ViewVC Help

View File | Revision Log | Show Annotations | Download File | Root Listing

root/jsr166/jsr166/src/main/java/util/concurrent/ConcurrentHashMap.java

Comparing jsr166/src/main/java/util/concurrent/ConcurrentHashMap.java (file contents):
Revision 1.83 by jsr166, Mon Aug 22 03:42:10 2005 UTC vs.
Revision 1.141 by jsr166, Tue Oct 30 16:54:26 2012 UTC

#	Line 1 | Line 1
1		/*
2		* Written by Doug Lea with assistance from members of JCP JSR-166
3		* Expert Group and released to the public domain, as explained at
4	<	* http://creativecommons.org/licenses/publicdomain
4	>	* http://creativecommons.org/publicdomain/zero/1.0/
5		*/
6
7		package java.util.concurrent;
8	<	import java.util.concurrent.locks.*;
9	<	import java.util.*;
8	>	import java.util.concurrent.atomic.LongAdder;
9	>	import java.util.concurrent.ForkJoinPool;
10	>	import java.util.concurrent.ForkJoinTask;
11	>
12	>	import java.util.Comparator;
13	>	import java.util.Arrays;
14	>	import java.util.Map;
15	>	import java.util.Set;
16	>	import java.util.Collection;
17	>	import java.util.AbstractMap;
18	>	import java.util.AbstractSet;
19	>	import java.util.AbstractCollection;
20	>	import java.util.Hashtable;
21	>	import java.util.HashMap;
22	>	import java.util.Iterator;
23	>	import java.util.Enumeration;
24	>	import java.util.ConcurrentModificationException;
25	>	import java.util.NoSuchElementException;
26	>	import java.util.concurrent.ConcurrentMap;
27	>	import java.util.concurrent.ThreadLocalRandom;
28	>	import java.util.concurrent.locks.LockSupport;
29	>	import java.util.concurrent.locks.AbstractQueuedSynchronizer;
30	>	import java.util.concurrent.atomic.AtomicReference;
31	>
32		import java.io.Serializable;
11	–	import java.io.IOException;
12	–	import java.io.ObjectInputStream;
13	–	import java.io.ObjectOutputStream;
33
34		/**
35		* A hash table supporting full concurrency of retrievals and
36	<	* adjustable expected concurrency for updates. This class obeys the
36	>	* high expected concurrency for updates. This class obeys the
37		* same functional specification as {@link java.util.Hashtable}, and
38		* includes versions of methods corresponding to each method of
39	<	* <tt>Hashtable</tt>. However, even though all operations are
39	>	* {@code Hashtable}. However, even though all operations are
40		* thread-safe, retrieval operations do <em>not</em> entail locking,
41		* and there is <em>not</em> any support for locking the entire table
42		* in a way that prevents all access.  This class is fully
43	<	* interoperable with <tt>Hashtable</tt> in programs that rely on its
43	>	* interoperable with {@code Hashtable} in programs that rely on its
44		* thread safety but not on its synchronization details.
45		*
46	<	* <p> Retrieval operations (including <tt>get</tt>) generally do not
47	<	* block, so may overlap with update operations (including
48	<	* <tt>put</tt> and <tt>remove</tt>). Retrievals reflect the results
49	<	* of the most recently <em>completed</em> update operations holding
50	<	* upon their onset.  For aggregate operations such as <tt>putAll</tt>
51	<	* and <tt>clear</tt>, concurrent retrievals may reflect insertion or
52	<	* removal of only some entries.  Similarly, Iterators and
53	<	* Enumerations return elements reflecting the state of the hash table
54	<	* at some point at or since the creation of the iterator/enumeration.
55	<	* They do <em>not</em> throw {@link ConcurrentModificationException}.
56	<	* However, iterators are designed to be used by only one thread at a time.
57	<	*
58	<	* <p> The allowed concurrency among update operations is guided by
59	<	* the optional <tt>concurrencyLevel</tt> constructor argument
60	<	* (default <tt>16</tt>), which is used as a hint for internal sizing.  The
61	<	* table is internally partitioned to try to permit the indicated
62	<	* number of concurrent updates without contention. Because placement
63	<	* in hash tables is essentially random, the actual concurrency will
64	<	* vary.  Ideally, you should choose a value to accommodate as many
65	<	* threads as will ever concurrently modify the table. Using a
66	<	* significantly higher value than you need can waste space and time,
67	<	* and a significantly lower value can lead to thread contention. But
68	<	* overestimates and underestimates within an order of magnitude do
69	<	* not usually have much noticeable impact. A value of one is
70	<	* appropriate when it is known that only one thread will modify and
71	<	* all others will only read. Also, resizing this or any other kind of
72	<	* hash table is a relatively slow operation, so, when possible, it is
73	<	* a good idea to provide estimates of expected table sizes in
74	<	* constructors.
46	>	* <p> Retrieval operations (including {@code get}) generally do not
47	>	* block, so may overlap with update operations (including {@code put}
48	>	* and {@code remove}). Retrievals reflect the results of the most
49	>	* recently <em>completed</em> update operations holding upon their
50	>	* onset. (More formally, an update operation for a given key bears a
51	>	* <em>happens-before</em> relation with any (non-null) retrieval for
52	>	* that key reporting the updated value.)  For aggregate operations
53	>	* such as {@code putAll} and {@code clear}, concurrent retrievals may
54	>	* reflect insertion or removal of only some entries.  Similarly,
55	>	* Iterators and Enumerations return elements reflecting the state of
56	>	* the hash table at some point at or since the creation of the
57	>	* iterator/enumeration.  They do <em>not</em> throw {@link
58	>	* ConcurrentModificationException}.  However, iterators are designed
59	>	* to be used by only one thread at a time.  Bear in mind that the
60	>	* results of aggregate status methods including {@code size}, {@code
61	>	* isEmpty}, and {@code containsValue} are typically useful only when
62	>	* a map is not undergoing concurrent updates in other threads.
63	>	* Otherwise the results of these methods reflect transient states
64	>	* that may be adequate for monitoring or estimation purposes, but not
65	>	* for program control.
66	>	*
67	>	* <p> The table is dynamically expanded when there are too many
68	>	* collisions (i.e., keys that have distinct hash codes but fall into
69	>	* the same slot modulo the table size), with the expected average
70	>	* effect of maintaining roughly two bins per mapping (corresponding
71	>	* to a 0.75 load factor threshold for resizing). There may be much
72	>	* variance around this average as mappings are added and removed, but
73	>	* overall, this maintains a commonly accepted time/space tradeoff for
74	>	* hash tables.  However, resizing this or any other kind of hash
75	>	* table may be a relatively slow operation. When possible, it is a
76	>	* good idea to provide a size estimate as an optional {@code
77	>	* initialCapacity} constructor argument. An additional optional
78	>	* {@code loadFactor} constructor argument provides a further means of
79	>	* customizing initial table capacity by specifying the table density
80	>	* to be used in calculating the amount of space to allocate for the
81	>	* given number of elements.  Also, for compatibility with previous
82	>	* versions of this class, constructors may optionally specify an
83	>	* expected {@code concurrencyLevel} as an additional hint for
84	>	* internal sizing.  Note that using many keys with exactly the same
85	>	* {@code hashCode()} is a sure way to slow down performance of any
86	>	* hash table.
87	>	*
88	>	* <p> A {@link Set} projection of a ConcurrentHashMap may be created
89	>	* (using {@link #newKeySet()} or {@link #newKeySet(int)}), or viewed
90	>	* (using {@link #keySet(Object)} when only keys are of interest, and the
91	>	* mapped values are (perhaps transiently) not used or all take the
92	>	* same mapping value.
93	>	*
94	>	* <p> A ConcurrentHashMap can be used as scalable frequency map (a
95	>	* form of histogram or multiset) by using {@link LongAdder} values
96	>	* and initializing via {@link #computeIfAbsent}. For example, to add
97	>	* a count to a {@code ConcurrentHashMap<String,LongAdder> freqs}, you
98	>	* can use {@code freqs.computeIfAbsent(k -> new
99	>	* LongAdder()).increment();}
100		*
101		* <p>This class and its views and iterators implement all of the
102		* <em>optional</em> methods of the {@link Map} and {@link Iterator}
103		* interfaces.
104		*
105		* <p> Like {@link Hashtable} but unlike {@link HashMap}, this class
106	<	* does <em>not</em> allow <tt>null</tt> to be used as a key or value.
106	>	* does <em>not</em> allow {@code null} to be used as a key or value.
107	>	*
108	>	* <p>ConcurrentHashMaps support parallel operations using the {@link
109	>	* ForkJoinPool#commonPool}. (Tasks that may be used in other contexts
110	>	* are available in class {@link ForkJoinTasks}). These operations are
111	>	* designed to be safely, and often sensibly, applied even with maps
112	>	* that are being concurrently updated by other threads; for example,
113	>	* when computing a snapshot summary of the values in a shared
114	>	* registry.  There are three kinds of operation, each with four
115	>	* forms, accepting functions with Keys, Values, Entries, and (Key,
116	>	* Value) arguments and/or return values. Because the elements of a
117	>	* ConcurrentHashMap are not ordered in any particular way, and may be
118	>	* processed in different orders in different parallel executions, the
119	>	* correctness of supplied functions should not depend on any
120	>	* ordering, or on any other objects or values that may transiently
121	>	* change while computation is in progress; and except for forEach
122	>	* actions, should ideally be side-effect-free.
123	>	*
124	>	* <ul>
125	>	* <li> forEach: Perform a given action on each element.
126	>	* A variant form applies a given transformation on each element
127	>	* before performing the action.</li>
128	>	*
129	>	* <li> search: Return the first available non-null result of
130	>	* applying a given function on each element; skipping further
131	>	* search when a result is found.</li>
132	>	*
133	>	* <li> reduce: Accumulate each element.  The supplied reduction
134	>	* function cannot rely on ordering (more formally, it should be
135	>	* both associative and commutative).  There are five variants:
136	>	*
137	>	* <ul>
138	>	*
139	>	* <li> Plain reductions. (There is not a form of this method for
140	>	* (key, value) function arguments since there is no corresponding
141	>	* return type.)</li>
142	>	*
143	>	* <li> Mapped reductions that accumulate the results of a given
144	>	* function applied to each element.</li>
145	>	*
146	>	* <li> Reductions to scalar doubles, longs, and ints, using a
147	>	* given basis value.</li>
148	>	*
149	>	* </li>
150	>	* </ul>
151	>	* </ul>
152	>	*
153	>	* <p>The concurrency properties of bulk operations follow
154	>	* from those of ConcurrentHashMap: Any non-null result returned
155	>	* from {@code get(key)} and related access methods bears a
156	>	* happens-before relation with the associated insertion or
157	>	* update.  The result of any bulk operation reflects the
158	>	* composition of these per-element relations (but is not
159	>	* necessarily atomic with respect to the map as a whole unless it
160	>	* is somehow known to be quiescent).  Conversely, because keys
161	>	* and values in the map are never null, null serves as a reliable
162	>	* atomic indicator of the current lack of any result.  To
163	>	* maintain this property, null serves as an implicit basis for
164	>	* all non-scalar reduction operations. For the double, long, and
165	>	* int versions, the basis should be one that, when combined with
166	>	* any other value, returns that other value (more formally, it
167	>	* should be the identity element for the reduction). Most common
168	>	* reductions have these properties; for example, computing a sum
169	>	* with basis 0 or a minimum with basis MAX_VALUE.
170	>	*
171	>	* <p>Search and transformation functions provided as arguments
172	>	* should similarly return null to indicate the lack of any result
173	>	* (in which case it is not used). In the case of mapped
174	>	* reductions, this also enables transformations to serve as
175	>	* filters, returning null (or, in the case of primitive
176	>	* specializations, the identity basis) if the element should not
177	>	* be combined. You can create compound transformations and
178	>	* filterings by composing them yourself under this "null means
179	>	* there is nothing there now" rule before using them in search or
180	>	* reduce operations.
181	>	*
182	>	* <p>Methods accepting and/or returning Entry arguments maintain
183	>	* key-value associations. They may be useful for example when
184	>	* finding the key for the greatest value. Note that "plain" Entry
185	>	* arguments can be supplied using {@code new
186	>	* AbstractMap.SimpleEntry(k,v)}.
187	>	*
188	>	* <p> Bulk operations may complete abruptly, throwing an
189	>	* exception encountered in the application of a supplied
190	>	* function. Bear in mind when handling such exceptions that other
191	>	* concurrently executing functions could also have thrown
192	>	* exceptions, or would have done so if the first exception had
193	>	* not occurred.
194	>	*
195	>	* <p>Parallel speedups for bulk operations compared to sequential
196	>	* processing are common but not guaranteed.  Operations involving
197	>	* brief functions on small maps may execute more slowly than
198	>	* sequential loops if the underlying work to parallelize the
199	>	* computation is more expensive than the computation itself.
200	>	* Similarly, parallelization may not lead to much actual parallelism
201	>	* if all processors are busy performing unrelated tasks.
202	>	*
203	>	* <p> All arguments to all task methods must be non-null.
204	>	*
205	>	* <p><em>jsr166e note: During transition, this class
206	>	* uses nested functional interfaces with different names but the
207	>	* same forms as those expected for JDK8.<em>
208		*
209		* <p>This class is a member of the
210	<	* <a href="{@docRoot}/../guide/collections/index.html">
210	>	* <a href="{@docRoot}/../technotes/guides/collections/index.html">
211		* Java Collections Framework</a>.
212		*
213		* @since 1.5
#	Line 70 | Line 215 | import java.io.ObjectOutputStream;
215		* @param <K> the type of keys maintained by this map
216		* @param <V> the type of mapped values
217		*/
218	<	public class ConcurrentHashMap<K, V> extends AbstractMap<K, V>
219	<	implements ConcurrentMap<K, V>, Serializable {
218	>	public class ConcurrentHashMap<K, V>
219	>	implements ConcurrentMap<K, V>, Serializable {
220		private static final long serialVersionUID = 7249069246763182397L;
221
222	+	/**
223	+	* A partitionable iterator. A Spliterator can be traversed
224	+	* directly, but can also be partitioned (before traversal) by
225	+	* creating another Spliterator that covers a non-overlapping
226	+	* portion of the elements, and so may be amenable to parallel
227	+	* execution.
228	+	*
229	+	* <p> This interface exports a subset of expected JDK8
230	+	* functionality.
231	+	*
232	+	* <p>Sample usage: Here is one (of the several) ways to compute
233	+	* the sum of the values held in a map using the ForkJoin
234	+	* framework. As illustrated here, Spliterators are well suited to
235	+	* designs in which a task repeatedly splits off half its work
236	+	* into forked subtasks until small enough to process directly,
237	+	* and then joins these subtasks. Variants of this style can also
238	+	* be used in completion-based designs.
239	+	*
240	+	* <pre>
241	+	* {@code ConcurrentHashMap<String, Long> m = ...
242	+	* // split as if have 8 * parallelism, for load balance
243	+	* int n = m.size();
244	+	* int p = aForkJoinPool.getParallelism() * 8;
245	+	* int split = (n < p)? n : p;
246	+	* long sum = aForkJoinPool.invoke(new SumValues(m.valueSpliterator(), split, null));
247	+	* // ...
248	+	* static class SumValues extends RecursiveTask<Long> {
249	+	*   final Spliterator<Long> s;
250	+	*   final int split;             // split while > 1
251	+	*   final SumValues nextJoin;    // records forked subtasks to join
252	+	*   SumValues(Spliterator<Long> s, int depth, SumValues nextJoin) {
253	+	*     this.s = s; this.depth = depth; this.nextJoin = nextJoin;
254	+	*   }
255	+	*   public Long compute() {
256	+	*     long sum = 0;
257	+	*     SumValues subtasks = null; // fork subtasks
258	+	*     for (int s = split >>> 1; s > 0; s >>>= 1)
259	+	*       (subtasks = new SumValues(s.split(), s, subtasks)).fork();
260	+	*     while (s.hasNext())        // directly process remaining elements
261	+	*       sum += s.next();
262	+	*     for (SumValues t = subtasks; t != null; t = t.nextJoin)
263	+	*       sum += t.join();         // collect subtask results
264	+	*     return sum;
265	+	*   }
266	+	* }
267	+	* }</pre>
268	+	*/
269	+	public static interface Spliterator<T> extends Iterator<T> {
270	+	/**
271	+	* Returns a Spliterator covering approximately half of the
272	+	* elements, guaranteed not to overlap with those subsequently
273	+	* returned by this Spliterator.  After invoking this method,
274	+	* the current Spliterator will <em>not</em> produce any of
275	+	* the elements of the returned Spliterator, but the two
276	+	* Spliterators together will produce all of the elements that
277	+	* would have been produced by this Spliterator had this
278	+	* method not been called. The exact number of elements
279	+	* produced by the returned Spliterator is not guaranteed, and
280	+	* may be zero (i.e., with {@code hasNext()} reporting {@code
281	+	* false}) if this Spliterator cannot be further split.
282	+	*
283	+	* @return a Spliterator covering approximately half of the
284	+	* elements
285	+	* @throws IllegalStateException if this Spliterator has
286	+	* already commenced traversing elements
287	+	*/
288	+	Spliterator<T> split();
289	+	}
290	+
291	+	/**
292	+	* A view of a ConcurrentHashMap as a {@link Set} of keys, in
293	+	* which additions may optionally be enabled by mapping to a
294	+	* common value.  This class cannot be directly instantiated. See
295	+	* {@link #keySet}, {@link #keySet(Object)}, {@link #newKeySet()},
296	+	* {@link #newKeySet(int)}.
297	+	*
298	+	* <p>The view's {@code iterator} is a "weakly consistent" iterator
299	+	* that will never throw {@link ConcurrentModificationException},
300	+	* and guarantees to traverse elements as they existed upon
301	+	* construction of the iterator, and may (but is not guaranteed to)
302	+	* reflect any modifications subsequent to construction.
303	+	*/
304	+	public static class KeySetView<K,V> extends CHMView<K,V> implements Set<K>, java.io.Serializable {
305	+	private static final long serialVersionUID = 7249069246763182397L;
306	+	private final V value;
307	+	KeySetView(ConcurrentHashMap<K, V> map, V value) {  // non-public
308	+	super(map);
309	+	this.value = value;
310	+	}
311	+
312	+	/**
313	+	* Returns the map backing this view.
314	+	*
315	+	* @return the map backing this view
316	+	*/
317	+	public ConcurrentHashMap<K,V> getMap() { return map; }
318	+
319	+	/**
320	+	* Returns the default mapped value for additions,
321	+	* or {@code null} if additions are not supported.
322	+	*
323	+	* @return the default mapped value for additions, or {@code null}
324	+	* if not supported.
325	+	*/
326	+	public V getMappedValue() { return value; }
327	+
328	+	// implement Set API
329	+
330	+	public boolean contains(Object o) { return map.containsKey(o); }
331	+	public boolean remove(Object o)   { return map.remove(o) != null; }
332	+	public Iterator<K> iterator()     { return new KeyIterator<K,V>(map); }
333	+	public boolean add(K e) {
334	+	V v;
335	+	if ((v = value) == null)
336	+	throw new UnsupportedOperationException();
337	+	if (e == null)
338	+	throw new NullPointerException();
339	+	return map.internalPutIfAbsent(e, v) == null;
340	+	}
341	+	public boolean addAll(Collection<? extends K> c) {
342	+	boolean added = false;
343	+	V v;
344	+	if ((v = value) == null)
345	+	throw new UnsupportedOperationException();
346	+	for (K e : c) {
347	+	if (e == null)
348	+	throw new NullPointerException();
349	+	if (map.internalPutIfAbsent(e, v) == null)
350	+	added = true;
351	+	}
352	+	return added;
353	+	}
354	+	public boolean equals(Object o) {
355	+	Set<?> c;
356	+	return ((o instanceof Set) &&
357	+	((c = (Set<?>)o) == this ||
358	+	(containsAll(c) && c.containsAll(this))));
359	+	}
360	+	}
361	+
362		/*
363	<	* The basic strategy is to subdivide the table among Segments,
364	<	* each of which itself is a concurrently readable hash table.
363	>	* Overview:
364	>	*
365	>	* The primary design goal of this hash table is to maintain
366	>	* concurrent readability (typically method get(), but also
367	>	* iterators and related methods) while minimizing update
368	>	* contention. Secondary goals are to keep space consumption about
369	>	* the same or better than java.util.HashMap, and to support high
370	>	* initial insertion rates on an empty table by many threads.
371	>	*
372	>	* Each key-value mapping is held in a Node.  Because Node fields
373	>	* can contain special values, they are defined using plain Object
374	>	* types. Similarly in turn, all internal methods that use them
375	>	* work off Object types. And similarly, so do the internal
376	>	* methods of auxiliary iterator and view classes.  All public
377	>	* generic typed methods relay in/out of these internal methods,
378	>	* supplying null-checks and casts as needed. This also allows
379	>	* many of the public methods to be factored into a smaller number
380	>	* of internal methods (although sadly not so for the five
381	>	* variants of put-related operations). The validation-based
382	>	* approach explained below leads to a lot of code sprawl because
383	>	* retry-control precludes factoring into smaller methods.
384	>	*
385	>	* The table is lazily initialized to a power-of-two size upon the
386	>	* first insertion.  Each bin in the table normally contains a
387	>	* list of Nodes (most often, the list has only zero or one Node).
388	>	* Table accesses require volatile/atomic reads, writes, and
389	>	* CASes.  Because there is no other way to arrange this without
390	>	* adding further indirections, we use intrinsics
391	>	* (sun.misc.Unsafe) operations.  The lists of nodes within bins
392	>	* are always accurately traversable under volatile reads, so long
393	>	* as lookups check hash code and non-nullness of value before
394	>	* checking key equality.
395	>	*
396	>	* We use the top two bits of Node hash fields for control
397	>	* purposes -- they are available anyway because of addressing
398	>	* constraints.  As explained further below, these top bits are
399	>	* used as follows:
400	>	*  00 - Normal
401	>	*  01 - Locked
402	>	*  11 - Locked and may have a thread waiting for lock
403	>	*  10 - Node is a forwarding node
404	>	*
405	>	* The lower 30 bits of each Node's hash field contain a
406	>	* transformation of the key's hash code, except for forwarding
407	>	* nodes, for which the lower bits are zero (and so always have
408	>	* hash field == MOVED).
409	>	*
410	>	* Insertion (via put or its variants) of the first node in an
411	>	* empty bin is performed by just CASing it to the bin.  This is
412	>	* by far the most common case for put operations under most
413	>	* key/hash distributions.  Other update operations (insert,
414	>	* delete, and replace) require locks.  We do not want to waste
415	>	* the space required to associate a distinct lock object with
416	>	* each bin, so instead use the first node of a bin list itself as
417	>	* a lock. Blocking support for these locks relies on the builtin
418	>	* "synchronized" monitors.  However, we also need a tryLock
419	>	* construction, so we overlay these by using bits of the Node
420	>	* hash field for lock control (see above), and so normally use
421	>	* builtin monitors only for blocking and signalling using
422	>	* wait/notifyAll constructions. See Node.tryAwaitLock.
423	>	*
424	>	* Using the first node of a list as a lock does not by itself
425	>	* suffice though: When a node is locked, any update must first
426	>	* validate that it is still the first node after locking it, and
427	>	* retry if not. Because new nodes are always appended to lists,
428	>	* once a node is first in a bin, it remains first until deleted
429	>	* or the bin becomes invalidated (upon resizing).  However,
430	>	* operations that only conditionally update may inspect nodes
431	>	* until the point of update. This is a converse of sorts to the
432	>	* lazy locking technique described by Herlihy & Shavit.
