a new array of the
562 * same runtime type is allocated for this purpose
563 * @return an array containing all of the elements in this queue
564 * @throws ArrayStoreException if the runtime type of the specified array
565 * is not a supertype of the runtime type of every element in
566 * this queue
567 * @throws NullPointerException if the specified array is null
568 */
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 that
708 * will never throw {@link java.util.ConcurrentModificationException
709 * ConcurrentModificationException},
710 * and guarantees to traverse elements as they existed upon
711 * 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 sequence
715 */
716 public Iterator<E> iterator() {
717 return new Itr();
718 }
719
720 /**
721 * Iterator for ArrayBlockingQueue. To maintain weak consistency
722 * with respect to puts and takes, we (1) read ahead one slot, so
723 * as to not report hasNext true but then not have an element to
724 * return -- however we later recheck this slot to use the most
725 * current value; (2) ensure that each array slot is traversed at
726 * most once (by tracking "remaining" elements); (3) skip over
727 * null slots, which can occur if takes race ahead of iterators.
728 * However, for circular array-based queues, we cannot rely on any
729 * well established definition of what it means to be weakly
730 * consistent with respect to interior removes since these may
731 * require slot overwrites in the process of sliding elements to
732 * cover gaps. So we settle for resiliency, operating on
733 * established apparent nexts, which may miss some elements that
734 * have moved between calls to next.
735 */
736 private class Itr implements Iterator<E> {
737 private int remaining; // Number of elements yet to be returned
738 private int nextIndex; // Index of element to be returned by next
739 private E nextItem; // Element to be returned by next call to next
740 private E lastItem; // Element returned by last call to next
741 private int lastRet; // Index of last element returned, or -1 if none
742
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 value
767 if (x == null) {
768 x = nextItem; // we are forced to report old value
769 lastItem = null; // but ensure remove fails
770 }
771 else
772 lastItem = x;
773 while (--remaining > 0 && // skip over nulls
774 (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 index
793 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。
Java并发包--ArrayBlockingQueue的更多相关文章
- Java并发包源码学习系列：阻塞队列实现之ArrayBlockingQueue源码解析
目录 ArrayBlockingQueue概述 类图结构及重要字段 构造器 出队和入队操作 入队enqueue 出队dequeue 阻塞式操作 E take() 阻塞式获取 void put(E e) ...
- Java并发包中常用类小结(一)
从JDK1.5以后,Java为我们引入了一个并发包,用于解决实际开发中经常用到的并发问题,那我们今天就来简单看一下相关的一些常见类的使用情况. 1.ConcurrentHashMap Concurre ...
- Java并发包分析——BlockingQueue
之前因为找实习的缘故,博客1个多月没有写了.找实习的经历总算告一段落,现在重新更新博客,这次的内容是分析Java并发包中的阻塞队列 关于阻塞队列,我之前是一直充满好奇,很好奇这个阻塞是怎么实现.现在我 ...
- 死磕 java集合之ArrayBlockingQueue源码分析
问题 (1)ArrayBlockingQueue的实现方式? (2)ArrayBlockingQueue是否需要扩容? (3)ArrayBlockingQueue有什么缺点? 简介 ArrayBloc ...
- java并发包&线程池原理分析&锁的深度化
java并发包&线程池原理分析&锁的深度化 并发包 同步容器类 Vector与ArrayList区别 1.ArrayList是最常用的List实现类,内部是通过数组实现的, ...
- Java并发包——Blockingqueue，ConcurrentLinkedQueue，Executors
背景 通过做以下一个小的接口系统gate,了解一下mina和java并发包里的东西.A系统为javaweb项目,B为C语言项目,gate是本篇须要完毕的系统. 需求 1. A为集群系统,并发较高,会批 ...
- Java并发包——线程安全的Collection相关类
Java并发包——线程安全的Collection相关类 摘要:本文主要学习了Java并发包下线程安全的Collection相关的类. 部分内容来自以下博客: https://www.cnblogs.c ...
- Java并发包——线程池
Java并发包——线程池 摘要:本文主要学习了Java并发包中的线程池. 部分内容来自以下博客: https://www.cnblogs.com/dolphin0520/p/3932921.html ...
- Java并发编程（您不知道的线程池操作）， 最受欢迎的 8 位 Java 大师，Java并发包中的同步队列SynchronousQueue实现原理
Java_并发编程培训 java并发程序设计教程 JUC Exchanger 一.概述 Exchanger 可以在对中对元素进行配对和交换的线程的同步点.每个线程将条目上的某个方法呈现给 exchan ...
随机推荐
- [转帖]如何禁止VirtualBox虚拟机和物理机之间的时间同步
如何禁止VirtualBox虚拟机和物理机之间的时间同步 乐者为王 2014-11-02 706 阅读 http://ju.outofmemory.cn/entry/97301 验证了下 修改注册 ...
- Docker部署Gitlab11.10.4
1.下载镜像 官方镜像地址:https://hub.docker.com/r/gitlab/gitlab-ce ,根据自己需要下载指定版本 [root@vanje-dev01 ~]# docker p ...
- Centos7.x 安装libevent2.x
1.在http://libevent.org/下载libevent-2.1.8-stable.tar.gz 2.tar -zxvf libevent-2.1.8-stable.tar.gz 3.cd ...
- 剑指offer58：对称的二叉树。判断一颗二叉树是不是对称的，如果一个二叉树同此二叉树的镜像是同样的，定义其为对称的
1 题目描述 请实现一个函数,用来判断一颗二叉树是不是对称的.注意,如果一个二叉树同此二叉树的镜像是同样的,定义其为对称的. 2 思路和方法 定义一种遍历算法,先遍历右子结点再遍历左子结点:如对称先序 ...
- WUSTOJ 1276: 峰峰不搞G（Java）
1276: 峰峰不搞G 题目 给 n 数量的油漆,写出最大的数,每个数对应有油漆的花费.更多内容点击标题. 分析 我读完题,就想到用动态规划,结果是Time Limit Exceed.然后看了 ...
- Error:Could not find method google() for arguments [] on repository container
Error:Could not find method google() for arguments [] on repository container. Consult IDE log for m ...
- redis5.0集群配置
介绍 redis自3.0版本以来支持主从模式的集群,可用哨兵监控集群健康状态,但这种方式的集群很不成熟,数据备份需要全量拷贝.在之后的版本才真正支持集群分片. 在redis5.0中去除了以redis- ...
- Java Web 深入分析（8） Servlet工作原理解析
Servlet Servlet(Server Applet)是Java Servlet的简称,称为小服务程序或服务连接器,用Java编写的服务器端程序,主要功能在于交互式地浏览和修改数据,生成动态We ...
- php 环境搭建问题
项目过程中需要用到 PHP环境 https://www.cnblogs.com/cyrfr/p/6483529.html APACHE无法启动:THE REQUEST OPERATION HAS FA ...
- INTEL_BIOS 编译—for-ATOM_E3800
INTEL_BIOS 编译—for-ATOM_E3800 ======================================================================= ...