433	>	*
434	>	* The main disadvantage of per-bin locks is that other update
435	>	* operations on other nodes in a bin list protected by the same
436	>	* lock can stall, for example when user equals() or mapping
437	>	* functions take a long time.  However, statistically, under
438	>	* random hash codes, this is not a common problem.  Ideally, the
439	>	* frequency of nodes in bins follows a Poisson distribution
440	>	* (http://en.wikipedia.org/wiki/Poisson_distribution) with a
441	>	* parameter of about 0.5 on average, given the resizing threshold
442	>	* of 0.75, although with a large variance because of resizing
443	>	* granularity. Ignoring variance, the expected occurrences of
444	>	* list size k are (exp(-0.5) * pow(0.5, k) / factorial(k)). The
445	>	* first values are:
446	>	*
447	>	* 0:    0.60653066
448	>	* 1:    0.30326533
449	>	* 2:    0.07581633
450	>	* 3:    0.01263606
451	>	* 4:    0.00157952
452	>	* 5:    0.00015795
453	>	* 6:    0.00001316
454	>	* 7:    0.00000094
455	>	* 8:    0.00000006
456	>	* more: less than 1 in ten million
457	>	*
458	>	* Lock contention probability for two threads accessing distinct
459	>	* elements is roughly 1 / (8 * #elements) under random hashes.
460	>	*
461	>	* Actual hash code distributions encountered in practice
462	>	* sometimes deviate significantly from uniform randomness.  This
463	>	* includes the case when N > (1<<30), so some keys MUST collide.
464	>	* Similarly for dumb or hostile usages in which multiple keys are
465	>	* designed to have identical hash codes. Also, although we guard
466	>	* against the worst effects of this (see method spread), sets of
467	>	* hashes may differ only in bits that do not impact their bin
468	>	* index for a given power-of-two mask.  So we use a secondary
469	>	* strategy that applies when the number of nodes in a bin exceeds
470	>	* a threshold, and at least one of the keys implements
471	>	* Comparable.  These TreeBins use a balanced tree to hold nodes
472	>	* (a specialized form of red-black trees), bounding search time
473	>	* to O(log N).  Each search step in a TreeBin is around twice as
474	>	* slow as in a regular list, but given that N cannot exceed
475	>	* (1<<64) (before running out of addresses) this bounds search
476	>	* steps, lock hold times, etc, to reasonable constants (roughly
477	>	* 100 nodes inspected per operation worst case) so long as keys
478	>	* are Comparable (which is very common -- String, Long, etc).
479	>	* TreeBin nodes (TreeNodes) also maintain the same "next"
480	>	* traversal pointers as regular nodes, so can be traversed in
481	>	* iterators in the same way.
482	>	*
483	>	* The table is resized when occupancy exceeds a percentage
484	>	* threshold (nominally, 0.75, but see below).  Only a single
485	>	* thread performs the resize (using field "sizeCtl", to arrange
486	>	* exclusion), but the table otherwise remains usable for reads
487	>	* and updates. Resizing proceeds by transferring bins, one by
488	>	* one, from the table to the next table.  Because we are using
489	>	* power-of-two expansion, the elements from each bin must either
490	>	* stay at same index, or move with a power of two offset. We
491	>	* eliminate unnecessary node creation by catching cases where old
492	>	* nodes can be reused because their next fields won't change.  On
493	>	* average, only about one-sixth of them need cloning when a table
494	>	* doubles. The nodes they replace will be garbage collectable as
495	>	* soon as they are no longer referenced by any reader thread that
496	>	* may be in the midst of concurrently traversing table.  Upon
497	>	* transfer, the old table bin contains only a special forwarding
498	>	* node (with hash field "MOVED") that contains the next table as
499	>	* its key. On encountering a forwarding node, access and update
500	>	* operations restart, using the new table.
501	>	*
502	>	* Each bin transfer requires its bin lock. However, unlike other
503	>	* cases, a transfer can skip a bin if it fails to acquire its
504	>	* lock, and revisit it later (unless it is a TreeBin). Method
505	>	* rebuild maintains a buffer of TRANSFER_BUFFER_SIZE bins that
506	>	* have been skipped because of failure to acquire a lock, and
507	>	* blocks only if none are available (i.e., only very rarely).
508	>	* The transfer operation must also ensure that all accessible
509	>	* bins in both the old and new table are usable by any traversal.
510	>	* When there are no lock acquisition failures, this is arranged
511	>	* simply by proceeding from the last bin (table.length - 1) up
512	>	* towards the first.  Upon seeing a forwarding node, traversals
513	>	* (see class Iter) arrange to move to the new table
514	>	* without revisiting nodes.  However, when any node is skipped
515	>	* during a transfer, all earlier table bins may have become
516	>	* visible, so are initialized with a reverse-forwarding node back
517	>	* to the old table until the new ones are established. (This
518	>	* sometimes requires transiently locking a forwarding node, which
519	>	* is possible under the above encoding.) These more expensive
520	>	* mechanics trigger only when necessary.
521	>	*
522	>	* The traversal scheme also applies to partial traversals of
523	>	* ranges of bins (via an alternate Traverser constructor)
524	>	* to support partitioned aggregate operations.  Also, read-only
525	>	* operations give up if ever forwarded to a null table, which
526	>	* provides support for shutdown-style clearing, which is also not
527	>	* currently implemented.
528	>	*
529	>	* Lazy table initialization minimizes footprint until first use,
530	>	* and also avoids resizings when the first operation is from a
531	>	* putAll, constructor with map argument, or deserialization.
532	>	* These cases attempt to override the initial capacity settings,
533	>	* but harmlessly fail to take effect in cases of races.
534	>	*
535	>	* The element count is maintained using a LongAdder, which avoids
536	>	* contention on updates but can encounter cache thrashing if read
537	>	* too frequently during concurrent access. To avoid reading so
538	>	* often, resizing is attempted either when a bin lock is
539	>	* contended, or upon adding to a bin already holding two or more
540	>	* nodes (checked before adding in the xIfAbsent methods, after
541	>	* adding in others). Under uniform hash distributions, the
542	>	* probability of this occurring at threshold is around 13%,
543	>	* meaning that only about 1 in 8 puts check threshold (and after
544	>	* resizing, many fewer do so). But this approximation has high
545	>	* variance for small table sizes, so we check on any collision
546	>	* for sizes <= 64. The bulk putAll operation further reduces
547	>	* contention by only committing count updates upon these size
548	>	* checks.
549	>	*
550	>	* Maintaining API and serialization compatibility with previous
551	>	* versions of this class introduces several oddities. Mainly: We
552	>	* leave untouched but unused constructor arguments refering to
553	>	* concurrencyLevel. We accept a loadFactor constructor argument,
554	>	* but apply it only to initial table capacity (which is the only
555	>	* time that we can guarantee to honor it.) We also declare an
556	>	* unused "Segment" class that is instantiated in minimal form
557	>	* only when serializing.
558		*/
559
560		/* ---------------- Constants -------------- */
561
562		/**
563	<	* The default initial capacity for this table,
564	<	* used when not otherwise specified in a constructor.
563	>	* The largest possible table capacity.  This value must be
564	>	* exactly 1<<30 to stay within Java array allocation and indexing
565	>	* bounds for power of two table sizes, and is further required
566	>	* because the top two bits of 32bit hash fields are used for
567	>	* control purposes.
568	>	*/
569	>	private static final int MAXIMUM_CAPACITY = 1 << 30;
570	>
571	>	/**
572	>	* The default initial table capacity.  Must be a power of 2
573	>	* (i.e., at least 1) and at most MAXIMUM_CAPACITY.
574		*/
575	<	static final int DEFAULT_INITIAL_CAPACITY = 16;
575	>	private static final int DEFAULT_CAPACITY = 16;
576
577		/**
578	<	* The default load factor for this table, used when not
579	<	* otherwise specified in a constructor.
578	>	* The largest possible (non-power of two) array size.
579	>	* Needed by toArray and related methods.
580		*/
581	<	static final float DEFAULT_LOAD_FACTOR = 0.75f;
581	>	static final int MAX_ARRAY_SIZE = Integer.MAX_VALUE - 8;
582
583		/**
584	<	* The default concurrency level for this table, used when not
585	<	* otherwise specified in a constructor.
584	>	* The default concurrency level for this table. Unused but
585	>	* defined for compatibility with previous versions of this class.
586		*/
587	<	static final int DEFAULT_CONCURRENCY_LEVEL = 16;
587	>	private static final int DEFAULT_CONCURRENCY_LEVEL = 16;
588
589		/**
590	<	* The maximum capacity, used if a higher value is implicitly
591	<	* specified by either of the constructors with arguments.  MUST
592	<	* be a power of two <= 1<<30 to ensure that entries are indexable
593	<	* using ints.
590	>	* The load factor for this table. Overrides of this value in
591	>	* constructors affect only the initial table capacity.  The
592	>	* actual floating point value isn't normally used -- it is
593	>	* simpler to use expressions such as {@code n - (n >>> 2)} for
594	>	* the associated resizing threshold.
595		*/
596	<	static final int MAXIMUM_CAPACITY = 1 << 30;
596	>	private static final float LOAD_FACTOR = 0.75f;
597
598		/**
599	<	* The maximum number of segments to allow; used to bound
600	<	* constructor arguments.
599	>	* The buffer size for skipped bins during transfers. The
600	>	* value is arbitrary but should be large enough to avoid
601	>	* most locking stalls during resizes.
602		*/
603	<	static final int MAX_SEGMENTS = 1 << 16; // slightly conservative
603	>	private static final int TRANSFER_BUFFER_SIZE = 32;
604
605		/**
606	<	* Number of unsynchronized retries in size and containsValue
607	<	* methods before resorting to locking. This is used to avoid
608	<	* unbounded retries if tables undergo continuous modification
609	<	* which would make it impossible to obtain an accurate result.
606	>	* The bin count threshold for using a tree rather than list for a
607	>	* bin.  The value reflects the approximate break-even point for
608	>	* using tree-based operations.
609	>	*/
610	>	private static final int TREE_THRESHOLD = 8;
611	>
612	>	/*
613	>	* Encodings for special uses of Node hash fields. See above for
614	>	* explanation.
615		*/
616	<	static final int RETRIES_BEFORE_LOCK = 2;
616	>	static final int MOVED     = 0x80000000; // hash field for forwarding nodes
617	>	static final int LOCKED    = 0x40000000; // set/tested only as a bit
618	>	static final int WAITING   = 0xc0000000; // both bits set/tested together
619	>	static final int HASH_BITS = 0x3fffffff; // usable bits of normal node hash
620
621		/* ---------------- Fields -------------- */
622
623		/**
624	<	* Mask value for indexing into segments. The upper bits of a
625	<	* key's hash code are used to choose the segment.
624	>	* The array of bins. Lazily initialized upon first insertion.
625	>	* Size is always a power of two. Accessed directly by iterators.
626		*/
627	<	final int segmentMask;
627	>	transient volatile Node[] table;
628
629		/**
630	<	* Shift value for indexing within segments.
630	>	* The counter maintaining number of elements.
631		*/
632	<	final int segmentShift;
632	>	private transient final LongAdder counter;
633
634		/**
635	<	* The segments, each of which is a specialized hash table
636	<	*/
637	<	final Segment<K,V>[] segments;
635	>	* Table initialization and resizing control.  When negative, the
636	>	* table is being initialized or resized. Otherwise, when table is
637	>	* null, holds the initial table size to use upon creation, or 0
638	>	* for default. After initialization, holds the next element count
639	>	* value upon which to resize the table.
640	>	*/
641	>	private transient volatile int sizeCtl;
642	>
643	>	// views
644	>	private transient KeySetView<K,V> keySet;
645	>	private transient Values<K,V> values;
646	>	private transient EntrySet<K,V> entrySet;
647
648	<	transient Set<K> keySet;
649	<	transient Set<Map.Entry<K,V>> entrySet;
144	<	transient Collection<V> values;
648	>	/** For serialization compatibility. Null unless serialized; see below */
649	>	private Segment<K,V>[] segments;
650
651	<	/* ---------------- Small Utilities -------------- */
651	>	/* ---------------- Table element access -------------- */
652
653	<	/**
654	<	* Returns a hash code for non-null Object x.
655	<	* Uses the same hash code spreader as most other java.util hash tables.
656	<	* @param x the object serving as a key
657	<	* @return the hash code
653	>	/*
654	>	* Volatile access methods are used for table elements as well as
655	>	* elements of in-progress next table while resizing.  Uses are
656	>	* null checked by callers, and implicitly bounds-checked, relying
657	>	* on the invariants that tab arrays have non-zero size, and all
658	>	* indices are masked with (tab.length - 1) which is never
659	>	* negative and always less than length. Note that, to be correct
660	>	* wrt arbitrary concurrency errors by users, bounds checks must
661	>	* operate on local variables, which accounts for some odd-looking
662	>	* inline assignments below.
663		*/
664	<	static int hash(Object x) {
665	<	int h = x.hashCode();
666	<	h += ~(h << 9);
157	<	h ^=  (h >>> 14);
158	<	h +=  (h << 4);
159	<	h ^=  (h >>> 10);
160	<	return h;
664	>
665	>	static final Node tabAt(Node[] tab, int i) { // used by Iter
666	>	return (Node)UNSAFE.getObjectVolatile(tab, ((long)i<<ASHIFT)+ABASE);
667		}
668
669	<	/**
670	<	* Returns the segment that should be used for key with given hash
165	<	* @param hash the hash code for the key
166	<	* @return the segment
167	<	*/
168	<	final Segment<K,V> segmentFor(int hash) {
169	<	return segments[(hash >>> segmentShift) & segmentMask];
669	>	private static final boolean casTabAt(Node[] tab, int i, Node c, Node v) {
670	>	return UNSAFE.compareAndSwapObject(tab, ((long)i<<ASHIFT)+ABASE, c, v);
671		}
672
673	<	/* ---------------- Inner Classes -------------- */
673	>	private static final void setTabAt(Node[] tab, int i, Node v) {
674	>	UNSAFE.putObjectVolatile(tab, ((long)i<<ASHIFT)+ABASE, v);
675	>	}
676	>
677	>	/* ---------------- Nodes -------------- */
678
679		/**
680	<	* ConcurrentHashMap list entry. Note that this is never exported
681	<	* out as a user-visible Map.Entry.
682	<	*
683	<	* Because the value field is volatile, not final, it is legal wrt
684	<	* the Java Memory Model for an unsynchronized reader to see null
685	<	* instead of initial value when read via a data race.  Although a
686	<	* reordering leading to this is not likely to ever actually
687	<	* occur, the Segment.readValueUnderLock method is used as a
688	<	* backup in case a null (pre-initialized) value is ever seen in
689	<	* an unsynchronized access method.
690	<	*/
691	<	static final class HashEntry<K,V> {
692	<	final K key;
693	<	final int hash;
189	<	volatile V value;
190	<	final HashEntry<K,V> next;
680	>	* Key-value entry. Note that this is never exported out as a
681	>	* user-visible Map.Entry (see MapEntry below). Nodes with a hash
682	>	* field of MOVED are special, and do not contain user keys or
683	>	* values.  Otherwise, keys are never null, and null val fields
684	>	* indicate that a node is in the process of being deleted or
685	>	* created. For purposes of read-only access, a key may be read
686	>	* before a val, but can only be used after checking val to be
687	>	* non-null.
688	>	*/
689	>	static class Node {
690	>	volatile int hash;
691	>	final Object key;
692	>	volatile Object val;
693	>	volatile Node next;
694
695	<	HashEntry(K key, int hash, HashEntry<K,V> next, V value) {
193	<	this.key = key;
695	>	Node(int hash, Object key, Object val, Node next) {
696		this.hash = hash;
697	+	this.key = key;
698	+	this.val = val;
699		this.next = next;
196	–	this.value = value;
700		}
701
702	<	@SuppressWarnings("unchecked")
703	<	static final <K,V> HashEntry<K,V>[] newArray(int i) {
704	<	return new HashEntry[i];
705	<	}
203	<	}
204	<
205	<	/**
206	<	* Segments are specialized versions of hash tables.  This
207	<	* subclasses from ReentrantLock opportunistically, just to
208	<	* simplify some locking and avoid separate construction.
209	<	*/
210	<	static final class Segment<K,V> extends ReentrantLock implements Serializable {
211	<	/*
212	<	* Segments maintain a table of entry lists that are ALWAYS
213	<	* kept in a consistent state, so can be read without locking.
214	<	* Next fields of nodes are immutable (final).  All list
215	<	* additions are performed at the front of each bin. This
216	<	* makes it easy to check changes, and also fast to traverse.
217	<	* When nodes would otherwise be changed, new nodes are
218	<	* created to replace them. This works well for hash tables
219	<	* since the bin lists tend to be short. (The average length
220	<	* is less than two for the default load factor threshold.)
221	<	*
222	<	* Read operations can thus proceed without locking, but rely
223	<	* on selected uses of volatiles to ensure that completed
224	<	* write operations performed by other threads are
225	<	* noticed. For most purposes, the "count" field, tracking the
226	<	* number of elements, serves as that volatile variable
227	<	* ensuring visibility.  This is convenient because this field
228	<	* needs to be read in many read operations anyway:
229	<	*
230	<	*   - All (unsynchronized) read operations must first read the
231	<	*     "count" field, and should not look at table entries if
232	<	*     it is 0.
233	<	*
234	<	*   - All (synchronized) write operations should write to
235	<	*     the "count" field after structurally changing any bin.
236	<	*     The operations must not take any action that could even
237	<	*     momentarily cause a concurrent read operation to see
238	<	*     inconsistent data. This is made easier by the nature of
239	<	*     the read operations in Map. For example, no operation
240	<	*     can reveal that the table has grown but the threshold
241	<	*     has not yet been updated, so there are no atomicity
242	<	*     requirements for this with respect to reads.
243	<	*
244	<	* As a guide, all critical volatile reads and writes to the
245	<	* count field are marked in code comments.
246	<	*/
702	>	/** CompareAndSet the hash field */
703	>	final boolean casHash(int cmp, int val) {
704	>	return UNSAFE.compareAndSwapInt(this, hashOffset, cmp, val);
705	>	}
706
707	<	private static final long serialVersionUID = 2249069246763182397L;
707	>	/** The number of spins before blocking for a lock */
708	>	static final int MAX_SPINS =
709	>	Runtime.getRuntime().availableProcessors() > 1 ? 64 : 1;
710
711		/**
712	<	* The number of elements in this segment's region.
712	>	* Spins a while if LOCKED bit set and this node is the first
713	>	* of its bin, and then sets WAITING bits on hash field and
714	>	* blocks (once) if they are still set.  It is OK for this
715	>	* method to return even if lock is not available upon exit,
716	>	* which enables these simple single-wait mechanics.
717	>	*
718	>	* The corresponding signalling operation is performed within
719	>	* callers: Upon detecting that WAITING has been set when
720	>	* unlocking lock (via a failed CAS from non-waiting LOCKED
721	>	* state), unlockers acquire the sync lock and perform a
722	>	* notifyAll.
723	>	*
724	>	* The initial sanity check on tab and bounds is not currently
725	>	* necessary in the only usages of this method, but enables
726	>	* use in other future contexts.
727		*/
728	<	transient volatile int count;
728	>	final void tryAwaitLock(Node[] tab, int i) {
729	>	if (tab != null && i >= 0 && i < tab.length) { // sanity check
730	>	int r = ThreadLocalRandom.current().nextInt(); // randomize spins
731	>	int spins = MAX_SPINS, h;
732	>	while (tabAt(tab, i) == this && ((h = hash) & LOCKED) != 0) {
733	>	if (spins >= 0) {
734	>	r ^= r << 1; r ^= r >>> 3; r ^= r << 10; // xorshift
735	>	if (r >= 0 && --spins == 0)
736	>	Thread.yield();  // yield before block
737	>	}
738	>	else if (casHash(h, h | WAITING)) {
739	>	synchronized (this) {
740	>	if (tabAt(tab, i) == this &&
741	>	(hash & WAITING) == WAITING) {
742	>	try {
743	>	wait();
744	>	} catch (InterruptedException ie) {
745	>	Thread.currentThread().interrupt();
746	>	}
747	>	}
748	>	else
749	>	notifyAll(); // possibly won race vs signaller
750	>	}
751	>	break;
752	>	}
753	>	}
754	>	}
755	>	}
756
757	<	/**
758	<	* Number of updates that alter the size of the table. This is
759	<	* used during bulk-read methods to make sure they see a
258	<	* consistent snapshot: If modCounts change during a traversal
259	<	* of segments computing size or checking containsValue, then
260	<	* we might have an inconsistent view of state so (usually)
261	<	* must retry.
262	<	*/
263	<	transient int modCount;
757	>	// Unsafe mechanics for casHash
758	>	private static final sun.misc.Unsafe UNSAFE;
759	>	private static final long hashOffset;
760
761	<	/**
762	<	* The table is rehashed when its size exceeds this threshold.
763	<	* (The value of this field is always <tt>(int)(capacity *
764	<	* loadFactor)</tt>.)
765	<	*/
766	<	transient int threshold;
761	>	static {
762	>	try {
763	>	UNSAFE = sun.misc.Unsafe.getUnsafe();
764	>	Class<?> k = Node.class;
765	>	hashOffset = UNSAFE.objectFieldOffset
766	>	(k.getDeclaredField("hash"));
767	>	} catch (Exception e) {
768	>	throw new Error(e);
769	>	}
770	>	}
771	>	}
772
773	<	/**
273	<	* The per-segment table.
274	<	*/
275	<	transient volatile HashEntry<K,V>[] table;
773	>	/* ---------------- TreeBins -------------- */
774
775	<	/**
776	<	* The load factor for the hash table.  Even though this value
777	<	* is same for all segments, it is replicated to avoid needing
778	<	* links to outer object.
779	<	* @serial
780	<	*/
781	<	final float loadFactor;
775	>	/**
776	>	* Nodes for use in TreeBins
777	>	*/
778	>	static final class TreeNode extends Node {
779	>	TreeNode parent;  // red-black tree links
780	>	TreeNode left;
781	>	TreeNode right;
782	>	TreeNode prev;    // needed to unlink next upon deletion
783	>	boolean red;
784	>
785	>	TreeNode(int hash, Object key, Object val, Node next, TreeNode parent) {
786	>	super(hash, key, val, next);
787	>	this.parent = parent;
788	>	}
789	>	}
790
791	<	Segment(int initialCapacity, float lf) {
792	<	loadFactor = lf;
793	<	setTable(HashEntry.<K,V>newArray(initialCapacity));
791	>	/**
792	>	* A specialized form of red-black tree for use in bins
793	>	* whose size exceeds a threshold.
794	>	*
795	>	* TreeBins use a special form of comparison for search and
796	>	* related operations (which is the main reason we cannot use
797	>	* existing collections such as TreeMaps). TreeBins contain
798	>	* Comparable elements, but may contain others, as well as
799	>	* elements that are Comparable but not necessarily Comparable<T>
800	>	* for the same T, so we cannot invoke compareTo among them. To
801	>	* handle this, the tree is ordered primarily by hash value, then
802	>	* by getClass().getName() order, and then by Comparator order
803	>	* among elements of the same class.  On lookup at a node, if
804	>	* elements are not comparable or compare as 0, both left and
805	>	* right children may need to be searched in the case of tied hash
806	>	* values. (This corresponds to the full list search that would be
807	>	* necessary if all elements were non-Comparable and had tied
808	>	* hashes.)  The red-black balancing code is updated from
809	>	* pre-jdk-collections
810	>	* (http://gee.cs.oswego.edu/dl/classes/collections/RBCell.java)
811	>	* based in turn on Cormen, Leiserson, and Rivest "Introduction to
812	>	* Algorithms" (CLR).
813	>	*
814	>	* TreeBins also maintain a separate locking discipline than
815	>	* regular bins. Because they are forwarded via special MOVED
816	>	* nodes at bin heads (which can never change once established),
817	>	* we cannot use those nodes as locks. Instead, TreeBin
818	>	* extends AbstractQueuedSynchronizer to support a simple form of
819	>	* read-write lock. For update operations and table validation,
820	>	* the exclusive form of lock behaves in the same way as bin-head
821	>	* locks. However, lookups use shared read-lock mechanics to allow
822	>	* multiple readers in the absence of writers.  Additionally,
823	>	* these lookups do not ever block: While the lock is not
824	>	* available, they proceed along the slow traversal path (via
825	>	* next-pointers) until the lock becomes available or the list is
826	>	* exhausted, whichever comes first. (These cases are not fast,
827	>	* but maximize aggregate expected throughput.)  The AQS mechanics
828	>	* for doing this are straightforward.  The lock state is held as
829	>	* AQS getState().  Read counts are negative; the write count (1)
830	>	* is positive.  There are no signalling preferences among readers
831	>	* and writers. Since we don't need to export full Lock API, we
832	>	* just override the minimal AQS methods and use them directly.
833	>	*/
834	>	static final class TreeBin extends AbstractQueuedSynchronizer {
835	>	private static final long serialVersionUID = 2249069246763182397L;
836	>	transient TreeNode root;  // root of tree
837	>	transient TreeNode first; // head of next-pointer list
838	>
839	>	/* AQS overrides */
840	>	public final boolean isHeldExclusively() { return getState() > 0; }
841	>	public final boolean tryAcquire(int ignore) {
842	>	if (compareAndSetState(0, 1)) {
843	>	setExclusiveOwnerThread(Thread.currentThread());
844	>	return true;
845	>	}
846	>	return false;
847	>	}
848	>	public final boolean tryRelease(int ignore) {
849	>	setExclusiveOwnerThread(null);
850	>	setState(0);
851	>	return true;
852	>	}
853	>	public final int tryAcquireShared(int ignore) {
854	>	for (int c;;) {
855	>	if ((c = getState()) > 0)
856	>	return -1;
857	>	if (compareAndSetState(c, c -1))
858	>	return 1;
859	>	}
860	>	}
861	>	public final boolean tryReleaseShared(int ignore) {
862	>	int c;
863	>	do {} while (!compareAndSetState(c = getState(), c + 1));
864	>	return c == -1;
865	>	}
866	>
867	>	/** From CLR */
868	>	private void rotateLeft(TreeNode p) {
869	>	if (p != null) {
870	>	TreeNode r = p.right, pp, rl;
871	>	if ((rl = p.right = r.left) != null)
872	>	rl.parent = p;
873	>	if ((pp = r.parent = p.parent) == null)
874	>	root = r;
875	>	else if (pp.left == p)
876	>	pp.left = r;
877	>	else
878	>	pp.right = r;
879	>	r.left = p;
880	>	p.parent = r;
881	>	}
882		}
883
884	<	@SuppressWarnings("unchecked")
885	<	static final <K,V> Segment<K,V>[] newArray(int i) {
886	<	return new Segment[i];
884	>	/** From CLR */
885	>	private void rotateRight(TreeNode p) {
886	>	if (p != null) {
887	>	TreeNode l = p.left, pp, lr;
888	>	if ((lr = p.left = l.right) != null)
889	>	lr.parent = p;
890	>	if ((pp = l.parent = p.parent) == null)
891	>	root = l;
892	>	else if (pp.right == p)
893	>	pp.right = l;
894	>	else
895	>	pp.left = l;
896	>	l.right = p;
897	>	p.parent = l;
898	>	}
899		}
900
901		/**
902	<	* Sets table to new HashEntry array.
903	<	* Call only while holding lock or in constructor.
902	>	* Returns the TreeNode (or null if not found) for the given key
903	>	* starting at given root.
904		*/
905	<	void setTable(HashEntry<K,V>[] newTable) {
906	<	threshold = (int)(newTable.length * loadFactor);
907	<	table = newTable;
905	>	@SuppressWarnings("unchecked") final TreeNode getTreeNode
906	>	(int h, Object k, TreeNode p) {
907	>	Class<?> c = k.getClass();
908	>	while (p != null) {
909	>	int dir, ph;  Object pk; Class<?> pc;
910	>	if ((ph = p.hash) == h) {
911	>	if ((pk = p.key) == k || k.equals(pk))
912	>	return p;
913	>	if (c != (pc = pk.getClass()) ||
914	>	!(k instanceof Comparable) ||
915	>	(dir = ((Comparable)k).compareTo((Comparable)pk)) == 0) {
916	>	dir = (c == pc) ? 0 : c.getName().compareTo(pc.getName());
917	>	TreeNode r = null, s = null, pl, pr;
918	>	if (dir >= 0) {
919	>	if ((pl = p.left) != null && h <= pl.hash)
920	>	s = pl;
921	>	}
922	>	else if ((pr = p.right) != null && h >= pr.hash)
923	>	s = pr;
924	>	if (s != null && (r = getTreeNode(h, k, s)) != null)
925	>	return r;
926	>	}
927	>	}
928	>	else
929	>	dir = (h < ph) ? -1 : 1;
930	>	p = (dir > 0) ? p.right : p.left;
931	>	}
932	>	return null;
933		}
934
935		/**
936	<	* Returns properly casted first entry of bin for given hash.
936	>	* Wrapper for getTreeNode used by CHM.get. Tries to obtain
937	>	* read-lock to call getTreeNode, but during failure to get
938	>	* lock, searches along next links.
939		*/
940	<	HashEntry<K,V> getFirst(int hash) {
941	<	HashEntry<K,V>[] tab = table;
942	<	return tab[hash & (tab.length - 1)];
940	>	final Object getValue(int h, Object k) {
941	>	Node r = null;
942	>	int c = getState(); // Must read lock state first
943	>	for (Node e = first; e != null; e = e.next) {
944	>	if (c <= 0 && compareAndSetState(c, c - 1)) {
945	>	try {
946	>	r = getTreeNode(h, k, root);
947	>	} finally {
948	>	releaseShared(0);
949	>	}
950	>	break;
951	>	}
952	>	else if ((e.hash & HASH_BITS) == h && k.equals(e.key)) {
953	>	r = e;
954	>	break;
955	>	}
956	>	else
957	>	c = getState();
958	>	}
959	>	return r == null ? null : r.val;
960		}
961
962		/**
963	<	* Reads value field of an entry under lock. Called if value
964	<	* field ever appears to be null. This is possible only if a
315	<	* compiler happens to reorder a HashEntry initialization with
316	<	* its table assignment, which is legal under memory model
317	<	* but is not known to ever occur.
963	>	* Finds or adds a node.
964	>	* @return null if added
965		*/
966	<	V readValueUnderLock(HashEntry<K,V> e) {
967	<	lock();
968	<	try {
969	<	return e.value;
970	<	} finally {
971	<	unlock();
966	>	@SuppressWarnings("unchecked") final TreeNode putTreeNode
967	>	(int h, Object k, Object v) {
968	>	Class<?> c = k.getClass();
969	>	TreeNode pp = root, p = null;
970	>	int dir = 0;
971	>	while (pp != null) { // find existing node or leaf to insert at
972	>	int ph;  Object pk; Class<?> pc;
973	>	p = pp;
974	>	if ((ph = p.hash) == h) {
975	>	if ((pk = p.key) == k || k.equals(pk))
976	>	return p;
977	>	if (c != (pc = pk.getClass()) ||
978	>	!(k instanceof Comparable) ||
979	>	(dir = ((Comparable)k).compareTo((Comparable)pk)) == 0) {
980	>	dir = (c == pc) ? 0 : c.getName().compareTo(pc.getName());
981	>	TreeNode r = null, s = null, pl, pr;
982	>	if (dir >= 0) {
983	>	if ((pl = p.left) != null && h <= pl.hash)
984	>	s = pl;
985	>	}
986	>	else if ((pr = p.right) != null && h >= pr.hash)
987	>	s = pr;
988	>	if (s != null && (r = getTreeNode(h, k, s)) != null)
989	>	return r;
990	>	}
991	>	}
992	>	else
993	>	dir = (h < ph) ? -1 : 1;
994	>	pp = (dir > 0) ? p.right : p.left;
995		}
326	–	}
327	–
328	–	/* Specialized implementations of map methods */
996
997	<	V get(Object key, int hash) {
998	<	if (count != 0) { // read-volatile
999	<	HashEntry<K,V> e = getFirst(hash);
1000	<	while (e != null) {
1001	<	if (e.hash == hash && key.equals(e.key)) {
1002	<	V v = e.value;
1003	<	if (v != null)
1004	<	return v;
1005	<	return readValueUnderLock(e); // recheck
997	>	TreeNode f = first;
998	>	TreeNode x = first = new TreeNode(h, k, v, f, p);
999	>	if (p == null)
1000	>	root = x;
1001	>	else { // attach and rebalance; adapted from CLR
1002	>	TreeNode xp, xpp;
1003	>	if (f != null)
1004	>	f.prev = x;
1005	>	if (dir <= 0)
1006	>	p.left = x;
1007	>	else
1008	>	p.right = x;
1009	>	x.red = true;
1010	>	while (x != null && (xp = x.parent) != null && xp.red &&
1011	>	(xpp = xp.parent) != null) {
1012	>	TreeNode xppl = xpp.left;
1013	>	if (xp == xppl) {
1014	>	TreeNode y = xpp.right;
1015	>	if (y != null && y.red) {
1016	>	y.red = false;
1017	>	xp.red = false;
1018	>	xpp.red = true;
1019	>	x = xpp;
1020	>	}
1021	>	else {
1022	>	if (x == xp.right) {
1023	>	rotateLeft(x = xp);
1024	>	xpp = (xp = x.parent) == null ? null : xp.parent;
1025	>	}
1026	>	if (xp != null) {
1027	>	xp.red = false;
1028	>	if (xpp != null) {
1029	>	xpp.red = true;
1030	>	rotateRight(xpp);
1031	>	}
1032	>	}
1033	>	}
1034	>	}
1035	>	else {
1036	>	TreeNode y = xppl;
1037	>	if (y != null && y.red) {
1038	>	y.red = false;
1039	>	xp.red = false;
1040	>	xpp.red = true;
1041	>	x = xpp;
1042	>	}
1043	>	else {
1044	>	if (x == xp.left) {
1045	>	rotateRight(x = xp);
1046	>	xpp = (xp = x.parent) == null ? null : xp.parent;
1047	>	}
1048	>	if (xp != null) {
1049	>	xp.red = false;
1050	>	if (xpp != null) {
1051	>	xpp.red = true;
1052	>	rotateLeft(xpp);
1053	>	}
1054	>	}
1055	>	}
1056		}
340	–	e = e.next;
1057		}
1058	+	TreeNode r = root;
1059	+	if (r != null && r.red)
1060	+	r.red = false;
1061		}
1062		return null;
1063		}
1064
1065	<	boolean containsKey(Object key, int hash) {
1066	<	if (count != 0) { // read-volatile
1067	<	HashEntry<K,V> e = getFirst(hash);
1068	<	while (e != null) {
1069	<	if (e.hash == hash && key.equals(e.key))
1070	<	return true;
1071	<	e = e.next;
1065	>	/**
1066	>	* Removes the given node, that must be present before this
1067	>	* call.  This is messier than typical red-black deletion code
1068	>	* because we cannot swap the contents of an interior node
1069	>	* with a leaf successor that is pinned by "next" pointers
1070	>	* that are accessible independently of lock. So instead we
1071	>	* swap the tree linkages.
1072	>	*/
1073	>	final void deleteTreeNode(TreeNode p) {
1074	>	TreeNode next = (TreeNode)p.next; // unlink traversal pointers
1075	>	TreeNode pred = p.prev;
1076	>	if (pred == null)
1077	>	first = next;
1078	>	else
1079	>	pred.next = next;
1080	>	if (next != null)
1081	>	next.prev = pred;
1082	>	TreeNode replacement;
1083	>	TreeNode pl = p.left;
1084	>	TreeNode pr = p.right;
1085	>	if (pl != null && pr != null) {
1086	>	TreeNode s = pr, sl;
1087	>	while ((sl = s.left) != null) // find successor
1088	>	s = sl;
1089	>	boolean c = s.red; s.red = p.red; p.red = c; // swap colors
1090	>	TreeNode sr = s.right;
1091	>	TreeNode pp = p.parent;
1092	>	if (s == pr) { // p was s's direct parent
1093	>	p.parent = s;
1094	>	s.right = p;
1095		}
1096	+	else {
1097	+	TreeNode sp = s.parent;
1098	+	if ((p.parent = sp) != null) {
1099	+	if (s == sp.left)
1100	+	sp.left = p;
1101	+	else
1102	+	sp.right = p;
1103	+	}
1104	+	if ((s.right = pr) != null)
1105	+	pr.parent = s;
1106	+	}
1107	+	p.left = null;
1108	+	if ((p.right = sr) != null)
1109	+	sr.parent = p;
1110	+	if ((s.left = pl) != null)
1111	+	pl.parent = s;
1112	+	if ((s.parent = pp) == null)
1113	+	root = s;
1114	+	else if (p == pp.left)
1115	+	pp.left = s;
1116	+	else
1117	+	pp.right = s;
1118	+	replacement = sr;
1119	+	}
1120	+	else
1121	+	replacement = (pl != null) ? pl : pr;
1122	+	TreeNode pp = p.parent;
1123	+	if (replacement == null) {
1124	+	if (pp == null) {
1125	+	root = null;
1126	+	return;
1127	+	}
1128	+	replacement = p;
1129	+	}
1130	+	else {
1131	+	replacement.parent = pp;
1132	+	if (pp == null)
1133	+	root = replacement;
1134	+	else if (p == pp.left)
1135	+	pp.left = replacement;
1136	+	else
1137	+	pp.right = replacement;
1138	+	p.left = p.right = p.parent = null;
1139	+	}
1140	+	if (!p.red) { // rebalance, from CLR
1141	+	TreeNode x = replacement;
1142	+	while (x != null) {
1143	+	TreeNode xp, xpl;
1144	+	if (x.red || (xp = x.parent) == null) {
1145	+	x.red = false;
1146	+	break;
1147	+	}
1148	+	if (x == (xpl = xp.left)) {
1149	+	TreeNode sib = xp.right;
1150	+	if (sib != null && sib.red) {
1151	+	sib.red = false;
1152	+	xp.red = true;
1153	+	rotateLeft(xp);
1154	+	sib = (xp = x.parent) == null ? null : xp.right;
1155	+	}
1156	+	if (sib == null)
1157	+	x = xp;
1158	+	else {
1159	+	TreeNode sl = sib.left, sr = sib.right;
1160	+	if ((sr == null || !sr.red) &&
1161	+	(sl == null || !sl.red)) {
1162	+	sib.red = true;
1163	+	x = xp;
1164	+	}
1165	+	else {
1166	+	if (sr == null || !sr.red) {
1167	+	if (sl != null)
1168	+	sl.red = false;
1169	+	sib.red = true;
1170	+	rotateRight(sib);
1171	+	sib = (xp = x.parent) == null ? null : xp.right;
1172	+	}
1173	+	if (sib != null) {
1174	+	sib.red = (xp == null) ? false : xp.red;
1175	+	if ((sr = sib.right) != null)
1176	+	sr.red = false;
1177	+	}
1178	+	if (xp != null) {
1179	+	xp.red = false;
1180	+	rotateLeft(xp);
1181	+	}
1182	+	x = root;
1183	+	}
1184	+	}
1185	+	}
1186	+	else { // symmetric
1187	+	TreeNode sib = xpl;
1188	+	if (sib != null && sib.red) {
1189	+	sib.red = false;
1190	+	xp.red = true;
1191	+	rotateRight(xp);
1192	+	sib = (xp = x.parent) == null ? null : xp.left;
1193	+	}
1194	+	if (sib == null)
1195	+	x = xp;
1196	+	else {
1197	+	TreeNode sl = sib.left, sr = sib.right;
1198	+	if ((sl == null || !sl.red) &&
1199	+	(sr == null || !sr.red)) {
1200	+	sib.red = true;
1201	+	x = xp;
1202	+	}
1203	+	else {
1204	+	if (sl == null || !sl.red) {
1205	+	if (sr != null)
1206	+	sr.red = false;
1207	+	sib.red = true;
1208	+	rotateLeft(sib);
1209	+	sib = (xp = x.parent) == null ? null : xp.left;
1210	+	}
1211	+	if (sib != null) {
1212	+	sib.red = (xp == null) ? false : xp.red;
1213	+	if ((sl = sib.left) != null)
1214	+	sl.red = false;
1215	+	}
1216	+	if (xp != null) {
1217	+	xp.red = false;
1218	+	rotateRight(xp);
1219	+	}
1220	+	x = root;
1221	+	}
1222	+	}
1223	+	}
1224	+	}
1225	+	}
1226	+	if (p == replacement && (pp = p.parent) != null) {
1227	+	if (p == pp.left) // detach pointers
1228	+	pp.left = null;
1229	+	else if (p == pp.right)
1230	+	pp.right = null;
1231	+	p.parent = null;
1232		}
355	–	return false;
1233		}
1234	+	}
1235	+
1236	+	/* ---------------- Collision reduction methods -------------- */
1237	+
1238	+	/**
1239	+	* Spreads higher bits to lower, and also forces top 2 bits to 0.
1240	+	* Because the table uses power-of-two masking, sets of hashes
1241	+	* that vary only in bits above the current mask will always
1242	+	* collide. (Among known examples are sets of Float keys holding
1243	+	* consecutive whole numbers in small tables.)  To counter this,
1244	+	* we apply a transform that spreads the impact of higher bits
1245	+	* downward. There is a tradeoff between speed, utility, and
1246	+	* quality of bit-spreading. Because many common sets of hashes
1247	+	* are already reasonably distributed across bits (so don't benefit
1248	+	* from spreading), and because we use trees to handle large sets
1249	+	* of collisions in bins, we don't need excessively high quality.
1250	+	*/
1251	+	private static final int spread(int h) {
1252	+	h ^= (h >>> 18) ^ (h >>> 12);
1253	+	return (h ^ (h >>> 10)) & HASH_BITS;
1254	+	}
1255	+
1256	+	/**
1257	+	* Replaces a list bin with a tree bin. Call only when locked.
1258	+	* Fails to replace if the given key is non-comparable or table
1259	+	* is, or needs, resizing.
1260	+	*/
1261	+	private final void replaceWithTreeBin(Node[] tab, int index, Object key) {
1262	+	if ((key instanceof Comparable) &&
1263	+	(tab.length >= MAXIMUM_CAPACITY || counter.sum() < (long)sizeCtl)) {
1264	+	TreeBin t = new TreeBin();
1265	+	for (Node e = tabAt(tab, index); e != null; e = e.next)
1266	+	t.putTreeNode(e.hash & HASH_BITS, e.key, e.val);
1267	+	setTabAt(tab, index, new Node(MOVED, t, null, null));
1268	+	}
1269	+	}
1270	+
1271	+	/* ---------------- Internal access and update methods -------------- */
1272
1273	<	boolean containsValue(Object value) {
1274	<	if (count != 0) { // read-volatile
1275	<	HashEntry<K,V>[] tab = table;
1276	<	int len = tab.length;
1277	<	for (int i = 0 ; i < len; i++) {
1278	<	for (HashEntry<K,V> e = tab[i]; e != null; e = e.next) {
1279	<	V v = e.value;
1280	<	if (v == null) // recheck
1281	<	v = readValueUnderLock(e);
1282	<	if (value.equals(v))
1283	<	return true;
1273	>	/** Implementation for get and containsKey */
1274	>	private final Object internalGet(Object k) {
1275	>	int h = spread(k.hashCode());
1276	>	retry: for (Node[] tab = table; tab != null;) {
1277	>	Node e, p; Object ek, ev; int eh;      // locals to read fields once
1278	>	for (e = tabAt(tab, (tab.length - 1) & h); e != null; e = e.next) {
1279	>	if ((eh = e.hash) == MOVED) {
1280	>	if ((ek = e.key) instanceof TreeBin)  // search TreeBin
1281	>	return ((TreeBin)ek).getValue(h, k);
1282	>	else {                        // restart with new table
1283	>	tab = (Node[])ek;
1284	>	continue retry;
1285		}
1286		}
1287	+	else if ((eh & HASH_BITS) == h && (ev = e.val) != null &&
1288	+	((ek = e.key) == k || k.equals(ek)))
1289	+	return ev;
1290		}
1291	<	return false;
1291	>	break;
1292		}
1293	+	return null;
1294	+	}
1295
1296	<	boolean replace(K key, int hash, V oldValue, V newValue) {
1297	<	lock();
1298	<	try {
1299	<	HashEntry<K,V> e = getFirst(hash);
1300	<	while (e != null && (e.hash != hash || !key.equals(e.key)))
1301	<	e = e.next;
1302	<
1303	<	boolean replaced = false;
1304	<	if (e != null && oldValue.equals(e.value)) {
1305	<	replaced = true;
1306	<	e.value = newValue;
1296	>	/**
1297	>	* Implementation for the four public remove/replace methods:
1298	>	* Replaces node value with v, conditional upon match of cv if
1299	>	* non-null.  If resulting value is null, delete.
1300	>	*/
1301	>	private final Object internalReplace(Object k, Object v, Object cv) {
1302	>	int h = spread(k.hashCode());
1303	>	Object oldVal = null;
1304	>	for (Node[] tab = table;;) {
1305	>	Node f; int i, fh; Object fk;
1306	>	if (tab == null ||
1307	>	(f = tabAt(tab, i = (tab.length - 1) & h)) == null)
1308	>	break;
1309	>	else if ((fh = f.hash) == MOVED) {
1310	>	if ((fk = f.key) instanceof TreeBin) {
1311	>	TreeBin t = (TreeBin)fk;
1312	>	boolean validated = false;
1313	>	boolean deleted = false;
1314	>	t.acquire(0);
1315	>	try {
1316	>	if (tabAt(tab, i) == f) {
1317	>	validated = true;
1318	>	TreeNode p = t.getTreeNode(h, k, t.root);
1319	>	if (p != null) {
1320	>	Object pv = p.val;
1321	>	if (cv == null || cv == pv || cv.equals(pv)) {
1322	>	oldVal = pv;
1323	>	if ((p.val = v) == null) {
1324	>	deleted = true;
1325	>	t.deleteTreeNode(p);
1326	>	}
1327	>	}
1328	>	}
1329	>	}
1330	>	} finally {
1331	>	t.release(0);
1332	>	}
1333	>	if (validated) {
1334	>	if (deleted)
1335	>	counter.add(-1L);
1336	>	break;
1337	>	}
1338	>	}
1339	>	else
1340	>	tab = (Node[])fk;
1341	>	}
1342	>	else if ((fh & HASH_BITS) != h && f.next == null) // precheck
1343	>	break;                          // rules out possible existence
1344	>	else if ((fh & LOCKED) != 0) {
1345	>	checkForResize();               // try resizing if can't get lock
1346	>	f.tryAwaitLock(tab, i);
1347	>	}
1348	>	else if (f.casHash(fh, fh | LOCKED)) {
1349	>	boolean validated = false;
1350	>	boolean deleted = false;
1351	>	try {
1352	>	if (tabAt(tab, i) == f) {
1353	>	validated = true;
1354	>	for (Node e = f, pred = null;;) {
1355	>	Object ek, ev;
1356	>	if ((e.hash & HASH_BITS) == h &&
1357	>	((ev = e.val) != null) &&
1358	>	((ek = e.key) == k || k.equals(ek))) {
1359	>	if (cv == null || cv == ev || cv.equals(ev)) {
1360	>	oldVal = ev;
1361	>	if ((e.val = v) == null) {
1362	>	deleted = true;
1363	>	Node en = e.next;
1364	>	if (pred != null)
1365	>	pred.next = en;
1366	>	else
1367	>	setTabAt(tab, i, en);
1368	>	}
1369	>	}
1370	>	break;
1371	>	}
1372	>	pred = e;
1373	>	if ((e = e.next) == null)
1374	>	break;
1375	>	}
1376	>	}
1377	>	} finally {
1378	>	if (!f.casHash(fh | LOCKED, fh)) {
1379	>	f.hash = fh;
1380	>	synchronized (f) { f.notifyAll(); };
1381	>	}
1382	>	}
1383	>	if (validated) {
1384	>	if (deleted)
1385	>	counter.add(-1L);
1386	>	break;
1387		}
387	–	return replaced;
388	–	} finally {
389	–	unlock();
1388		}
1389		}
1390	+	return oldVal;
1391	+	}
1392
1393	<	V replace(K key, int hash, V newValue) {
1394	<	lock();
1395	<	try {
1396	<	HashEntry<K,V> e = getFirst(hash);
1397	<	while (e != null && (e.hash != hash || !key.equals(e.key)))
1398	<	e = e.next;
1393	>	/*
1394	>	* Internal versions of the six insertion methods, each a
1395	>	* little more complicated than the last. All have
1396	>	* the same basic structure as the first (internalPut):
1397	>	*  1. If table uninitialized, create
1398	>	*  2. If bin empty, try to CAS new node
1399	>	*  3. If bin stale, use new table
1400	>	*  4. if bin converted to TreeBin, validate and relay to TreeBin methods
1401	>	*  5. Lock and validate; if valid, scan and add or update
1402	>	*
1403	>	* The others interweave other checks and/or alternative actions:
1404	>	*  * Plain put checks for and performs resize after insertion.
1405	>	*  * putIfAbsent prescans for mapping without lock (and fails to add
1406	>	*    if present), which also makes pre-emptive resize checks worthwhile.
1407	>	*  * computeIfAbsent extends form used in putIfAbsent with additional
1408	>	*    mechanics to deal with, calls, potential exceptions and null
1409	>	*    returns from function call.
1410	>	*  * compute uses the same function-call mechanics, but without
1411	>	*    the prescans
1412	>	*  * merge acts as putIfAbsent in the absent case, but invokes the
1413	>	*    update function if present
1414	>	*  * putAll attempts to pre-allocate enough table space
1415	>	*    and more lazily performs count updates and checks.
1416	>	*
1417	>	* Someday when details settle down a bit more, it might be worth
1418	>	* some factoring to reduce sprawl.
1419	>	*/
1420
1421	<	V oldValue = null;
1422	<	if (e != null) {
1423	<	oldValue = e.value;
1424	<	e.value = newValue;
1421	>	/** Implementation for put */
1422	>	private final Object internalPut(Object k, Object v) {
1423	>	int h = spread(k.hashCode());
1424	>	int count = 0;
1425	>	for (Node[] tab = table;;) {
1426	>	int i; Node f; int fh; Object fk;
1427	>	if (tab == null)
1428	>	tab = initTable();
1429	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1430	>	if (casTabAt(tab, i, null, new Node(h, k, v, null)))
1431	>	break;                   // no lock when adding to empty bin
1432	>	}
1433	>	else if ((fh = f.hash) == MOVED) {
1434	>	if ((fk = f.key) instanceof TreeBin) {
1435	>	TreeBin t = (TreeBin)fk;
1436	>	Object oldVal = null;
1437	>	t.acquire(0);
1438	>	try {
1439	>	if (tabAt(tab, i) == f) {
1440	>	count = 2;
1441	>	TreeNode p = t.putTreeNode(h, k, v);
1442	>	if (p != null) {
1443	>	oldVal = p.val;
1444	>	p.val = v;
1445	>	}
1446	>	}
1447	>	} finally {
1448	>	t.release(0);
1449	>	}
1450	>	if (count != 0) {
1451	>	if (oldVal != null)
1452	>	return oldVal;
1453	>	break;
1454	>	}
1455	>	}
1456	>	else
1457	>	tab = (Node[])fk;
1458	>	}
1459	>	else if ((fh & LOCKED) != 0) {
1460	>	checkForResize();
1461	>	f.tryAwaitLock(tab, i);
1462	>	}
1463	>	else if (f.casHash(fh, fh | LOCKED)) {
1464	>	Object oldVal = null;
1465	>	try {                        // needed in case equals() throws
1466	>	if (tabAt(tab, i) == f) {
1467	>	count = 1;
1468	>	for (Node e = f;; ++count) {
1469	>	Object ek, ev;
1470	>	if ((e.hash & HASH_BITS) == h &&
1471	>	(ev = e.val) != null &&
1472	>	((ek = e.key) == k || k.equals(ek))) {
1473	>	oldVal = ev;
1474	>	e.val = v;
1475	>	break;
1476	>	}
1477	>	Node last = e;
1478	>	if ((e = e.next) == null) {
1479	>	last.next = new Node(h, k, v, null);
1480	>	if (count >= TREE_THRESHOLD)
1481	>	replaceWithTreeBin(tab, i, k);
1482	>	break;
1483	>	}
1484	>	}
1485	>	}
1486	>	} finally {                  // unlock and signal if needed
1487	>	if (!f.casHash(fh | LOCKED, fh)) {
1488	>	f.hash = fh;
1489	>	synchronized (f) { f.notifyAll(); };
1490	>	}
1491	>	}
1492	>	if (count != 0) {
1493	>	if (oldVal != null)
1494	>	return oldVal;
1495	>	if (tab.length <= 64)
1496	>	count = 2;
1497	>	break;
1498		}
405	–	return oldValue;
406	–	} finally {
407	–	unlock();
1499		}
1500		}
1501	+	counter.add(1L);
1502	+	if (count > 1)
1503	+	checkForResize();
1504	+	return null;
1505	+	}
1506
1507	+	/** Implementation for putIfAbsent */
1508	+	private final Object internalPutIfAbsent(Object k, Object v) {
1509	+	int h = spread(k.hashCode());
1510	+	int count = 0;
1511	+	for (Node[] tab = table;;) {
1512	+	int i; Node f; int fh; Object fk, fv;
1513	+	if (tab == null)
1514	+	tab = initTable();
1515	+	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1516	+	if (casTabAt(tab, i, null, new Node(h, k, v, null)))
1517	+	break;
1518	+	}
1519	+	else if ((fh = f.hash) == MOVED) {
1520	+	if ((fk = f.key) instanceof TreeBin) {
1521	+	TreeBin t = (TreeBin)fk;
1522	+	Object oldVal = null;
1523	+	t.acquire(0);
1524	+	try {
1525	+	if (tabAt(tab, i) == f) {
1526	+	count = 2;
1527	+	TreeNode p = t.putTreeNode(h, k, v);
1528	+	if (p != null)
1529	+	oldVal = p.val;
1530	+	}
1531	+	} finally {
1532	+	t.release(0);
1533	+	}
1534	+	if (count != 0) {
1535	+	if (oldVal != null)
1536	+	return oldVal;
1537	+	break;
1538	+	}
1539	+	}
1540	+	else
1541	+	tab = (Node[])fk;
1542	+	}
1543	+	else if ((fh & HASH_BITS) == h && (fv = f.val) != null &&
1544	+	((fk = f.key) == k || k.equals(fk)))
1545	+	return fv;
1546	+	else {
1547	+	Node g = f.next;
1548	+	if (g != null) { // at least 2 nodes -- search and maybe resize
1549	+	for (Node e = g;;) {
1550	+	Object ek, ev;
1551	+	if ((e.hash & HASH_BITS) == h && (ev = e.val) != null &&
1552	+	((ek = e.key) == k || k.equals(ek)))
1553	+	return ev;
1554	+	if ((e = e.next) == null) {
1555	+	checkForResize();
1556	+	break;
1557	+	}
1558	+	}
1559	+	}
1560	+	if (((fh = f.hash) & LOCKED) != 0) {
1561	+	checkForResize();
1562	+	f.tryAwaitLock(tab, i);
1563	+	}
1564	+	else if (tabAt(tab, i) == f && f.casHash(fh, fh | LOCKED)) {
1565	+	Object oldVal = null;
1566	+	try {
1567	+	if (tabAt(tab, i) == f) {
1568	+	count = 1;
1569	+	for (Node e = f;; ++count) {
1570	+	Object ek, ev;
1571	+	if ((e.hash & HASH_BITS) == h &&
1572	+	(ev = e.val) != null &&
1573	+	((ek = e.key) == k || k.equals(ek))) {
1574	+	oldVal = ev;
1575	+	break;
1576	+	}
1577	+	Node last = e;
1578	+	if ((e = e.next) == null) {
1579	+	last.next = new Node(h, k, v, null);
1580	+	if (count >= TREE_THRESHOLD)
1581	+	replaceWithTreeBin(tab, i, k);
1582	+	break;
1583	+	}
1584	+	}
1585	+	}
1586	+	} finally {
1587	+	if (!f.casHash(fh | LOCKED, fh)) {
1588	+	f.hash = fh;
1589	+	synchronized (f) { f.notifyAll(); };
1590	+	}
1591	+	}
1592	+	if (count != 0) {
1593	+	if (oldVal != null)
1594	+	return oldVal;
1595	+	if (tab.length <= 64)
1596	+	count = 2;
1597	+	break;
1598	+	}
1599	+	}
1600	+	}
1601	+	}
1602	+	counter.add(1L);
1603	+	if (count > 1)
1604	+	checkForResize();
1605	+	return null;
1606	+	}
1607
1608	<	V put(K key, int hash, V value, boolean onlyIfAbsent) {
1609	<	lock();
1610	<	try {
1611	<	int c = count;
1612	<	if (c++ > threshold) // ensure capacity
1613	<	rehash();
1614	<	HashEntry<K,V>[] tab = table;
1615	<	int index = hash & (tab.length - 1);
1616	<	HashEntry<K,V> first = tab[index];
1617	<	HashEntry<K,V> e = first;
1618	<	while (e != null && (e.hash != hash || !key.equals(e.key)))
1619	<	e = e.next;
1620	<
1621	<	V oldValue;
1622	<	if (e != null) {
1623	<	oldValue = e.value;
1624	<	if (!onlyIfAbsent)
1625	<	e.value = value;
1608	>	/** Implementation for computeIfAbsent */
1609	>	private final Object internalComputeIfAbsent(K k,
1610	>	Fun<? super K, ?> mf) {
1611	>	int h = spread(k.hashCode());
1612	>	Object val = null;
1613	>	int count = 0;
1614	>	for (Node[] tab = table;;) {
1615	>	Node f; int i, fh; Object fk, fv;
1616	>	if (tab == null)
1617	>	tab = initTable();
1618	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1619	>	Node node = new Node(fh = h | LOCKED, k, null, null);
1620	>	if (casTabAt(tab, i, null, node)) {
1621	>	count = 1;
1622	>	try {
1623	>	if ((val = mf.apply(k)) != null)
1624	>	node.val = val;
1625	>	} finally {
1626	>	if (val == null)
1627	>	setTabAt(tab, i, null);
1628	>	if (!node.casHash(fh, h)) {
1629	>	node.hash = h;
1630	>	synchronized (node) { node.notifyAll(); };
1631	>	}
1632	>	}
1633		}
1634	<	else {
1635	<	oldValue = null;
1636	<	++modCount;
1637	<	tab[index] = new HashEntry<K,V>(key, hash, first, value);
1638	<	count = c; // write-volatile
1634	>	if (count != 0)
1635	>	break;
1636	>	}
1637	>	else if ((fh = f.hash) == MOVED) {
1638	>	if ((fk = f.key) instanceof TreeBin) {
1639	>	TreeBin t = (TreeBin)fk;
1640	>	boolean added = false;
1641	>	t.acquire(0);
1642	>	try {
1643	>	if (tabAt(tab, i) == f) {
1644	>	count = 1;
1645	>	TreeNode p = t.getTreeNode(h, k, t.root);
1646	>	if (p != null)
1647	>	val = p.val;
1648	>	else if ((val = mf.apply(k)) != null) {
1649	>	added = true;
1650	>	count = 2;
1651	>	t.putTreeNode(h, k, val);
1652	>	}
1653	>	}
1654	>	} finally {
1655	>	t.release(0);
1656	>	}
1657	>	if (count != 0) {
1658	>	if (!added)
1659	>	return val;
1660	>	break;
1661	>	}
1662	>	}
1663	>	else
1664	>	tab = (Node[])fk;
1665	>	}
1666	>	else if ((fh & HASH_BITS) == h && (fv = f.val) != null &&
1667	>	((fk = f.key) == k || k.equals(fk)))
1668	>	return fv;
1669	>	else {
1670	>	Node g = f.next;
1671	>	if (g != null) {
1672	>	for (Node e = g;;) {
1673	>	Object ek, ev;
1674	>	if ((e.hash & HASH_BITS) == h && (ev = e.val) != null &&
1675	>	((ek = e.key) == k || k.equals(ek)))
1676	>	return ev;
1677	>	if ((e = e.next) == null) {
1678	>	checkForResize();
1679	>	break;
1680	>	}
1681	>	}
1682	>	}
1683	>	if (((fh = f.hash) & LOCKED) != 0) {
1684	>	checkForResize();
1685	>	f.tryAwaitLock(tab, i);
1686	>	}
1687	>	else if (tabAt(tab, i) == f && f.casHash(fh, fh | LOCKED)) {
1688	>	boolean added = false;
1689	>	try {
1690	>	if (tabAt(tab, i) == f) {
1691	>	count = 1;
1692	>	for (Node e = f;; ++count) {
1693	>	Object ek, ev;
1694	>	if ((e.hash & HASH_BITS) == h &&
1695	>	(ev = e.val) != null &&
1696	>	((ek = e.key) == k || k.equals(ek))) {
1697	>	val = ev;
1698	>	break;
1699	>	}
1700	>	Node last = e;
1701	>	if ((e = e.next) == null) {
1702	>	if ((val = mf.apply(k)) != null) {
1703	>	added = true;
1704	>	last.next = new Node(h, k, val, null);
1705	>	if (count >= TREE_THRESHOLD)
1706	>	replaceWithTreeBin(tab, i, k);
1707	>	}
1708	>	break;
1709	>	}
1710	>	}
1711	>	}
1712	>	} finally {
1713	>	if (!f.casHash(fh | LOCKED, fh)) {
1714	>	f.hash = fh;
1715	>	synchronized (f) { f.notifyAll(); };
1716	>	}
1717	>	}
1718	>	if (count != 0) {
1719	>	if (!added)
1720	>	return val;
1721	>	if (tab.length <= 64)
1722	>	count = 2;
1723	>	break;
1724	>	}
1725		}
437	–	return oldValue;
438	–	} finally {
439	–	unlock();
1726		}
1727		}
1728	+	if (val != null) {
1729	+	counter.add(1L);
1730	+	if (count > 1)
1731	+	checkForResize();
1732	+	}
1733	+	return val;
1734	+	}
1735
1736	<	void rehash() {
1737	<	HashEntry<K,V>[] oldTable = table;
1738	<	int oldCapacity = oldTable.length;
1739	<	if (oldCapacity >= MAXIMUM_CAPACITY)
1740	<	return;
1741	<
1742	<	/*
1743	<	* Reclassify nodes in each list to new Map.  Because we are
1744	<	* using power-of-two expansion, the elements from each bin
1745	<	* must either stay at same index, or move with a power of two
1746	<	* offset. We eliminate unnecessary node creation by catching
1747	<	* cases where old nodes can be reused because their next
1748	<	* fields won't change. Statistically, at the default
1749	<	* threshold, only about one-sixth of them need cloning when
1750	<	* a table doubles. The nodes they replace will be garbage
1751	<	* collectable as soon as they are no longer referenced by any
1752	<	* reader thread that may be in the midst of traversing table
1753	<	* right now.
1754	<	*/
1755	<
1756	<	HashEntry<K,V>[] newTable = HashEntry.newArray(oldCapacity<<1);
1757	<	threshold = (int)(newTable.length * loadFactor);
1758	<	int sizeMask = newTable.length - 1;
1759	<	for (int i = 0; i < oldCapacity ; i++) {
1760	<	// We need to guarantee that any existing reads of old Map can
1761	<	//  proceed. So we cannot yet null out each bin.
1762	<	HashEntry<K,V> e = oldTable[i];
1763	<
1764	<	if (e != null) {
1765	<	HashEntry<K,V> next = e.next;
1766	<	int idx = e.hash & sizeMask;
1767	<
1768	<	//  Single node on list
1769	<	if (next == null)
1770	<	newTable[idx] = e;
1736	>	/** Implementation for compute */
1737	>	@SuppressWarnings("unchecked") private final Object internalCompute
1738	>	(K k, boolean onlyIfPresent, BiFun<? super K, ? super V, ? extends V> mf) {
1739	>	int h = spread(k.hashCode());
1740	>	Object val = null;
1741	>	int delta = 0;
1742	>	int count = 0;
1743	>	for (Node[] tab = table;;) {
1744	>	Node f; int i, fh; Object fk;
1745	>	if (tab == null)
1746	>	tab = initTable();
1747	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1748	>	if (onlyIfPresent)
1749	>	break;
1750	>	Node node = new Node(fh = h | LOCKED, k, null, null);
1751	>	if (casTabAt(tab, i, null, node)) {
1752	>	try {
1753	>	count = 1;
1754	>	if ((val = mf.apply(k, null)) != null) {
1755	>	node.val = val;
1756	>	delta = 1;
1757	>	}
1758	>	} finally {
1759	>	if (delta == 0)
1760	>	setTabAt(tab, i, null);
1761	>	if (!node.casHash(fh, h)) {
1762	>	node.hash = h;
1763	>	synchronized (node) { node.notifyAll(); };
1764	>	}
1765	>	}
1766	>	}
1767	>	if (count != 0)
1768	>	break;
1769	>	}
1770	>	else if ((fh = f.hash) == MOVED) {
1771	>	if ((fk = f.key) instanceof TreeBin) {
1772	>	TreeBin t = (TreeBin)fk;
1773	>	t.acquire(0);
1774	>	try {
1775	>	if (tabAt(tab, i) == f) {
1776	>	count = 1;
1777	>	TreeNode p = t.getTreeNode(h, k, t.root);
1778	>	Object pv = (p == null) ? null : p.val;
1779	>	if ((val = mf.apply(k, (V)pv)) != null) {
1780	>	if (p != null)
1781	>	p.val = val;
1782	>	else {
1783	>	count = 2;
1784	>	delta = 1;
1785	>	t.putTreeNode(h, k, val);
1786	>	}
1787	>	}
1788	>	else if (p != null) {
1789	>	delta = -1;
1790	>	t.deleteTreeNode(p);
1791	>	}
1792	>	}
1793	>	} finally {
1794	>	t.release(0);
1795	>	}
1796	>	if (count != 0)
1797	>	break;
1798	>	}
1799	>	else
1800	>	tab = (Node[])fk;
1801	>	}
1802	>	else if ((fh & LOCKED) != 0) {
1803	>	checkForResize();
1804	>	f.tryAwaitLock(tab, i);
1805	>	}
1806	>	else if (f.casHash(fh, fh | LOCKED)) {
1807	>	try {
1808	>	if (tabAt(tab, i) == f) {
1809	>	count = 1;
1810	>	for (Node e = f, pred = null;; ++count) {
1811	>	Object ek, ev;
1812	>	if ((e.hash & HASH_BITS) == h &&
1813	>	(ev = e.val) != null &&
1814	>	((ek = e.key) == k || k.equals(ek))) {
1815	>	val = mf.apply(k, (V)ev);
1816	>	if (val != null)
1817	>	e.val = val;
1818	>	else {
1819	>	delta = -1;
1820	>	Node en = e.next;
1821	>	if (pred != null)
1822	>	pred.next = en;
1823	>	else
1824	>	setTabAt(tab, i, en);
1825	>	}
1826	>	break;
1827	>	}
1828	>	pred = e;
1829	>	if ((e = e.next) == null) {
1830	>	if (!onlyIfPresent && (val = mf.apply(k, null)) != null) {
1831	>	pred.next = new Node(h, k, val, null);
1832	>	delta = 1;
1833	>	if (count >= TREE_THRESHOLD)
1834	>	replaceWithTreeBin(tab, i, k);
1835	>	}
1836	>	break;
1837	>	}
1838	>	}
1839	>	}
1840	>	} finally {
1841	>	if (!f.casHash(fh | LOCKED, fh)) {
1842	>	f.hash = fh;
1843	>	synchronized (f) { f.notifyAll(); };
1844	>	}
1845	>	}
1846	>	if (count != 0) {
1847	>	if (tab.length <= 64)
1848	>	count = 2;
1849	>	break;
1850	>	}
1851	>	}
1852	>	}
1853	>	if (delta != 0) {
1854	>	counter.add((long)delta);
1855	>	if (count > 1)
1856	>	checkForResize();
1857	>	}
1858	>	return val;
1859	>	}
1860
1861	<	else {
1862	<	// Reuse trailing consecutive sequence at same slot
1863	<	HashEntry<K,V> lastRun = e;
1864	<	int lastIdx = idx;
1865	<	for (HashEntry<K,V> last = next;
1866	<	last != null;
1867	<	last = last.next) {
1868	<	int k = last.hash & sizeMask;
1869	<	if (k != lastIdx) {
1870	<	lastIdx = k;
1871	<	lastRun = last;
1861	>	/** Implementation for merge */
1862	>	@SuppressWarnings("unchecked") private final Object internalMerge
1863	>	(K k, V v, BiFun<? super V, ? super V, ? extends V> mf) {
1864	>	int h = spread(k.hashCode());
1865	>	Object val = null;
1866	>	int delta = 0;
1867	>	int count = 0;
1868	>	for (Node[] tab = table;;) {
1869	>	int i; Node f; int fh; Object fk, fv;
1870	>	if (tab == null)
1871	>	tab = initTable();
1872	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1873	>	if (casTabAt(tab, i, null, new Node(h, k, v, null))) {
1874	>	delta = 1;
1875	>	val = v;
1876	>	break;
1877	>	}
1878	>	}
1879	>	else if ((fh = f.hash) == MOVED) {
1880	>	if ((fk = f.key) instanceof TreeBin) {
1881	>	TreeBin t = (TreeBin)fk;
1882	>	t.acquire(0);
1883	>	try {
1884	>	if (tabAt(tab, i) == f) {
1885	>	count = 1;
1886	>	TreeNode p = t.getTreeNode(h, k, t.root);
1887	>	val = (p == null) ? v : mf.apply((V)p.val, v);
1888	>	if (val != null) {
1889	>	if (p != null)
1890	>	p.val = val;
1891	>	else {
1892	>	count = 2;
1893	>	delta = 1;
1894	>	t.putTreeNode(h, k, val);
1895	>	}
1896	>	}
1897	>	else if (p != null) {
1898	>	delta = -1;
1899	>	t.deleteTreeNode(p);
1900		}
1901		}
1902	<	newTable[lastIdx] = lastRun;
1902	>	} finally {
1903	>	t.release(0);
1904	>	}
1905	>	if (count != 0)
1906	>	break;
1907	>	}
1908	>	else
1909	>	tab = (Node[])fk;
1910	>	}
1911	>	else if ((fh & LOCKED) != 0) {
1912	>	checkForResize();
1913	>	f.tryAwaitLock(tab, i);
1914	>	}
1915	>	else if (f.casHash(fh, fh | LOCKED)) {
1916	>	try {
1917	>	if (tabAt(tab, i) == f) {
1918	>	count = 1;
1919	>	for (Node e = f, pred = null;; ++count) {
1920	>	Object ek, ev;
1921	>	if ((e.hash & HASH_BITS) == h &&
1922	>	(ev = e.val) != null &&
1923	>	((ek = e.key) == k || k.equals(ek))) {
1924	>	val = mf.apply(v, (V)ev);
1925	>	if (val != null)
1926	>	e.val = val;
1927	>	else {
1928	>	delta = -1;
1929	>	Node en = e.next;
1930	>	if (pred != null)
1931	>	pred.next = en;
1932	>	else
1933	>	setTabAt(tab, i, en);
1934	>	}
1935	>	break;
1936	>	}
1937	>	pred = e;
1938	>	if ((e = e.next) == null) {
1939	>	val = v;
1940	>	pred.next = new Node(h, k, val, null);
1941	>	delta = 1;
1942	>	if (count >= TREE_THRESHOLD)
1943	>	replaceWithTreeBin(tab, i, k);
1944	>	break;
1945	>	}
1946	>	}
1947	>	}
1948	>	} finally {
1949	>	if (!f.casHash(fh | LOCKED, fh)) {
1950	>	f.hash = fh;
1951	>	synchronized (f) { f.notifyAll(); };
1952	>	}
1953	>	}
1954	>	if (count != 0) {
1955	>	if (tab.length <= 64)
1956	>	count = 2;
1957	>	break;
1958	>	}
1959	>	}
1960	>	}
1961	>	if (delta != 0) {
1962	>	counter.add((long)delta);
1963	>	if (count > 1)
1964	>	checkForResize();
1965	>	}
1966	>	return val;
1967	>	}
1968
1969	<	// Clone all remaining nodes
1970	<	for (HashEntry<K,V> p = e; p != lastRun; p = p.next) {
1971	<	int k = p.hash & sizeMask;
1972	<	HashEntry<K,V> n = newTable[k];
1973	<	newTable[k] = new HashEntry<K,V>(p.key, p.hash,
1974	<	n, p.value);
1969	>	/** Implementation for putAll */
1970	>	private final void internalPutAll(Map<?, ?> m) {
1971	>	tryPresize(m.size());
1972	>	long delta = 0L;     // number of uncommitted additions
1973	>	boolean npe = false; // to throw exception on exit for nulls
1974	>	try {                // to clean up counts on other exceptions
1975	>	for (Map.Entry<?, ?> entry : m.entrySet()) {
1976	>	Object k, v;
1977	>	if (entry == null || (k = entry.getKey()) == null ||
1978	>	(v = entry.getValue()) == null) {
1979	>	npe = true;
1980	>	break;
1981	>	}
1982	>	int h = spread(k.hashCode());
1983	>	for (Node[] tab = table;;) {
1984	>	int i; Node f; int fh; Object fk;
1985	>	if (tab == null)
1986	>	tab = initTable();
1987	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null){
1988	>	if (casTabAt(tab, i, null, new Node(h, k, v, null))) {
1989	>	++delta;
1990	>	break;
1991	>	}
1992	>	}
1993	>	else if ((fh = f.hash) == MOVED) {
1994	>	if ((fk = f.key) instanceof TreeBin) {
1995	>	TreeBin t = (TreeBin)fk;
1996	>	boolean validated = false;
1997	>	t.acquire(0);
1998	>	try {
1999	>	if (tabAt(tab, i) == f) {
2000	>	validated = true;
2001	>	TreeNode p = t.getTreeNode(h, k, t.root);
2002	>	if (p != null)
2003	>	p.val = v;
2004	>	else {
2005	>	t.putTreeNode(h, k, v);
2006	>	++delta;
2007	>	}
2008	>	}
2009	>	} finally {
2010	>	t.release(0);
2011	>	}
2012	>	if (validated)
2013	>	break;
2014	>	}
2015	>	else
2016	>	tab = (Node[])fk;
2017	>	}
2018	>	else if ((fh & LOCKED) != 0) {
2019	>	counter.add(delta);
2020	>	delta = 0L;
2021	>	checkForResize();
2022	>	f.tryAwaitLock(tab, i);
2023	>	}
2024	>	else if (f.casHash(fh, fh | LOCKED)) {
2025	>	int count = 0;
2026	>	try {
2027	>	if (tabAt(tab, i) == f) {
2028	>	count = 1;
2029	>	for (Node e = f;; ++count) {
2030	>	Object ek, ev;
2031	>	if ((e.hash & HASH_BITS) == h &&
2032	>	(ev = e.val) != null &&
2033	>	((ek = e.key) == k || k.equals(ek))) {
2034	>	e.val = v;
2035	>	break;
2036	>	}
2037	>	Node last = e;
2038	>	if ((e = e.next) == null) {
2039	>	++delta;
2040	>	last.next = new Node(h, k, v, null);
2041	>	if (count >= TREE_THRESHOLD)
2042	>	replaceWithTreeBin(tab, i, k);
2043	>	break;
2044	>	}
2045	>	}
2046	>	}
2047	>	} finally {
2048	>	if (!f.casHash(fh | LOCKED, fh)) {
2049	>	f.hash = fh;
2050	>	synchronized (f) { f.notifyAll(); };
2051	>	}
2052	>	}
2053	>	if (count != 0) {
2054	>	if (count > 1) {
2055	>	counter.add(delta);
2056	>	delta = 0L;
2057	>	checkForResize();
2058	>	}
2059	>	break;
2060		}
2061		}
2062		}
2063		}
2064	<	table = newTable;
2064	>	} finally {
2065	>	if (delta != 0)
2066	>	counter.add(delta);
2067		}
2068	+	if (npe)
2069	+	throw new NullPointerException();
2070	+	}
2071
2072	<	/**
508	<	* Remove; match on key only if value null, else match both.
509	<	*/
510	<	V remove(Object key, int hash, Object value) {
511	<	lock();
512	<	try {
513	<	int c = count - 1;
514	<	HashEntry<K,V>[] tab = table;
515	<	int index = hash & (tab.length - 1);
516	<	HashEntry<K,V> first = tab[index];
517	<	HashEntry<K,V> e = first;
518	<	while (e != null && (e.hash != hash || !key.equals(e.key)))
519	<	e = e.next;
2072	>	/* ---------------- Table Initialization and Resizing -------------- */
2073
2074	<	V oldValue = null;
2075	<	if (e != null) {
2076	<	V v = e.value;
2077	<	if (value == null || value.equals(v)) {
2078	<	oldValue = v;
2079	<	// All entries following removed node can stay
2080	<	// in list, but all preceding ones need to be
2081	<	// cloned.
2082	<	++modCount;
2083	<	HashEntry<K,V> newFirst = e.next;
2084	<	for (HashEntry<K,V> p = first; p != e; p = p.next)
2085	<	newFirst = new HashEntry<K,V>(p.key, p.hash,
2086	<	newFirst, p.value);
2087	<	tab[index] = newFirst;
2088	<	count = c; // write-volatile
2074	>	/**
2075	>	* Returns a power of two table size for the given desired capacity.
2076	>	* See Hackers Delight, sec 3.2
2077	>	*/
2078	>	private static final int tableSizeFor(int c) {
2079	>	int n = c - 1;
2080	>	n |= n >>> 1;
2081	>	n |= n >>> 2;
2082	>	n |= n >>> 4;
2083	>	n |= n >>> 8;
2084	>	n |= n >>> 16;
2085	>	return (n < 0) ? 1 : (n >= MAXIMUM_CAPACITY) ? MAXIMUM_CAPACITY : n + 1;
2086	>	}
2087	>
2088	>	/**
2089	>	* Initializes table, using the size recorded in sizeCtl.
2090	>	*/
2091	>	private final Node[] initTable() {
2092	>	Node[] tab; int sc;
2093	>	while ((tab = table) == null) {
2094	>	if ((sc = sizeCtl) < 0)
2095	>	Thread.yield(); // lost initialization race; just spin
2096	>	else if (UNSAFE.compareAndSwapInt(this, sizeCtlOffset, sc, -1)) {
2097	>	try {
2098	>	if ((tab = table) == null) {
2099	>	int n = (sc > 0) ? sc : DEFAULT_CAPACITY;
2100	>	tab = table = new Node[n];
2101	>	sc = n - (n >>> 2);
2102		}
2103	+	} finally {
2104	+	sizeCtl = sc;
2105	+	}
2106	+	break;
2107	+	}
2108	+	}
2109	+	return tab;
2110	+	}
2111	+
2112	+	/**
2113	+	* If table is too small and not already resizing, creates next
2114	+	* table and transfers bins.  Rechecks occupancy after a transfer
2115	+	* to see if another resize is already needed because resizings
2116	+	* are lagging additions.
2117	+	*/
2118	+	private final void checkForResize() {
2119	+	Node[] tab; int n, sc;
2120	+	while ((tab = table) != null &&
2121	+	(n = tab.length) < MAXIMUM_CAPACITY &&
2122	+	(sc = sizeCtl) >= 0 && counter.sum() >= (long)sc &&
2123	+	UNSAFE.compareAndSwapInt(this, sizeCtlOffset, sc, -1)) {
2124	+	try {
2125	+	if (tab == table) {
2126	+	table = rebuild(tab);
2127	+	sc = (n << 1) - (n >>> 1);
2128		}
538	–	return oldValue;
2129		} finally {
2130	<	unlock();
2130	>	sizeCtl = sc;
2131		}
2132		}
2133	+	}
2134
2135	<	void clear() {
2136	<	if (count != 0) {
2137	<	lock();
2135	>	/**
2136	>	* Tries to presize table to accommodate the given number of elements.
2137	>	*
2138	>	* @param size number of elements (doesn't need to be perfectly accurate)
2139	>	*/
2140	>	private final void tryPresize(int size) {
2141	>	int c = (size >= (MAXIMUM_CAPACITY >>> 1)) ? MAXIMUM_CAPACITY :
2142	>	tableSizeFor(size + (size >>> 1) + 1);
2143	>	int sc;
2144	>	while ((sc = sizeCtl) >= 0) {
2145	>	Node[] tab = table; int n;
2146	>	if (tab == null || (n = tab.length) == 0) {
2147	>	n = (sc > c) ? sc : c;
2148	>	if (UNSAFE.compareAndSwapInt(this, sizeCtlOffset, sc, -1)) {
2149	>	try {
2150	>	if (table == tab) {
2151	>	table = new Node[n];
2152	>	sc = n - (n >>> 2);
2153	>	}
2154	>	} finally {
2155	>	sizeCtl = sc;
2156	>	}
2157	>	}
2158	>	}
2159	>	else if (c <= sc || n >= MAXIMUM_CAPACITY)
2160	>	break;
2161	>	else if (UNSAFE.compareAndSwapInt(this, sizeCtlOffset, sc, -1)) {
2162		try {
2163	<	HashEntry<K,V>[] tab = table;
2164	<	for (int i = 0; i < tab.length ; i++)
2165	<	tab[i] = null;
2166	<	++modCount;
552	<	count = 0; // write-volatile
2163	>	if (table == tab) {
2164	>	table = rebuild(tab);
2165	>	sc = (n << 1) - (n >>> 1);
2166	>	}
2167		} finally {
2168	<	unlock();
2168	>	sizeCtl = sc;
2169		}
2170		}
2171		}
2172		}
2173
2174	+	/*
2175	+	* Moves and/or copies the nodes in each bin to new table. See
2176	+	* above for explanation.
2177	+	*
2178	+	* @return the new table
2179	+	*/
2180	+	private static final Node[] rebuild(Node[] tab) {
2181	+	int n = tab.length;
2182	+	Node[] nextTab = new Node[n << 1];
2183	+	Node fwd = new Node(MOVED, nextTab, null, null);
2184	+	int[] buffer = null;       // holds bins to revisit; null until needed
2185	+	Node rev = null;           // reverse forwarder; null until needed
2186	+	int nbuffered = 0;         // the number of bins in buffer list
2187	+	int bufferIndex = 0;       // buffer index of current buffered bin
2188	+	int bin = n - 1;           // current non-buffered bin or -1 if none
2189	+
2190	+	for (int i = bin;;) {      // start upwards sweep
2191	+	int fh; Node f;
2192	+	if ((f = tabAt(tab, i)) == null) {
2193	+	if (bin >= 0) {    // Unbuffered; no lock needed (or available)
2194	+	if (!casTabAt(tab, i, f, fwd))
2195	+	continue;
2196	+	}
2197	+	else {             // transiently use a locked forwarding node
2198	+	Node g = new Node(MOVED|LOCKED, nextTab, null, null);
2199	+	if (!casTabAt(tab, i, f, g))
2200	+	continue;
2201	+	setTabAt(nextTab, i, null);
2202	+	setTabAt(nextTab, i + n, null);
2203	+	setTabAt(tab, i, fwd);
2204	+	if (!g.casHash(MOVED|LOCKED, MOVED)) {
2205	+	g.hash = MOVED;
2206	+	synchronized (g) { g.notifyAll(); }
2207	+	}
2208	+	}
2209	+	}
2210	+	else if ((fh = f.hash) == MOVED) {
2211	+	Object fk = f.key;
2212	+	if (fk instanceof TreeBin) {
2213	+	TreeBin t = (TreeBin)fk;
2214	+	boolean validated = false;
2215	+	t.acquire(0);
2216	+	try {
2217	+	if (tabAt(tab, i) == f) {
2218	+	validated = true;
2219	+	splitTreeBin(nextTab, i, t);
2220	+	setTabAt(tab, i, fwd);
2221	+	}
2222	+	} finally {
2223	+	t.release(0);
2224	+	}
2225	+	if (!validated)
2226	+	continue;
2227	+	}
2228	+	}
2229	+	else if ((fh & LOCKED) == 0 && f.casHash(fh, fh|LOCKED)) {
2230	+	boolean validated = false;
2231	+	try {              // split to lo and hi lists; copying as needed
2232	+	if (tabAt(tab, i) == f) {
2233	+	validated = true;
2234	+	splitBin(nextTab, i, f);
2235	+	setTabAt(tab, i, fwd);
2236	+	}
2237	+	} finally {
2238	+	if (!f.casHash(fh | LOCKED, fh)) {
2239	+	f.hash = fh;
2240	+	synchronized (f) { f.notifyAll(); };
2241	+	}
2242	+	}
2243	+	if (!validated)
2244	+	continue;
2245	+	}
2246	+	else {
2247	+	if (buffer == null) // initialize buffer for revisits
2248	+	buffer = new int[TRANSFER_BUFFER_SIZE];
2249	+	if (bin < 0 && bufferIndex > 0) {
2250	+	int j = buffer[--bufferIndex];
2251	+	buffer[bufferIndex] = i;
2252	+	i = j;         // swap with another bin
2253	+	continue;
2254	+	}
2255	+	if (bin < 0 || nbuffered >= TRANSFER_BUFFER_SIZE) {
2256	+	f.tryAwaitLock(tab, i);
2257	+	continue;      // no other options -- block
2258	+	}
2259	+	if (rev == null)   // initialize reverse-forwarder
2260	+	rev = new Node(MOVED, tab, null, null);
2261	+	if (tabAt(tab, i) != f || (f.hash & LOCKED) == 0)
2262	+	continue;      // recheck before adding to list
2263	+	buffer[nbuffered++] = i;
2264	+	setTabAt(nextTab, i, rev);     // install place-holders
2265	+	setTabAt(nextTab, i + n, rev);
2266	+	}
2267
2268	+	if (bin > 0)
2269	+	i = --bin;
2270	+	else if (buffer != null && nbuffered > 0) {
2271	+	bin = -1;
2272	+	i = buffer[bufferIndex = --nbuffered];
2273	+	}
2274	+	else
2275	+	return nextTab;
2276	+	}
2277	+	}
2278
2279	<	/* ---------------- Public operations -------------- */
2279	>	/**
2280	>	* Splits a normal bin with list headed by e into lo and hi parts;
2281	>	* installs in given table.
2282	>	*/
2283	>	private static void splitBin(Node[] nextTab, int i, Node e) {
2284	>	int bit = nextTab.length >>> 1; // bit to split on
2285	>	int runBit = e.hash & bit;
2286	>	Node lastRun = e, lo = null, hi = null;
2287	>	for (Node p = e.next; p != null; p = p.next) {
2288	>	int b = p.hash & bit;
2289	>	if (b != runBit) {
2290	>	runBit = b;
2291	>	lastRun = p;
2292	>	}
2293	>	}
2294	>	if (runBit == 0)
2295	>	lo = lastRun;
2296	>	else
2297	>	hi = lastRun;
2298	>	for (Node p = e; p != lastRun; p = p.next) {
2299	>	int ph = p.hash & HASH_BITS;
2300	>	Object pk = p.key, pv = p.val;
2301	>	if ((ph & bit) == 0)
2302	>	lo = new Node(ph, pk, pv, lo);
2303	>	else
2304	>	hi = new Node(ph, pk, pv, hi);
2305	>	}
2306	>	setTabAt(nextTab, i, lo);
2307	>	setTabAt(nextTab, i + bit, hi);
2308	>	}
2309
2310		/**
2311	<	* Creates a new, empty map with the specified initial
566	<	* capacity, load factor and concurrency level.
567	<	*
568	<	* @param initialCapacity the initial capacity. The implementation
569	<	* performs internal sizing to accommodate this many elements.
570	<	* @param loadFactor  the load factor threshold, used to control resizing.
571	<	* Resizing may be performed when the average number of elements per
572	<	* bin exceeds this threshold.
573	<	* @param concurrencyLevel the estimated number of concurrently
574	<	* updating threads. The implementation performs internal sizing
575	<	* to try to accommodate this many threads.
576	<	* @throws IllegalArgumentException if the initial capacity is
577	<	* negative or the load factor or concurrencyLevel are
578	<	* nonpositive.
2311	>	* Splits a tree bin into lo and hi parts; installs in given table.
2312		*/
2313	<	public ConcurrentHashMap(int initialCapacity,
2314	<	float loadFactor, int concurrencyLevel) {
2315	<	if (!(loadFactor > 0) || initialCapacity < 0 || concurrencyLevel <= 0)
2316	<	throw new IllegalArgumentException();
2313	>	private static void splitTreeBin(Node[] nextTab, int i, TreeBin t) {
2314	>	int bit = nextTab.length >>> 1;
2315	>	TreeBin lt = new TreeBin();
2316	>	TreeBin ht = new TreeBin();
2317	>	int lc = 0, hc = 0;
2318	>	for (Node e = t.first; e != null; e = e.next) {
2319	>	int h = e.hash & HASH_BITS;
2320	>	Object k = e.key, v = e.val;
2321	>	if ((h & bit) == 0) {
2322	>	++lc;
2323	>	lt.putTreeNode(h, k, v);
2324	>	}
2325	>	else {
2326	>	++hc;
2327	>	ht.putTreeNode(h, k, v);
2328	>	}
2329	>	}
2330	>	Node ln, hn; // throw away trees if too small
2331	>	if (lc <= (TREE_THRESHOLD >>> 1)) {
2332	>	ln = null;
2333	>	for (Node p = lt.first; p != null; p = p.next)
2334	>	ln = new Node(p.hash, p.key, p.val, ln);
2335	>	}
2336	>	else
2337	>	ln = new Node(MOVED, lt, null, null);
2338	>	setTabAt(nextTab, i, ln);
2339	>	if (hc <= (TREE_THRESHOLD >>> 1)) {
2340	>	hn = null;
2341	>	for (Node p = ht.first; p != null; p = p.next)
2342	>	hn = new Node(p.hash, p.key, p.val, hn);
2343	>	}
2344	>	else
2345	>	hn = new Node(MOVED, ht, null, null);
2346	>	setTabAt(nextTab, i + bit, hn);
2347	>	}
2348
2349	<	if (concurrencyLevel > MAX_SEGMENTS)
2350	<	concurrencyLevel = MAX_SEGMENTS;
2349	>	/**
2350	>	* Implementation for clear. Steps through each bin, removing all
2351	>	* nodes.
2352	>	*/
2353	>	private final void internalClear() {
2354	>	long delta = 0L; // negative number of deletions
2355	>	int i = 0;
2356	>	Node[] tab = table;
2357	>	while (tab != null && i < tab.length) {
2358	>	int fh; Object fk;
2359	>	Node f = tabAt(tab, i);
2360	>	if (f == null)
2361	>	++i;
2362	>	else if ((fh = f.hash) == MOVED) {
2363	>	if ((fk = f.key) instanceof TreeBin) {
2364	>	TreeBin t = (TreeBin)fk;
2365	>	t.acquire(0);
2366	>	try {
2367	>	if (tabAt(tab, i) == f) {
2368	>	for (Node p = t.first; p != null; p = p.next) {
2369	>	if (p.val != null) { // (currently always true)
2370	>	p.val = null;
2371	>	--delta;
2372	>	}
2373	>	}
2374	>	t.first = null;
2375	>	t.root = null;
2376	>	++i;
2377	>	}
2378	>	} finally {
2379	>	t.release(0);
2380	>	}
2381	>	}
2382	>	else
2383	>	tab = (Node[])fk;
2384	>	}
2385	>	else if ((fh & LOCKED) != 0) {
2386	>	counter.add(delta); // opportunistically update count
2387	>	delta = 0L;
2388	>	f.tryAwaitLock(tab, i);
2389	>	}
2390	>	else if (f.casHash(fh, fh | LOCKED)) {
2391	>	try {
2392	>	if (tabAt(tab, i) == f) {
2393	>	for (Node e = f; e != null; e = e.next) {
2394	>	if (e.val != null) {  // (currently always true)
2395	>	e.val = null;
2396	>	--delta;
2397	>	}
2398	>	}
2399	>	setTabAt(tab, i, null);
2400	>	++i;
2401	>	}
2402	>	} finally {
2403	>	if (!f.casHash(fh | LOCKED, fh)) {
2404	>	f.hash = fh;
2405	>	synchronized (f) { f.notifyAll(); };
2406	>	}
2407	>	}
2408	>	}
2409	>	}
2410	>	if (delta != 0)
2411	>	counter.add(delta);
2412	>	}
2413	>
2414	>	/* ----------------Table Traversal -------------- */
2415	>
2416	>	/**
2417	>	* Encapsulates traversal for methods such as containsValue; also
2418	>	* serves as a base class for other iterators and bulk tasks.
2419	>	*
2420	>	* At each step, the iterator snapshots the key ("nextKey") and
2421	>	* value ("nextVal") of a valid node (i.e., one that, at point of
2422	>	* snapshot, has a non-null user value). Because val fields can
2423	>	* change (including to null, indicating deletion), field nextVal
2424	>	* might not be accurate at point of use, but still maintains the
2425	>	* weak consistency property of holding a value that was once
2426	>	* valid. To support iterator.remove, the nextKey field is not
2427	>	* updated (nulled out) when the iterator cannot advance.
2428	>	*
2429	>	* Internal traversals directly access these fields, as in:
2430	>	* {@code while (it.advance() != null) { process(it.nextKey); }}
2431	>	*
2432	>	* Exported iterators must track whether the iterator has advanced
2433	>	* (in hasNext vs next) (by setting/checking/nulling field
2434	>	* nextVal), and then extract key, value, or key-value pairs as
2435	>	* return values of next().
2436	>	*
2437	>	* The iterator visits once each still-valid node that was
2438	>	* reachable upon iterator construction. It might miss some that
2439	>	* were added to a bin after the bin was visited, which is OK wrt
2440	>	* consistency guarantees. Maintaining this property in the face
2441	>	* of possible ongoing resizes requires a fair amount of
2442	>	* bookkeeping state that is difficult to optimize away amidst
2443	>	* volatile accesses.  Even so, traversal maintains reasonable
2444	>	* throughput.
2445	>	*
2446	>	* Normally, iteration proceeds bin-by-bin traversing lists.
2447	>	* However, if the table has been resized, then all future steps
2448	>	* must traverse both the bin at the current index as well as at
2449	>	* (index + baseSize); and so on for further resizings. To
2450	>	* paranoically cope with potential sharing by users of iterators
2451	>	* across threads, iteration terminates if a bounds checks fails
2452	>	* for a table read.
2453	>	*
2454	>	* This class extends ForkJoinTask to streamline parallel
2455	>	* iteration in bulk operations (see BulkTask). This adds only an
2456	>	* int of space overhead, which is close enough to negligible in
2457	>	* cases where it is not needed to not worry about it.  Because
2458	>	* ForkJoinTask is Serializable, but iterators need not be, we
2459	>	* need to add warning suppressions.
2460	>	*/
2461	>	@SuppressWarnings("serial") static class Traverser<K,V,R> extends ForkJoinTask<R> {
2462	>	final ConcurrentHashMap<K, V> map;
2463	>	Node next;           // the next entry to use
2464	>	Object nextKey;      // cached key field of next
2465	>	Object nextVal;      // cached val field of next
2466	>	Node[] tab;          // current table; updated if resized
2467	>	int index;           // index of bin to use next
2468	>	int baseIndex;       // current index of initial table
2469	>	int baseLimit;       // index bound for initial table
2470	>	int baseSize;        // initial table size
2471	>
2472	>	/** Creates iterator for all entries in the table. */
2473	>	Traverser(ConcurrentHashMap<K, V> map) {
2474	>	this.map = map;
2475	>	}
2476	>
2477	>	/** Creates iterator for split() methods */
2478	>	Traverser(Traverser<K,V,?> it) {
2479	>	ConcurrentHashMap<K, V> m; Node[] t;
2480	>	if ((m = this.map = it.map) == null)
2481	>	t = null;
2482	>	else if ((t = it.tab) == null && // force parent tab initialization
2483	>	(t = it.tab = m.table) != null)
2484	>	it.baseLimit = it.baseSize = t.length;
2485	>	this.tab = t;
2486	>	this.baseSize = it.baseSize;
2487	>	it.baseLimit = this.index = this.baseIndex =
2488	>	((this.baseLimit = it.baseLimit) + it.baseIndex + 1) >>> 1;
2489	>	}
2490	>
2491	>	/**
2492	>	* Advances next; returns nextVal or null if terminated.
2493	>	* See above for explanation.
2494	>	*/
2495	>	final Object advance() {
2496	>	Node e = next;
2497	>	Object ev = null;
2498	>	outer: do {
2499	>	if (e != null)                  // advance past used/skipped node
2500	>	e = e.next;
2501	>	while (e == null) {             // get to next non-null bin
2502	>	ConcurrentHashMap<K, V> m;
2503	>	Node[] t; int b, i, n; Object ek; // checks must use locals
2504	>	if ((t = tab) != null)
2505	>	n = t.length;
2506	>	else if ((m = map) != null && (t = tab = m.table) != null)
2507	>	n = baseLimit = baseSize = t.length;
2508	>	else
2509	>	break outer;
2510	>	if ((b = baseIndex) >= baseLimit ||
2511	>	(i = index) < 0 || i >= n)
2512	>	break outer;
2513	>	if ((e = tabAt(t, i)) != null && e.hash == MOVED) {
2514	>	if ((ek = e.key) instanceof TreeBin)
2515	>	e = ((TreeBin)ek).first;
2516	>	else {
2517	>	tab = (Node[])ek;
2518	>	continue;           // restarts due to null val
2519	>	}
2520	>	}                           // visit upper slots if present
2521	>	index = (i += baseSize) < n ? i : (baseIndex = b + 1);
2522	>	}
2523	>	nextKey = e.key;
2524	>	} while ((ev = e.val) == null);    // skip deleted or special nodes
2525	>	next = e;
2526	>	return nextVal = ev;
2527	>	}
2528	>
2529	>	public final void remove() {
2530	>	Object k = nextKey;
2531	>	if (k == null && (advance() == null || (k = nextKey) == null))
2532	>	throw new IllegalStateException();
2533	>	map.internalReplace(k, null, null);
2534	>	}
2535
2536	<	// Find power-of-two sizes best matching arguments
2537	<	int sshift = 0;
2538	<	int ssize = 1;
591	<	while (ssize < concurrencyLevel) {
592	<	++sshift;
593	<	ssize <<= 1;
594	<	}
595	<	segmentShift = 32 - sshift;
596	<	segmentMask = ssize - 1;
597	<	this.segments = Segment.newArray(ssize);
598	<
599	<	if (initialCapacity > MAXIMUM_CAPACITY)
600	<	initialCapacity = MAXIMUM_CAPACITY;
601	<	int c = initialCapacity / ssize;
602	<	if (c * ssize < initialCapacity)
603	<	++c;
604	<	int cap = 1;
605	<	while (cap < c)
606	<	cap <<= 1;
2536	>	public final boolean hasNext() {
2537	>	return nextVal != null || advance() != null;
2538	>	}
2539
2540	<	for (int i = 0; i < this.segments.length; ++i)
2541	<	this.segments[i] = new Segment<K,V>(cap, loadFactor);
2540	>	public final boolean hasMoreElements() { return hasNext(); }
2541	>	public final void setRawResult(Object x) { }
2542	>	public R getRawResult() { return null; }
2543	>	public boolean exec() { return true; }
2544		}
2545
2546	+	/* ---------------- Public operations -------------- */
2547	+
2548		/**
2549	<	* Creates a new, empty map with the specified initial capacity
2550	<	* and load factor and with the default concurrencyLevel (16).
2549	>	* Creates a new, empty map with the default initial table size (16).
2550	>	*/
2551	>	public ConcurrentHashMap() {
2552	>	this.counter = new LongAdder();
2553	>	}
2554	>
2555	>	/**
2556	>	* Creates a new, empty map with an initial table size
2557	>	* accommodating the specified number of elements without the need
2558	>	* to dynamically resize.
2559		*
2560		* @param initialCapacity The implementation performs internal
2561		* sizing to accommodate this many elements.
2562	<	* @param loadFactor  the load factor threshold, used to control resizing.
2563	<	* Resizing may be performed when the average number of elements per
2564	<	* bin exceeds this threshold.
2562	>	* @throws IllegalArgumentException if the initial capacity of
2563	>	* elements is negative
2564	>	*/
2565	>	public ConcurrentHashMap(int initialCapacity) {
2566	>	if (initialCapacity < 0)
2567	>	throw new IllegalArgumentException();
2568	>	int cap = ((initialCapacity >= (MAXIMUM_CAPACITY >>> 1)) ?
2569	>	MAXIMUM_CAPACITY :
2570	>	tableSizeFor(initialCapacity + (initialCapacity >>> 1) + 1));
2571	>	this.counter = new LongAdder();
2572	>	this.sizeCtl = cap;
2573	>	}
2574	>
2575	>	/**
2576	>	* Creates a new map with the same mappings as the given map.
2577	>	*
2578	>	* @param m the map
2579	>	*/
2580	>	public ConcurrentHashMap(Map<? extends K, ? extends V> m) {
2581	>	this.counter = new LongAdder();
2582	>	this.sizeCtl = DEFAULT_CAPACITY;
2583	>	internalPutAll(m);
2584	>	}
2585	>
2586	>	/**
2587	>	* Creates a new, empty map with an initial table size based on
2588	>	* the given number of elements ({@code initialCapacity}) and
2589	>	* initial table density ({@code loadFactor}).
2590	>	*
2591	>	* @param initialCapacity the initial capacity. The implementation
2592	>	* performs internal sizing to accommodate this many elements,
2593	>	* given the specified load factor.
2594	>	* @param loadFactor the load factor (table density) for
2595	>	* establishing the initial table size
2596		* @throws IllegalArgumentException if the initial capacity of
2597		* elements is negative or the load factor is nonpositive
2598		*
2599		* @since 1.6
2600		*/
2601		public ConcurrentHashMap(int initialCapacity, float loadFactor) {
2602	<	this(initialCapacity, loadFactor, DEFAULT_CONCURRENCY_LEVEL);
2602	>	this(initialCapacity, loadFactor, 1);
2603		}
2604
2605		/**
2606	<	* Creates a new, empty map with the specified initial capacity,
2607	<	* and with default load factor (0.75) and concurrencyLevel (16).
2606	>	* Creates a new, empty map with an initial table size based on
2607	>	* the given number of elements ({@code initialCapacity}), table
2608	>	* density ({@code loadFactor}), and number of concurrently
2609	>	* updating threads ({@code concurrencyLevel}).
2610		*
2611		* @param initialCapacity the initial capacity. The implementation
2612	<	* performs internal sizing to accommodate this many elements.
2613	<	* @throws IllegalArgumentException if the initial capacity of
2614	<	* elements is negative.
2612	>	* performs internal sizing to accommodate this many elements,
2613	>	* given the specified load factor.
2614	>	* @param loadFactor the load factor (table density) for
2615	>	* establishing the initial table size
2616	>	* @param concurrencyLevel the estimated number of concurrently
2617	>	* updating threads. The implementation may use this value as
2618	>	* a sizing hint.
2619	>	* @throws IllegalArgumentException if the initial capacity is
2620	>	* negative or the load factor or concurrencyLevel are
2621	>	* nonpositive
2622		*/
2623	<	public ConcurrentHashMap(int initialCapacity) {
2624	<	this(initialCapacity, DEFAULT_LOAD_FACTOR, DEFAULT_CONCURRENCY_LEVEL);
2623	>	public ConcurrentHashMap(int initialCapacity,
2624	>	float loadFactor, int concurrencyLevel) {
2625	>	if (!(loadFactor > 0.0f) || initialCapacity < 0 || concurrencyLevel <= 0)
2626	>	throw new IllegalArgumentException();
2627	>	if (initialCapacity < concurrencyLevel)   // Use at least as many bins
2628	>	initialCapacity = concurrencyLevel;   // as estimated threads
2629	>	long size = (long)(1.0 + (long)initialCapacity / loadFactor);
2630	>	int cap = (size >= (long)MAXIMUM_CAPACITY) ?
2631	>	MAXIMUM_CAPACITY : tableSizeFor((int)size);
2632	>	this.counter = new LongAdder();
2633	>	this.sizeCtl = cap;
2634		}
2635
2636		/**
2637	<	* Creates a new, empty map with a default initial capacity (16),
2638	<	* load factor (0.75) and concurrencyLevel (16).
2637	>	* Creates a new {@link Set} backed by a ConcurrentHashMap
2638	>	* from the given type to {@code Boolean.TRUE}.
2639	>	*
2640	>	* @return the new set
2641		*/
2642	<	public ConcurrentHashMap() {
2643	<	this(DEFAULT_INITIAL_CAPACITY, DEFAULT_LOAD_FACTOR, DEFAULT_CONCURRENCY_LEVEL);
2642	>	public static <K> KeySetView<K,Boolean> newKeySet() {
2643	>	return new KeySetView<K,Boolean>(new ConcurrentHashMap<K,Boolean>(),
2644	>	Boolean.TRUE);
2645		}
2646
2647		/**
2648	<	* Creates a new map with the same mappings as the given map.
2649	<	* The map is created with a capacity of 1.5 times the number
654	<	* of mappings in the given map or 16 (whichever is greater),
655	<	* and a default load factor (0.75) and concurrencyLevel (16).
2648	>	* Creates a new {@link Set} backed by a ConcurrentHashMap
2649	>	* from the given type to {@code Boolean.TRUE}.
2650		*
2651	<	* @param m the map
2651	>	* @param initialCapacity The implementation performs internal
2652	>	* sizing to accommodate this many elements.
2653	>	* @throws IllegalArgumentException if the initial capacity of
2654	>	* elements is negative
2655	>	* @return the new set
2656		*/
2657	<	public ConcurrentHashMap(Map<? extends K, ? extends V> m) {
2658	<	this(Math.max((int) (m.size() / DEFAULT_LOAD_FACTOR) + 1,
2659	<	DEFAULT_INITIAL_CAPACITY),
662	<	DEFAULT_LOAD_FACTOR, DEFAULT_CONCURRENCY_LEVEL);
663	<	putAll(m);
2657	>	public static <K> KeySetView<K,Boolean> newKeySet(int initialCapacity) {
2658	>	return new KeySetView<K,Boolean>(new ConcurrentHashMap<K,Boolean>(initialCapacity),
2659	>	Boolean.TRUE);
2660		}
2661
2662		/**
2663	<	* Returns <tt>true</tt> if this map contains no key-value mappings.
668	<	*
669	<	* @return <tt>true</tt> if this map contains no key-value mappings
2663	>	* {@inheritDoc}
2664		*/
2665		public boolean isEmpty() {
2666	<	final Segment<K,V>[] segments = this.segments;
673	<	/*
674	<	* We keep track of per-segment modCounts to avoid ABA
675	<	* problems in which an element in one segment was added and
676	<	* in another removed during traversal, in which case the
677	<	* table was never actually empty at any point. Note the
678	<	* similar use of modCounts in the size() and containsValue()
679	<	* methods, which are the only other methods also susceptible
680	<	* to ABA problems.
681	<	*/
682	<	int[] mc = new int[segments.length];
683	<	int mcsum = 0;
684	<	for (int i = 0; i < segments.length; ++i) {
685	<	if (segments[i].count != 0)
686	<	return false;
687	<	else
688	<	mcsum += mc[i] = segments[i].modCount;
689	<	}
690	<	// If mcsum happens to be zero, then we know we got a snapshot
691	<	// before any modifications at all were made.  This is
692	<	// probably common enough to bother tracking.
693	<	if (mcsum != 0) {
694	<	for (int i = 0; i < segments.length; ++i) {
695	<	if (segments[i].count != 0 ||
696	<	mc[i] != segments[i].modCount)
697	<	return false;
698	<	}
699	<	}
700	<	return true;
2666	>	return counter.sum() <= 0L; // ignore transient negative values
2667		}
2668
2669		/**
2670	<	* Returns the number of key-value mappings in this map.  If the
705	<	* map contains more than <tt>Integer.MAX_VALUE</tt> elements, returns
706	<	* <tt>Integer.MAX_VALUE</tt>.
707	<	*
708	<	* @return the number of key-value mappings in this map
2670	>	* {@inheritDoc}
2671		*/
2672		public int size() {
2673	<	final Segment<K,V>[] segments = this.segments;
2674	<	long sum = 0;
2675	<	long check = 0;
2676	<	int[] mc = new int[segments.length];
2677	<	// Try a few times to get accurate count. On failure due to
2678	<	// continuous async changes in table, resort to locking.
2679	<	for (int k = 0; k < RETRIES_BEFORE_LOCK; ++k) {
2680	<	check = 0;
2681	<	sum = 0;
2682	<	int mcsum = 0;
2683	<	for (int i = 0; i < segments.length; ++i) {
2684	<	sum += segments[i].count;
2685	<	mcsum += mc[i] = segments[i].modCount;
2686	<	}
2687	<	if (mcsum != 0) {
2688	<	for (int i = 0; i < segments.length; ++i) {
2689	<	check += segments[i].count;
2690	<	if (mc[i] != segments[i].modCount) {
729	<	check = -1; // force retry
730	<	break;
731	<	}
732	<	}
733	<	}
734	<	if (check == sum)
735	<	break;
736	<	}
737	<	if (check != sum) { // Resort to locking all segments
738	<	sum = 0;
739	<	for (int i = 0; i < segments.length; ++i)
740	<	segments[i].lock();
741	<	for (int i = 0; i < segments.length; ++i)
742	<	sum += segments[i].count;
743	<	for (int i = 0; i < segments.length; ++i)
744	<	segments[i].unlock();
745	<	}
746	<	if (sum > Integer.MAX_VALUE)
747	<	return Integer.MAX_VALUE;
748	<	else
749	<	return (int)sum;
2673	>	long n = counter.sum();
2674	>	return ((n < 0L) ? 0 :
2675	>	(n > (long)Integer.MAX_VALUE) ? Integer.MAX_VALUE :
2676	>	(int)n);
2677	>	}
2678	>
2679	>	/**
2680	>	* Returns the number of mappings. This method should be used
2681	>	* instead of {@link #size} because a ConcurrentHashMap may
2682	>	* contain more mappings than can be represented as an int. The
2683	>	* value returned is a snapshot; the actual count may differ if
2684	>	* there are ongoing concurrent insertions or removals.
2685	>	*
2686	>	* @return the number of mappings
2687	>	*/
2688	>	public long mappingCount() {
2689	>	long n = counter.sum();
2690	>	return (n < 0L) ? 0L : n; // ignore transient negative values
2691		}
2692
2693		/**
2694	<	* Returns the value to which this map maps the specified key, or
2695	<	* <tt>null</tt> if the map contains no mapping for the key.
2694	>	* Returns the value to which the specified key is mapped,
2695	>	* or {@code null} if this map contains no mapping for the key.
2696	>	*
2697	>	* <p>More formally, if this map contains a mapping from a key
2698	>	* {@code k} to a value {@code v} such that {@code key.equals(k)},
2699	>	* then this method returns {@code v}; otherwise it returns
2700	>	* {@code null}.  (There can be at most one such mapping.)
2701		*
756	–	* @param key key whose associated value is to be returned
757	–	* @return the value to which this map maps the specified key, or
758	–	*         <tt>null</tt> if the map contains no mapping for the key
2702		* @throws NullPointerException if the specified key is null
2703		*/
2704	<	public V get(Object key) {
2705	<	int hash = hash(key); // throws NullPointerException if key null
2706	<	return segmentFor(hash).get(key, hash);
2704	>	@SuppressWarnings("unchecked") public V get(Object key) {
2705	>	if (key == null)
2706	>	throw new NullPointerException();
2707	>	return (V)internalGet(key);
2708	>	}
2709	>
2710	>	/**
2711	>	* Returns the value to which the specified key is mapped,
2712	>	* or the given defaultValue if this map contains no mapping for the key.
2713	>	*
2714	>	* @param key the key
2715	>	* @param defaultValue the value to return if this map contains
2716	>	* no mapping for the given key
2717	>	* @return the mapping for the key, if present; else the defaultValue
2718	>	* @throws NullPointerException if the specified key is null
2719	>	*/
2720	>	@SuppressWarnings("unchecked") public V getValueOrDefault(Object key, V defaultValue) {
2721	>	if (key == null)
2722	>	throw new NullPointerException();
2723	>	V v = (V) internalGet(key);
2724	>	return v == null ? defaultValue : v;
2725		}
2726
2727		/**
2728		* Tests if the specified object is a key in this table.
2729		*
2730		* @param  key   possible key
2731	<	* @return <tt>true</tt> if and only if the specified object
2731	>	* @return {@code true} if and only if the specified object
2732		*         is a key in this table, as determined by the
2733	<	*         <tt>equals</tt> method; <tt>false</tt> otherwise.
2733	>	*         {@code equals} method; {@code false} otherwise
2734		* @throws NullPointerException if the specified key is null
2735		*/
2736		public boolean containsKey(Object key) {
2737	<	int hash = hash(key); // throws NullPointerException if key null
2738	<	return segmentFor(hash).containsKey(key, hash);
2737	>	if (key == null)
2738	>	throw new NullPointerException();
2739	>	return internalGet(key) != null;
2740		}
2741
2742		/**
2743	<	* Returns <tt>true</tt> if this map maps one or more keys to the
2744	<	* specified value. Note: This method requires a full internal
2745	<	* traversal of the hash table, and so is much slower than
784	<	* method <tt>containsKey</tt>.
2743	>	* Returns {@code true} if this map maps one or more keys to the
2744	>	* specified value. Note: This method may require a full traversal
2745	>	* of the map, and is much slower than method {@code containsKey}.
2746		*
2747		* @param value value whose presence in this map is to be tested
2748	<	* @return <tt>true</tt> if this map maps one or more keys to the
2748	>	* @return {@code true} if this map maps one or more keys to the
2749		*         specified value
2750		* @throws NullPointerException if the specified value is null
2751		*/
2752		public boolean containsValue(Object value) {
2753		if (value == null)
2754		throw new NullPointerException();
2755	<
2756	<	// See explanation of modCount use above
2757	<
2758	<	final Segment<K,V>[] segments = this.segments;
2759	<	int[] mc = new int[segments.length];
799	<
800	<	// Try a few times without locking
801	<	for (int k = 0; k < RETRIES_BEFORE_LOCK; ++k) {
802	<	int sum = 0;
803	<	int mcsum = 0;
804	<	for (int i = 0; i < segments.length; ++i) {
805	<	int c = segments[i].count;
806	<	mcsum += mc[i] = segments[i].modCount;
807	<	if (segments[i].containsValue(value))
808	<	return true;
809	<	}
810	<	boolean cleanSweep = true;
811	<	if (mcsum != 0) {
812	<	for (int i = 0; i < segments.length; ++i) {
813	<	int c = segments[i].count;
814	<	if (mc[i] != segments[i].modCount) {
815	<	cleanSweep = false;
816	<	break;
817	<	}
818	<	}
819	<	}
820	<	if (cleanSweep)
821	<	return false;
822	<	}
823	<	// Resort to locking all segments
824	<	for (int i = 0; i < segments.length; ++i)
825	<	segments[i].lock();
826	<	boolean found = false;
827	<	try {
828	<	for (int i = 0; i < segments.length; ++i) {
829	<	if (segments[i].containsValue(value)) {
830	<	found = true;
831	<	break;
832	<	}
833	<	}
834	<	} finally {
835	<	for (int i = 0; i < segments.length; ++i)
836	<	segments[i].unlock();
2755	>	Object v;
2756	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
2757	>	while ((v = it.advance()) != null) {
2758	>	if (v == value || value.equals(v))
2759	>	return true;
2760		}
2761	<	return found;
2761	>	return false;
2762		}
2763
2764		/**
#	Line 845 | Line 2768 | public class ConcurrentHashMap<K, V> ext
2768		* full compatibility with class {@link java.util.Hashtable},
2769		* which supported this method prior to introduction of the
2770		* Java Collections framework.
2771	<
2771	>	*
2772		* @param  value a value to search for
2773	<	* @return <tt>true</tt> if and only if some key maps to the
2774	<	*         <tt>value</tt> argument in this table as
2775	<	*         determined by the <tt>equals</tt> method;
2776	<	*         <tt>false</tt> otherwise
2773	>	* @return {@code true} if and only if some key maps to the
2774	>	*         {@code value} argument in this table as
2775	>	*         determined by the {@code equals} method;
2776	>	*         {@code false} otherwise
2777		* @throws NullPointerException if the specified value is null
2778		*/
2779		public boolean contains(Object value) {
#	Line 861 | Line 2784 | public class ConcurrentHashMap<K, V> ext
2784		* Maps the specified key to the specified value in this table.
2785		* Neither the key nor the value can be null.
2786		*
2787	<	* <p> The value can be retrieved by calling the <tt>get</tt> method
2787	>	* <p> The value can be retrieved by calling the {@code get} method
2788		* with a key that is equal to the original key.
2789		*
2790		* @param key key with which the specified value is to be associated
2791		* @param value value to be associated with the specified key
2792	<	* @return the previous value associated with <tt>key</tt>, or
2793	<	*         <tt>null</tt> if there was no mapping for <tt>key</tt>
2792	>	* @return the previous value associated with {@code key}, or
2793	>	*         {@code null} if there was no mapping for {@code key}
2794		* @throws NullPointerException if the specified key or value is null
2795		*/
2796	<	public V put(K key, V value) {
2797	<	if (value == null)
2796	>	@SuppressWarnings("unchecked") public V put(K key, V value) {
2797	>	if (key == null || value == null)
2798		throw new NullPointerException();
2799	<	int hash = hash(key);
877	<	return segmentFor(hash).put(key, hash, value, false);
2799	>	return (V)internalPut(key, value);
2800		}
2801
2802		/**
2803		* {@inheritDoc}
2804		*
2805		* @return the previous value associated with the specified key,
2806	<	*         or <tt>null</tt> if there was no mapping for the key
2806	>	*         or {@code null} if there was no mapping for the key
2807		* @throws NullPointerException if the specified key or value is null
2808		*/
2809	<	public V putIfAbsent(K key, V value) {
2810	<	if (value == null)
2809	>	@SuppressWarnings("unchecked") public V putIfAbsent(K key, V value) {
2810	>	if (key == null || value == null)
2811		throw new NullPointerException();
2812	<	int hash = hash(key);
891	<	return segmentFor(hash).put(key, hash, value, true);
2812	>	return (V)internalPutIfAbsent(key, value);
2813		}
2814
2815		/**
#	Line 899 | Line 2820 | public class ConcurrentHashMap<K, V> ext
2820		* @param m mappings to be stored in this map
2821		*/
2822		public void putAll(Map<? extends K, ? extends V> m) {
2823	<	for (Iterator<? extends Map.Entry<? extends K, ? extends V>> it = (Iterator<? extends Map.Entry<? extends K, ? extends V>>) m.entrySet().iterator(); it.hasNext(); ) {
2824	<	Entry<? extends K, ? extends V> e = it.next();
2825	<	put(e.getKey(), e.getValue());
2826	<	}
2823	>	internalPutAll(m);
2824	>	}
2825	>
2826	>	/**
2827	>	* If the specified key is not already associated with a value,
2828	>	* computes its value using the given mappingFunction and enters
2829	>	* it into the map unless null.  This is equivalent to
2830	>	* <pre> {@code
2831	>	* if (map.containsKey(key))
2832	>	*   return map.get(key);
2833	>	* value = mappingFunction.apply(key);
2834	>	* if (value != null)
2835	>	*   map.put(key, value);
2836	>	* return value;}</pre>
2837	>	*
2838	>	* except that the action is performed atomically.  If the
2839	>	* function returns {@code null} no mapping is recorded. If the
2840	>	* function itself throws an (unchecked) exception, the exception
2841	>	* is rethrown to its caller, and no mapping is recorded.  Some
2842	>	* attempted update operations on this map by other threads may be
2843	>	* blocked while computation is in progress, so the computation
2844	>	* should be short and simple, and must not attempt to update any
2845	>	* other mappings of this Map. The most appropriate usage is to
2846	>	* construct a new object serving as an initial mapped value, or
2847	>	* memoized result, as in:
2848	>	*
2849	>	*  <pre> {@code
2850	>	* map.computeIfAbsent(key, new Fun<K, V>() {
2851	>	*   public V map(K k) { return new Value(f(k)); }});}</pre>
2852	>	*
2853	>	* @param key key with which the specified value is to be associated
2854	>	* @param mappingFunction the function to compute a value
2855	>	* @return the current (existing or computed) value associated with
2856	>	*         the specified key, or null if the computed value is null
2857	>	* @throws NullPointerException if the specified key or mappingFunction
2858	>	*         is null
2859	>	* @throws IllegalStateException if the computation detectably
2860	>	*         attempts a recursive update to this map that would
2861	>	*         otherwise never complete
2862	>	* @throws RuntimeException or Error if the mappingFunction does so,
2863	>	*         in which case the mapping is left unestablished
2864	>	*/
2865	>	@SuppressWarnings("unchecked") public V computeIfAbsent
2866	>	(K key, Fun<? super K, ? extends V> mappingFunction) {
2867	>	if (key == null || mappingFunction == null)
2868	>	throw new NullPointerException();
2869	>	return (V)internalComputeIfAbsent(key, mappingFunction);
2870	>	}
2871	>
2872	>	/**
2873	>	* If the given key is present, computes a new mapping value given a key and
2874	>	* its current mapped value. This is equivalent to
2875	>	*  <pre> {@code
2876	>	*   if (map.containsKey(key)) {
2877	>	*     value = remappingFunction.apply(key, map.get(key));
2878	>	*     if (value != null)
2879	>	*       map.put(key, value);
2880	>	*     else
2881	>	*       map.remove(key);
2882	>	*   }
2883	>	* }</pre>
2884	>	*
2885	>	* except that the action is performed atomically.  If the
2886	>	* function returns {@code null}, the mapping is removed.  If the
2887	>	* function itself throws an (unchecked) exception, the exception
2888	>	* is rethrown to its caller, and the current mapping is left
2889	>	* unchanged.  Some attempted update operations on this map by
2890	>	* other threads may be blocked while computation is in progress,
2891	>	* so the computation should be short and simple, and must not
2892	>	* attempt to update any other mappings of this Map. For example,
2893	>	* to either create or append new messages to a value mapping:
2894	>	*
2895	>	* @param key key with which the specified value is to be associated
2896	>	* @param remappingFunction the function to compute a value
2897	>	* @return the new value associated with the specified key, or null if none
2898	>	* @throws NullPointerException if the specified key or remappingFunction
2899	>	*         is null
2900	>	* @throws IllegalStateException if the computation detectably
2901	>	*         attempts a recursive update to this map that would
2902	>	*         otherwise never complete
2903	>	* @throws RuntimeException or Error if the remappingFunction does so,
2904	>	*         in which case the mapping is unchanged
2905	>	*/
2906	>	@SuppressWarnings("unchecked") public V computeIfPresent
2907	>	(K key, BiFun<? super K, ? super V, ? extends V> remappingFunction) {
2908	>	if (key == null || remappingFunction == null)
2909	>	throw new NullPointerException();
2910	>	return (V)internalCompute(key, true, remappingFunction);
2911	>	}
2912	>
2913	>	/**
2914	>	* Computes a new mapping value given a key and
2915	>	* its current mapped value (or {@code null} if there is no current
2916	>	* mapping). This is equivalent to
2917	>	*  <pre> {@code
2918	>	*   value = remappingFunction.apply(key, map.get(key));
2919	>	*   if (value != null)
2920	>	*     map.put(key, value);
2921	>	*   else
2922	>	*     map.remove(key);
2923	>	* }</pre>
2924	>	*
2925	>	* except that the action is performed atomically.  If the
2926	>	* function returns {@code null}, the mapping is removed.  If the
2927	>	* function itself throws an (unchecked) exception, the exception
2928	>	* is rethrown to its caller, and the current mapping is left
2929	>	* unchanged.  Some attempted update operations on this map by
2930	>	* other threads may be blocked while computation is in progress,
2931	>	* so the computation should be short and simple, and must not
2932	>	* attempt to update any other mappings of this Map. For example,
2933	>	* to either create or append new messages to a value mapping:
2934	>	*
2935	>	* <pre> {@code
2936	>	* Map<Key, String> map = ...;
2937	>	* final String msg = ...;
2938	>	* map.compute(key, new BiFun<Key, String, String>() {
2939	>	*   public String apply(Key k, String v) {
2940	>	*    return (v == null) ? msg : v + msg;});}}</pre>
2941	>	*
2942	>	* @param key key with which the specified value is to be associated
2943	>	* @param remappingFunction the function to compute a value
2944	>	* @return the new value associated with the specified key, or null if none
2945	>	* @throws NullPointerException if the specified key or remappingFunction
2946	>	*         is null
2947	>	* @throws IllegalStateException if the computation detectably
2948	>	*         attempts a recursive update to this map that would
2949	>	*         otherwise never complete
2950	>	* @throws RuntimeException or Error if the remappingFunction does so,
2951	>	*         in which case the mapping is unchanged
2952	>	*/
2953	>	@SuppressWarnings("unchecked") public V compute
2954	>	(K key, BiFun<? super K, ? super V, ? extends V> remappingFunction) {
2955	>	if (key == null || remappingFunction == null)
2956	>	throw new NullPointerException();
2957	>	return (V)internalCompute(key, false, remappingFunction);
2958	>	}
2959	>
2960	>	/**
2961	>	* If the specified key is not already associated
2962	>	* with a value, associate it with the given value.
2963	>	* Otherwise, replace the value with the results of
2964	>	* the given remapping function. This is equivalent to:
2965	>	*  <pre> {@code
2966	>	*   if (!map.containsKey(key))
2967	>	*     map.put(value);
2968	>	*   else {
2969	>	*     newValue = remappingFunction.apply(map.get(key), value);
2970	>	*     if (value != null)
2971	>	*       map.put(key, value);
2972	>	*     else
2973	>	*       map.remove(key);
2974	>	*   }
2975	>	* }</pre>
2976	>	* except that the action is performed atomically.  If the
2977	>	* function returns {@code null}, the mapping is removed.  If the
2978	>	* function itself throws an (unchecked) exception, the exception
2979	>	* is rethrown to its caller, and the current mapping is left
2980	>	* unchanged.  Some attempted update operations on this map by
2981	>	* other threads may be blocked while computation is in progress,
2982	>	* so the computation should be short and simple, and must not
2983	>	* attempt to update any other mappings of this Map.
2984	>	*/
2985	>	@SuppressWarnings("unchecked") public V merge
2986	>	(K key, V value, BiFun<? super V, ? super V, ? extends V> remappingFunction) {
2987	>	if (key == null || value == null || remappingFunction == null)
2988	>	throw new NullPointerException();
2989	>	return (V)internalMerge(key, value, remappingFunction);
2990		}
2991
2992		/**
#	Line 910 | Line 2994 | public class ConcurrentHashMap<K, V> ext
2994		* This method does nothing if the key is not in the map.
2995		*
2996		* @param  key the key that needs to be removed
2997	<	* @return the previous value associated with <tt>key</tt>, or
2998	<	*         <tt>null</tt> if there was no mapping for <tt>key</tt>.
2997	>	* @return the previous value associated with {@code key}, or
2998	>	*         {@code null} if there was no mapping for {@code key}
2999		* @throws NullPointerException if the specified key is null
3000		*/
3001	<	public V remove(Object key) {
3002	<	int hash = hash(key);
3003	<	return segmentFor(hash).remove(key, hash, null);
3001	>	@SuppressWarnings("unchecked") public V remove(Object key) {
3002	>	if (key == null)
3003	>	throw new NullPointerException();
3004	>	return (V)internalReplace(key, null, null);
3005		}
3006
3007		/**
#	Line 925 | Line 3010 | public class ConcurrentHashMap<K, V> ext
3010		* @throws NullPointerException if the specified key is null
3011		*/
3012		public boolean remove(Object key, Object value) {
3013	+	if (key == null)
3014	+	throw new NullPointerException();
3015		if (value == null)
3016		return false;
3017	<	int hash = hash(key);
931	<	return segmentFor(hash).remove(key, hash, value) != null;
3017	>	return internalReplace(key, null, value) != null;
3018		}
3019
3020		/**
#	Line 937 | Line 3023 | public class ConcurrentHashMap<K, V> ext
3023		* @throws NullPointerException if any of the arguments are null
3024		*/
3025		public boolean replace(K key, V oldValue, V newValue) {
3026	<	if (oldValue == null || newValue == null)
3026	>	if (key == null || oldValue == null || newValue == null)
3027		throw new NullPointerException();
3028	<	int hash = hash(key);
943	<	return segmentFor(hash).replace(key, hash, oldValue, newValue);
3028	>	return internalReplace(key, newValue, oldValue) != null;
3029		}
3030
3031		/**
3032		* {@inheritDoc}
3033		*
3034		* @return the previous value associated with the specified key,
3035	<	*         or <tt>null</tt> if there was no mapping for the key
3035	>	*         or {@code null} if there was no mapping for the key
3036		* @throws NullPointerException if the specified key or value is null
3037		*/
3038	<	public V replace(K key, V value) {
3039	<	if (value == null)
3038	>	@SuppressWarnings("unchecked") public V replace(K key, V value) {
3039	>	if (key == null || value == null)
3040		throw new NullPointerException();
3041	<	int hash = hash(key);
957	<	return segmentFor(hash).replace(key, hash, value);
3041	>	return (V)internalReplace(key, value, null);
3042		}
3043
3044		/**
3045		* Removes all of the mappings from this map.
3046		*/
3047		public void clear() {
3048	<	for (int i = 0; i < segments.length; ++i)
965	<	segments[i].clear();
3048	>	internalClear();
3049		}
3050
3051		/**
3052		* Returns a {@link Set} view of the keys contained in this map.
3053		* The set is backed by the map, so changes to the map are
3054	<	* reflected in the set, and vice-versa.  The set supports element
972	<	* removal, which removes the corresponding mapping from this map,
973	<	* via the <tt>Iterator.remove</tt>, <tt>Set.remove</tt>,
974	<	* <tt>removeAll</tt>, <tt>retainAll</tt>, and <tt>clear</tt>
975	<	* operations.  It does not support the <tt>add</tt> or
976	<	* <tt>addAll</tt> operations.
3054	>	* reflected in the set, and vice-versa.
3055		*
3056	<	* <p>The view's <tt>iterator</tt> is a "weakly consistent" iterator
979	<	* that will never throw {@link ConcurrentModificationException},
980	<	* and guarantees to traverse elements as they existed upon
981	<	* construction of the iterator, and may (but is not guaranteed to)
982	<	* reflect any modifications subsequent to construction.
3056	>	* @return the set view
3057		*/
3058	<	public Set<K> keySet() {
3059	<	Set<K> ks = keySet;
3060	<	return (ks != null) ? ks : (keySet = new KeySet());
3058	>	public KeySetView<K,V> keySet() {
3059	>	KeySetView<K,V> ks = keySet;
3060	>	return (ks != null) ? ks : (keySet = new KeySetView<K,V>(this, null));
3061	>	}
3062	>
3063	>	/**
3064	>	* Returns a {@link Set} view of the keys in this map, using the
3065	>	* given common mapped value for any additions (i.e., {@link
3066	>	* Collection#add} and {@link Collection#addAll}). This is of
3067	>	* course only appropriate if it is acceptable to use the same
3068	>	* value for all additions from this view.
3069	>	*
3070	>	* @param mappedValue the mapped value to use for any
3071	>	* additions.
3072	>	* @return the set view
3073	>	* @throws NullPointerException if the mappedValue is null
3074	>	*/
3075	>	public KeySetView<K,V> keySet(V mappedValue) {
3076	>	if (mappedValue == null)
3077	>	throw new NullPointerException();
3078	>	return new KeySetView<K,V>(this, mappedValue);
3079		}
3080
3081		/**
#	Line 991 | Line 3083 | public class ConcurrentHashMap<K, V> ext
3083		* The collection is backed by the map, so changes to the map are
3084		* reflected in the collection, and vice-versa.  The collection
3085		* supports element removal, which removes the corresponding
3086	<	* mapping from this map, via the <tt>Iterator.remove</tt>,
3087	<	* <tt>Collection.remove</tt>, <tt>removeAll</tt>,
3088	<	* <tt>retainAll</tt>, and <tt>clear</tt> operations.  It does not
3089	<	* support the <tt>add</tt> or <tt>addAll</tt> operations.
3086	>	* mapping from this map, via the {@code Iterator.remove},
3087	>	* {@code Collection.remove}, {@code removeAll},
3088	>	* {@code retainAll}, and {@code clear} operations.  It does not
3089	>	* support the {@code add} or {@code addAll} operations.
3090		*
3091	<	* <p>The view's <tt>iterator</tt> is a "weakly consistent" iterator
3091	>	* <p>The view's {@code iterator} is a "weakly consistent" iterator
3092		* that will never throw {@link ConcurrentModificationException},
3093		* and guarantees to traverse elements as they existed upon
3094		* construction of the iterator, and may (but is not guaranteed to)
3095		* reflect any modifications subsequent to construction.
3096		*/
3097		public Collection<V> values() {
3098	<	Collection<V> vs = values;
3099	<	return (vs != null) ? vs : (values = new Values());
3098	>	Values<K,V> vs = values;
3099	>	return (vs != null) ? vs : (values = new Values<K,V>(this));
3100		}
3101
3102		/**
#	Line 1012 | Line 3104 | public class ConcurrentHashMap<K, V> ext
3104		* The set is backed by the map, so changes to the map are
3105		* reflected in the set, and vice-versa.  The set supports element
3106		* removal, which removes the corresponding mapping from the map,
3107	<	* via the <tt>Iterator.remove</tt>, <tt>Set.remove</tt>,
3108	<	* <tt>removeAll</tt>, <tt>retainAll</tt>, and <tt>clear</tt>
3109	<	* operations.  It does not support the <tt>add</tt> or
3110	<	* <tt>addAll</tt> operations.
3107	>	* via the {@code Iterator.remove}, {@code Set.remove},
3108	>	* {@code removeAll}, {@code retainAll}, and {@code clear}
3109	>	* operations.  It does not support the {@code add} or
3110	>	* {@code addAll} operations.
3111		*
3112	<	* <p>The view's <tt>iterator</tt> is a "weakly consistent" iterator
3112	>	* <p>The view's {@code iterator} is a "weakly consistent" iterator
3113		* that will never throw {@link ConcurrentModificationException},
3114		* and guarantees to traverse elements as they existed upon
3115		* construction of the iterator, and may (but is not guaranteed to)
3116		* reflect any modifications subsequent to construction.
3117		*/
3118		public Set<Map.Entry<K,V>> entrySet() {
3119	<	Set<Map.Entry<K,V>> es = entrySet;
3120	<	return (es != null) ? es : (entrySet = new EntrySet());
3119	>	EntrySet<K,V> es = entrySet;
3120	>	return (es != null) ? es : (entrySet = new EntrySet<K,V>(this));
3121		}
3122
3123		/**
3124		* Returns an enumeration of the keys in this table.
3125		*
3126		* @return an enumeration of the keys in this table
3127	<	* @see #keySet
3127	>	* @see #keySet()
3128		*/
3129		public Enumeration<K> keys() {
3130	<	return new KeyIterator();
3130	>	return new KeyIterator<K,V>(this);
3131		}
3132
3133		/**
3134		* Returns an enumeration of the values in this table.
3135		*
3136		* @return an enumeration of the values in this table
3137	<	* @see #values
3137	>	* @see #values()
3138		*/
3139		public Enumeration<V> elements() {
3140	<	return new ValueIterator();
3140	>	return new ValueIterator<K,V>(this);
3141		}
3142
3143	<	/* ---------------- Iterator Support -------------- */
3144	<
3145	<	abstract class HashIterator {
3146	<	int nextSegmentIndex;
3147	<	int nextTableIndex;
3148	<	HashEntry<K,V>[] currentTable;
3149	<	HashEntry<K, V> nextEntry;
3150	<	HashEntry<K, V> lastReturned;
3143	>	/**
3144	>	* Returns a partitionable iterator of the keys in this map.
3145	>	*
3146	>	* @return a partitionable iterator of the keys in this map
3147	>	*/
3148	>	public Spliterator<K> keySpliterator() {
3149	>	return new KeyIterator<K,V>(this);
3150	>	}
3151
3152	<	HashIterator() {
3153	<	nextSegmentIndex = segments.length - 1;
3154	<	nextTableIndex = -1;
3155	<	advance();
3156	<	}
3152	>	/**
3153	>	* Returns a partitionable iterator of the values in this map.
3154	>	*
3155	>	* @return a partitionable iterator of the values in this map
3156	>	*/
3157	>	public Spliterator<V> valueSpliterator() {
3158	>	return new ValueIterator<K,V>(this);
3159	>	}
3160
3161	<	public boolean hasMoreElements() { return hasNext(); }
3161	>	/**
3162	>	* Returns a partitionable iterator of the entries in this map.
3163	>	*
3164	>	* @return a partitionable iterator of the entries in this map
3165	>	*/
3166	>	public Spliterator<Map.Entry<K,V>> entrySpliterator() {
3167	>	return new EntryIterator<K,V>(this);
3168	>	}
3169
3170	<	final void advance() {
3171	<	if (nextEntry != null && (nextEntry = nextEntry.next) != null)
3172	<	return;
3170	>	/**
3171	>	* Returns the hash code value for this {@link Map}, i.e.,
3172	>	* the sum of, for each key-value pair in the map,
3173	>	* {@code key.hashCode() ^ value.hashCode()}.
3174	>	*
3175	>	* @return the hash code value for this map
3176	>	*/
3177	>	public int hashCode() {
3178	>	int h = 0;
3179	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
3180	>	Object v;
3181	>	while ((v = it.advance()) != null) {
3182	>	h += it.nextKey.hashCode() ^ v.hashCode();
3183	>	}
3184	>	return h;
3185	>	}
3186
3187	<	while (nextTableIndex >= 0) {
3188	<	if ( (nextEntry = currentTable[nextTableIndex--]) != null)
3189	<	return;
3187	>	/**
3188	>	* Returns a string representation of this map.  The string
3189	>	* representation consists of a list of key-value mappings (in no
3190	>	* particular order) enclosed in braces ("{@code {}}").  Adjacent
3191	>	* mappings are separated by the characters {@code ", "} (comma
3192	>	* and space).  Each key-value mapping is rendered as the key
3193	>	* followed by an equals sign ("{@code =}") followed by the
3194	>	* associated value.
3195	>	*
3196	>	* @return a string representation of this map
3197	>	*/
3198	>	public String toString() {
3199	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
3200	>	StringBuilder sb = new StringBuilder();
3201	>	sb.append('{');
3202	>	Object v;
3203	>	if ((v = it.advance()) != null) {
3204	>	for (;;) {
3205	>	Object k = it.nextKey;
3206	>	sb.append(k == this ? "(this Map)" : k);
3207	>	sb.append('=');
3208	>	sb.append(v == this ? "(this Map)" : v);
3209	>	if ((v = it.advance()) == null)
3210	>	break;
3211	>	sb.append(',').append(' ');
3212		}
3213	+	}
3214	+	return sb.append('}').toString();
3215	+	}
3216
3217	<	while (nextSegmentIndex >= 0) {
3218	<	Segment<K,V> seg = segments[nextSegmentIndex--];
3219	<	if (seg.count != 0) {
3220	<	currentTable = seg.table;
3221	<	for (int j = currentTable.length - 1; j >= 0; --j) {
3222	<	if ( (nextEntry = currentTable[j]) != null) {
3223	<	nextTableIndex = j - 1;
3224	<	return;
3225	<	}
3226	<	}
3227	<	}
3217	>	/**
3218	>	* Compares the specified object with this map for equality.
3219	>	* Returns {@code true} if the given object is a map with the same
3220	>	* mappings as this map.  This operation may return misleading
3221	>	* results if either map is concurrently modified during execution
3222	>	* of this method.
3223	>	*
3224	>	* @param o object to be compared for equality with this map
3225	>	* @return {@code true} if the specified object is equal to this map
3226	>	*/
3227	>	public boolean equals(Object o) {
3228	>	if (o != this) {
3229	>	if (!(o instanceof Map))
3230	>	return false;
3231	>	Map<?,?> m = (Map<?,?>) o;
3232	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
3233	>	Object val;
3234	>	while ((val = it.advance()) != null) {
3235	>	Object v = m.get(it.nextKey);
3236	>	if (v == null || (v != val && !v.equals(val)))
3237	>	return false;
3238	>	}
3239	>	for (Map.Entry<?,?> e : m.entrySet()) {
3240	>	Object mk, mv, v;
3241	>	if ((mk = e.getKey()) == null ||
3242	>	(mv = e.getValue()) == null ||
3243	>	(v = internalGet(mk)) == null ||
3244	>	(mv != v && !mv.equals(v)))
3245	>	return false;
3246		}
3247		}
3248	+	return true;
3249	+	}
3250
3251	<	public boolean hasNext() { return nextEntry != null; }
3251	>	/* ----------------Iterators -------------- */
3252
3253	<	HashEntry<K,V> nextEntry() {
3254	<	if (nextEntry == null)
3253	>	@SuppressWarnings("serial") static final class KeyIterator<K,V> extends Traverser<K,V,Object>
3254	>	implements Spliterator<K>, Enumeration<K> {
3255	>	KeyIterator(ConcurrentHashMap<K, V> map) { super(map); }
3256	>	KeyIterator(Traverser<K,V,Object> it) {
3257	>	super(it);
3258	>	}
3259	>	public KeyIterator<K,V> split() {
3260	>	if (nextKey != null)
3261	>	throw new IllegalStateException();
3262	>	return new KeyIterator<K,V>(this);
3263	>	}
3264	>	@SuppressWarnings("unchecked") public final K next() {
3265	>	if (nextVal == null && advance() == null)
3266		throw new NoSuchElementException();
3267	<	lastReturned = nextEntry;
3268	<	advance();
3269	<	return lastReturned;
3267	>	Object k = nextKey;
3268	>	nextVal = null;
3269	>	return (K) k;
3270		}
3271
3272	<	public void remove() {
3273	<	if (lastReturned == null)
3272	>	public final K nextElement() { return next(); }
3273	>	}
3274	>
3275	>	@SuppressWarnings("serial") static final class ValueIterator<K,V> extends Traverser<K,V,Object>
3276	>	implements Spliterator<V>, Enumeration<V> {
3277	>	ValueIterator(ConcurrentHashMap<K, V> map) { super(map); }
3278	>	ValueIterator(Traverser<K,V,Object> it) {
3279	>	super(it);
3280	>	}
3281	>	public ValueIterator<K,V> split() {
3282	>	if (nextKey != null)
3283		throw new IllegalStateException();
3284	<	ConcurrentHashMap.this.remove(lastReturned.key);
1105	<	lastReturned = null;
3284	>	return new ValueIterator<K,V>(this);
3285		}
1107	–	}
3286
3287	<	final class KeyIterator
3288	<	extends HashIterator
3289	<	implements Iterator<K>, Enumeration<K>
3290	<	{
3291	<	public K next()        { return super.nextEntry().key; }
3292	<	public K nextElement() { return super.nextEntry().key; }
3287	>	@SuppressWarnings("unchecked") public final V next() {
3288	>	Object v;
3289	>	if ((v = nextVal) == null && (v = advance()) == null)
3290	>	throw new NoSuchElementException();
3291	>	nextVal = null;
3292	>	return (V) v;
3293	>	}
3294	>
3295	>	public final V nextElement() { return next(); }
3296		}
3297
3298	<	final class ValueIterator
3299	<	extends HashIterator
3300	<	implements Iterator<V>, Enumeration<V>
3301	<	{
3302	<	public V next()        { return super.nextEntry().value; }
3303	<	public V nextElement() { return super.nextEntry().value; }
3298	>	@SuppressWarnings("serial") static final class EntryIterator<K,V> extends Traverser<K,V,Object>
3299	>	implements Spliterator<Map.Entry<K,V>> {
3300	>	EntryIterator(ConcurrentHashMap<K, V> map) { super(map); }
3301	>	EntryIterator(Traverser<K,V,Object> it) {
3302	>	super(it);
3303	>	}
3304	>	public EntryIterator<K,V> split() {
3305	>	if (nextKey != null)
3306	>	throw new IllegalStateException();
3307	>	return new EntryIterator<K,V>(this);
3308	>	}
3309	>
3310	>	@SuppressWarnings("unchecked") public final Map.Entry<K,V> next() {
3311	>	Object v;
3312	>	if ((v = nextVal) == null && (v = advance()) == null)
3313	>	throw new NoSuchElementException();
3314	>	Object k = nextKey;
3315	>	nextVal = null;
3316	>	return new MapEntry<K,V>((K)k, (V)v, map);
3317	>	}
3318		}
3319
3320		/**
3321	<	* Custom Entry class used by EntryIterator.next(), that relays
1127	<	* setValue changes to the underlying map.
3321	>	* Exported Entry for iterators
3322		*/
3323	<	final class WriteThroughEntry
3324	<	extends AbstractMap.SimpleEntry<K,V>
3325	<	{
3326	<	WriteThroughEntry(K k, V v) {
3327	<	super(k,v);
3323	>	static final class MapEntry<K,V> implements Map.Entry<K, V> {
3324	>	final K key; // non-null
3325	>	V val;       // non-null
3326	>	final ConcurrentHashMap<K, V> map;
3327	>	MapEntry(K key, V val, ConcurrentHashMap<K, V> map) {
3328	>	this.key = key;
3329	>	this.val = val;
3330	>	this.map = map;
3331	>	}
3332	>	public final K getKey()       { return key; }
3333	>	public final V getValue()     { return val; }
3334	>	public final int hashCode()   { return key.hashCode() ^ val.hashCode(); }
3335	>	public final String toString(){ return key + "=" + val; }
3336	>
3337	>	public final boolean equals(Object o) {
3338	>	Object k, v; Map.Entry<?,?> e;
3339	>	return ((o instanceof Map.Entry) &&
3340	>	(k = (e = (Map.Entry<?,?>)o).getKey()) != null &&
3341	>	(v = e.getValue()) != null &&
3342	>	(k == key || k.equals(key)) &&
3343	>	(v == val || v.equals(val)));
3344		}
3345
3346		/**
3347	<	* Set our entry's value and write through to the map. The
3348	<	* value to return is somewhat arbitrary here. Since a
3349	<	* WriteThroughEntry does not necessarily track asynchronous
3350	<	* changes, the most recent "previous" value could be
3351	<	* different from what we return (or could even have been
3352	<	* removed in which case the put will re-establish). We do not
1143	<	* and cannot guarantee more.
3347	>	* Sets our entry's value and writes through to the map. The
3348	>	* value to return is somewhat arbitrary here. Since we do not
3349	>	* necessarily track asynchronous changes, the most recent
3350	>	* "previous" value could be different from what we return (or
3351	>	* could even have been removed in which case the put will
3352	>	* re-establish). We do not and cannot guarantee more.
3353		*/
3354	<	public V setValue(V value) {
3354	>	public final V setValue(V value) {
3355		if (value == null) throw new NullPointerException();
3356	<	V v = super.setValue(value);
3357	<	ConcurrentHashMap.this.put(getKey(), value);
3356	>	V v = val;
3357	>	val = value;
3358	>	map.put(key, value);
3359		return v;
3360		}
3361		}
3362
3363	<	final class EntryIterator
3364	<	extends HashIterator
3365	<	implements Iterator<Entry<K,V>>
3366	<	{
3367	<	public Map.Entry<K,V> next() {
3368	<	HashEntry<K,V> e = super.nextEntry();
3369	<	return new WriteThroughEntry(e.key, e.value);
3363	>	/* ----------------Views -------------- */
3364	>
3365	>	/**
3366	>	* Base class for views.
3367	>	*/
3368	>	static abstract class CHMView<K, V> {
3369	>	final ConcurrentHashMap<K, V> map;
3370	>	CHMView(ConcurrentHashMap<K, V> map)  { this.map = map; }
3371	>	public final int size()                 { return map.size(); }
3372	>	public final boolean isEmpty()          { return map.isEmpty(); }
3373	>	public final void clear()               { map.clear(); }
3374	>
3375	>	// implementations below rely on concrete classes supplying these
3376	>	abstract public Iterator<?> iterator();
3377	>	abstract public boolean contains(Object o);
3378	>	abstract public boolean remove(Object o);
3379	>
3380	>	private static final String oomeMsg = "Required array size too large";
3381	>
3382	>	public final Object[] toArray() {
3383	>	long sz = map.mappingCount();
3384	>	if (sz > (long)(MAX_ARRAY_SIZE))
3385	>	throw new OutOfMemoryError(oomeMsg);
3386	>	int n = (int)sz;
3387	>	Object[] r = new Object[n];
3388	>	int i = 0;
3389	>	Iterator<?> it = iterator();
3390	>	while (it.hasNext()) {
3391	>	if (i == n) {
3392	>	if (n >= MAX_ARRAY_SIZE)
3393	>	throw new OutOfMemoryError(oomeMsg);
3394	>	if (n >= MAX_ARRAY_SIZE - (MAX_ARRAY_SIZE >>> 1) - 1)
3395	>	n = MAX_ARRAY_SIZE;
3396	>	else
3397	>	n += (n >>> 1) + 1;
3398	>	r = Arrays.copyOf(r, n);
3399	>	}
3400	>	r[i++] = it.next();
3401	>	}
3402	>	return (i == n) ? r : Arrays.copyOf(r, i);
3403	>	}
3404	>
3405	>	@SuppressWarnings("unchecked") public final <T> T[] toArray(T[] a) {
3406	>	long sz = map.mappingCount();
3407	>	if (sz > (long)(MAX_ARRAY_SIZE))
3408	>	throw new OutOfMemoryError(oomeMsg);
3409	>	int m = (int)sz;
3410	>	T[] r = (a.length >= m) ? a :
3411	>	(T[])java.lang.reflect.Array
3412	>	.newInstance(a.getClass().getComponentType(), m);
3413	>	int n = r.length;
3414	>	int i = 0;
3415	>	Iterator<?> it = iterator();
3416	>	while (it.hasNext()) {
3417	>	if (i == n) {
3418	>	if (n >= MAX_ARRAY_SIZE)
3419	>	throw new OutOfMemoryError(oomeMsg);
3420	>	if (n >= MAX_ARRAY_SIZE - (MAX_ARRAY_SIZE >>> 1) - 1)
3421	>	n = MAX_ARRAY_SIZE;
3422	>	else
3423	>	n += (n >>> 1) + 1;
3424	>	r = Arrays.copyOf(r, n);
3425	>	}
3426	>	r[i++] = (T)it.next();
3427	>	}
3428	>	if (a == r && i < n) {
3429	>	r[i] = null; // null-terminate
3430	>	return r;
3431	>	}
3432	>	return (i == n) ? r : Arrays.copyOf(r, i);
3433		}
3434	+
3435	+	public final int hashCode() {
3436	+	int h = 0;
3437	+	for (Iterator<?> it = iterator(); it.hasNext();)
3438	+	h += it.next().hashCode();
3439	+	return h;
3440	+	}
3441	+
3442	+	public final String toString() {
3443	+	StringBuilder sb = new StringBuilder();
3444	+	sb.append('[');
3445	+	Iterator<?> it = iterator();
3446	+	if (it.hasNext()) {
3447	+	for (;;) {
3448	+	Object e = it.next();
3449	+	sb.append(e == this ? "(this Collection)" : e);
3450	+	if (!it.hasNext())
3451	+	break;
3452	+	sb.append(',').append(' ');
3453	+	}
3454	+	}
3455	+	return sb.append(']').toString();
3456	+	}
3457	+
3458	+	public final boolean containsAll(Collection<?> c) {
3459	+	if (c != this) {
3460	+	for (Iterator<?> it = c.iterator(); it.hasNext();) {
3461	+	Object e = it.next();
3462	+	if (e == null || !contains(e))
3463	+	return false;
3464	+	}
3465	+	}
3466	+	return true;
3467	+	}
3468	+
3469	+	public final boolean removeAll(Collection<?> c) {
3470	+	boolean modified = false;
3471	+	for (Iterator<?> it = iterator(); it.hasNext();) {
3472	+	if (c.contains(it.next())) {
3473	+	it.remove();
3474	+	modified = true;
3475	+	}
3476	+	}
3477	+	return modified;
3478	+	}
3479	+
3480	+	public final boolean retainAll(Collection<?> c) {
3481	+	boolean modified = false;
3482	+	for (Iterator<?> it = iterator(); it.hasNext();) {
3483	+	if (!c.contains(it.next())) {
3484	+	it.remove();
3485	+	modified = true;
3486	+	}
3487	+	}
3488	+	return modified;
3489	+	}
3490	+
3491		}
3492
3493	<	final class KeySet extends AbstractSet<K> {
3494	<	public Iterator<K> iterator() {
3495	<	return new KeyIterator();
3493	>	static final class Values<K,V> extends CHMView<K,V>
3494	>	implements Collection<V> {
3495	>	Values(ConcurrentHashMap<K, V> map)   { super(map); }
3496	>	public final boolean contains(Object o) { return map.containsValue(o); }
3497	>	public final boolean remove(Object o) {
3498	>	if (o != null) {
3499	>	Iterator<V> it = new ValueIterator<K,V>(map);
3500	>	while (it.hasNext()) {
3501	>	if (o.equals(it.next())) {
3502	>	it.remove();
3503	>	return true;
3504	>	}
3505	>	}
3506	>	}
3507	>	return false;
3508		}
3509	<	public int size() {
3510	<	return ConcurrentHashMap.this.size();
3509	>	public final Iterator<V> iterator() {
3510	>	return new ValueIterator<K,V>(map);
3511		}
3512	<	public boolean contains(Object o) {
3513	<	return ConcurrentHashMap.this.containsKey(o);
3512	>	public final boolean add(V e) {
3513	>	throw new UnsupportedOperationException();
3514		}
3515	<	public boolean remove(Object o) {
3516	<	return ConcurrentHashMap.this.remove(o) != null;
3515	>	public final boolean addAll(Collection<? extends V> c) {
3516	>	throw new UnsupportedOperationException();
3517		}
3518	<	public void clear() {
3519	<	ConcurrentHashMap.this.clear();
3518	>
3519	>	}
3520	>
3521	>	static final class EntrySet<K,V> extends CHMView<K,V>
3522	>	implements Set<Map.Entry<K,V>> {
3523	>	EntrySet(ConcurrentHashMap<K, V> map) { super(map); }
3524	>	public final boolean contains(Object o) {
3525	>	Object k, v, r; Map.Entry<?,?> e;
3526	>	return ((o instanceof Map.Entry) &&
3527	>	(k = (e = (Map.Entry<?,?>)o).getKey()) != null &&
3528	>	(r = map.get(k)) != null &&
3529	>	(v = e.getValue()) != null &&
3530	>	(v == r || v.equals(r)));
3531	>	}
3532	>	public final boolean remove(Object o) {
3533	>	Object k, v; Map.Entry<?,?> e;
3534	>	return ((o instanceof Map.Entry) &&
3535	>	(k = (e = (Map.Entry<?,?>)o).getKey()) != null &&
3536	>	(v = e.getValue()) != null &&
3537	>	map.remove(k, v));
3538	>	}
3539	>	public final Iterator<Map.Entry<K,V>> iterator() {
3540	>	return new EntryIterator<K,V>(map);
3541	>	}
3542	>	public final boolean add(Entry<K,V> e) {
3543	>	throw new UnsupportedOperationException();
3544	>	}
3545	>	public final boolean addAll(Collection<? extends Entry<K,V>> c) {
3546	>	throw new UnsupportedOperationException();
3547	>	}
3548	>	public boolean equals(Object o) {
3549	>	Set<?> c;
3550	>	return ((o instanceof Set) &&
3551	>	((c = (Set<?>)o) == this ||
3552	>	(containsAll(c) && c.containsAll(this))));
3553		}
3554	<	public Object[] toArray() {
3555	<	Collection<K> c = new ArrayList<K>(size());
3556	<	for (K k : this)
3557	<	c.add(k);
3558	<	return c.toArray();
3554	>	}
3555	>
3556	>	/* ---------------- Serialization Support -------------- */
3557	>
3558	>	/**
3559	>	* Stripped-down version of helper class used in previous version,
3560	>	* declared for the sake of serialization compatibility
3561	>	*/
3562	>	static class Segment<K,V> implements Serializable {
3563	>	private static final long serialVersionUID = 2249069246763182397L;
3564	>	final float loadFactor;
3565	>	Segment(float lf) { this.loadFactor = lf; }
3566	>	}
3567	>
3568	>	/**
3569	>	* Saves the state of the {@code ConcurrentHashMap} instance to a
3570	>	* stream (i.e., serializes it).
3571	>	* @param s the stream
3572	>	* @serialData
3573	>	* the key (Object) and value (Object)
3574	>	* for each key-value mapping, followed by a null pair.
3575	>	* The key-value mappings are emitted in no particular order.
3576	>	*/
3577	>	@SuppressWarnings("unchecked") private void writeObject(java.io.ObjectOutputStream s)
3578	>	throws java.io.IOException {
3579	>	if (segments == null) { // for serialization compatibility
3580	>	segments = (Segment<K,V>[])
3581	>	new Segment<?,?>[DEFAULT_CONCURRENCY_LEVEL];
3582	>	for (int i = 0; i < segments.length; ++i)
3583	>	segments[i] = new Segment<K,V>(LOAD_FACTOR);
3584		}
3585	<	public <T> T[] toArray(T[] a) {
3586	<	Collection<K> c = new ArrayList<K>();
3587	<	for (K k : this)
3588	<	c.add(k);
3589	<	return c.toArray(a);
3585	>	s.defaultWriteObject();
3586	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
3587	>	Object v;
3588	>	while ((v = it.advance()) != null) {
3589	>	s.writeObject(it.nextKey);
3590	>	s.writeObject(v);
3591		}
3592	+	s.writeObject(null);
3593	+	s.writeObject(null);
3594	+	segments = null; // throw away
3595		}
3596
3597	<	final class Values extends AbstractCollection<V> {
3598	<	public Iterator<V> iterator() {
3599	<	return new ValueIterator();
3597	>	/**
3598	>	* Reconstitutes the instance from a stream (that is, deserializes it).
3599	>	* @param s the stream
3600	>	*/
3601	>	@SuppressWarnings("unchecked") private void readObject(java.io.ObjectInputStream s)
3602	>	throws java.io.IOException, ClassNotFoundException {
3603	>	s.defaultReadObject();
3604	>	this.segments = null; // unneeded
3605	>	// initialize transient final field
3606	>	UNSAFE.putObjectVolatile(this, counterOffset, new LongAdder());
3607	>
3608	>	// Create all nodes, then place in table once size is known
3609	>	long size = 0L;
3610	>	Node p = null;
3611	>	for (;;) {
3612	>	K k = (K) s.readObject();
3613	>	V v = (V) s.readObject();
3614	>	if (k != null && v != null) {
3615	>	int h = spread(k.hashCode());
3616	>	p = new Node(h, k, v, p);
3617	>	++size;
3618	>	}
3619	>	else
3620	>	break;
3621		}
3622	<	public int size() {
3623	<	return ConcurrentHashMap.this.size();
3622	>	if (p != null) {
3623	>	boolean init = false;
3624	>	int n;
3625	>	if (size >= (long)(MAXIMUM_CAPACITY >>> 1))
3626	>	n = MAXIMUM_CAPACITY;
3627	>	else {
3628	>	int sz = (int)size;
3629	>	n = tableSizeFor(sz + (sz >>> 1) + 1);
3630	>	}
3631	>	int sc = sizeCtl;
3632	>	boolean collide = false;
3633	>	if (n > sc &&
3634	>	UNSAFE.compareAndSwapInt(this, sizeCtlOffset, sc, -1)) {
3635	>	try {
3636	>	if (table == null) {
3637	>	init = true;
3638	>	Node[] tab = new Node[n];
3639	>	int mask = n - 1;
3640	>	while (p != null) {
3641	>	int j = p.hash & mask;
3642	>	Node next = p.next;
3643	>	Node q = p.next = tabAt(tab, j);
3644	>	setTabAt(tab, j, p);
3645	>	if (!collide && q != null && q.hash == p.hash)
3646	>	collide = true;
3647	>	p = next;
3648	>	}
3649	>	table = tab;
3650	>	counter.add(size);
3651	>	sc = n - (n >>> 2);
3652	>	}
3653	>	} finally {
3654	>	sizeCtl = sc;
3655	>	}
3656	>	if (collide) { // rescan and convert to TreeBins
3657	>	Node[] tab = table;
3658	>	for (int i = 0; i < tab.length; ++i) {
3659	>	int c = 0;
3660	>	for (Node e = tabAt(tab, i); e != null; e = e.next) {
3661	>	if (++c > TREE_THRESHOLD &&
3662	>	(e.key instanceof Comparable)) {
3663	>	replaceWithTreeBin(tab, i, e.key);
3664	>	break;
3665	>	}
3666	>	}
3667	>	}
3668	>	}
3669	>	}
3670	>	if (!init) { // Can only happen if unsafely published.
3671	>	while (p != null) {
3672	>	internalPut(p.key, p.val);
3673	>	p = p.next;
3674	>	}
3675	>	}
3676		}
3677	<	public boolean contains(Object o) {
3678	<	return ConcurrentHashMap.this.containsValue(o);
3677	>	}
3678	>
3679	>
3680	>	// -------------------------------------------------------
3681	>
3682	>	// Sams
3683	>	/** Interface describing a void action of one argument */
3684	>	public interface Action<A> { void apply(A a); }
3685	>	/** Interface describing a void action of two arguments */
3686	>	public interface BiAction<A,B> { void apply(A a, B b); }
3687	>	/** Interface describing a function of one argument */
3688	>	public interface Fun<A,T> { T apply(A a); }
3689	>	/** Interface describing a function of two arguments */
3690	>	public interface BiFun<A,B,T> { T apply(A a, B b); }
3691	>	/** Interface describing a function of no arguments */
3692	>	public interface Generator<T> { T apply(); }
3693	>	/** Interface describing a function mapping its argument to a double */
3694	>	public interface ObjectToDouble<A> { double apply(A a); }
3695	>	/** Interface describing a function mapping its argument to a long */
3696	>	public interface ObjectToLong<A> { long apply(A a); }
3697	>	/** Interface describing a function mapping its argument to an int */
3698	>	public interface ObjectToInt<A> {int apply(A a); }
3699	>	/** Interface describing a function mapping two arguments to a double */
3700	>	public interface ObjectByObjectToDouble<A,B> { double apply(A a, B b); }
3701	>	/** Interface describing a function mapping two arguments to a long */
3702	>	public interface ObjectByObjectToLong<A,B> { long apply(A a, B b); }
3703	>	/** Interface describing a function mapping two arguments to an int */
3704	>	public interface ObjectByObjectToInt<A,B> {int apply(A a, B b); }
3705	>	/** Interface describing a function mapping a double to a double */
3706	>	public interface DoubleToDouble { double apply(double a); }
3707	>	/** Interface describing a function mapping a long to a long */
3708	>	public interface LongToLong { long apply(long a); }
3709	>	/** Interface describing a function mapping an int to an int */
3710	>	public interface IntToInt { int apply(int a); }
3711	>	/** Interface describing a function mapping two doubles to a double */
3712	>	public interface DoubleByDoubleToDouble { double apply(double a, double b); }
3713	>	/** Interface describing a function mapping two longs to a long */
3714	>	public interface LongByLongToLong { long apply(long a, long b); }
3715	>	/** Interface describing a function mapping two ints to an int */
3716	>	public interface IntByIntToInt { int apply(int a, int b); }
3717	>
3718	>
3719	>	// -------------------------------------------------------
3720	>
3721	>	/**
3722	>	* Performs the given action for each (key, value).
3723	>	*
3724	>	* @param action the action
3725	>	*/
3726	>	public void forEach(BiAction<K,V> action) {
3727	>	ForkJoinTasks.forEach
3728	>	(this, action).invoke();
3729	>	}
3730	>
3731	>	/**
3732	>	* Performs the given action for each non-null transformation
3733	>	* of each (key, value).
3734	>	*
3735	>	* @param transformer a function returning the transformation
3736	>	* for an element, or null of there is no transformation (in
3737	>	* which case the action is not applied).
3738	>	* @param action the action
3739	>	*/
3740	>	public <U> void forEach(BiFun<? super K, ? super V, ? extends U> transformer,
3741	>	Action<U> action) {
3742	>	ForkJoinTasks.forEach
3743	>	(this, transformer, action).invoke();
3744	>	}
3745	>
3746	>	/**
3747	>	* Returns a non-null result from applying the given search
3748	>	* function on each (key, value), or null if none.  Upon
3749	>	* success, further element processing is suppressed and the
3750	>	* results of any other parallel invocations of the search
3751	>	* function are ignored.
3752	>	*
3753	>	* @param searchFunction a function returning a non-null
3754	>	* result on success, else null
3755	>	* @return a non-null result from applying the given search
3756	>	* function on each (key, value), or null if none
3757	>	*/
3758	>	public <U> U search(BiFun<? super K, ? super V, ? extends U> searchFunction) {
3759	>	return ForkJoinTasks.search
3760	>	(this, searchFunction).invoke();
3761	>	}
3762	>
3763	>	/**
3764	>	* Returns the result of accumulating the given transformation
3765	>	* of all (key, value) pairs using the given reducer to
3766	>	* combine values, or null if none.
3767	>	*
3768	>	* @param transformer a function returning the transformation
3769	>	* for an element, or null of there is no transformation (in
3770	>	* which case it is not combined).
3771	>	* @param reducer a commutative associative combining function
3772	>	* @return the result of accumulating the given transformation
3773	>	* of all (key, value) pairs
3774	>	*/
3775	>	public <U> U reduce(BiFun<? super K, ? super V, ? extends U> transformer,
3776	>	BiFun<? super U, ? super U, ? extends U> reducer) {
3777	>	return ForkJoinTasks.reduce
3778	>	(this, transformer, reducer).invoke();
3779	>	}
3780	>
3781	>	/**
3782	>	* Returns the result of accumulating the given transformation
3783	>	* of all (key, value) pairs using the given reducer to
3784	>	* combine values, and the given basis as an identity value.
3785	>	*
3786	>	* @param transformer a function returning the transformation
3787	>	* for an element
3788	>	* @param basis the identity (initial default value) for the reduction
3789	>	* @param reducer a commutative associative combining function
3790	>	* @return the result of accumulating the given transformation
3791	>	* of all (key, value) pairs
3792	>	*/
3793	>	public double reduceToDouble(ObjectByObjectToDouble<? super K, ? super V> transformer,
3794	>	double basis,
3795	>	DoubleByDoubleToDouble reducer) {
3796	>	return ForkJoinTasks.reduceToDouble
3797	>	(this, transformer, basis, reducer).invoke();
3798	>	}
3799	>
3800	>	/**
3801	>	* Returns the result of accumulating the given transformation
3802	>	* of all (key, value) pairs using the given reducer to
3803	>	* combine values, and the given basis as an identity value.
3804	>	*
3805	>	* @param transformer a function returning the transformation
3806	>	* for an element
3807	>	* @param basis the identity (initial default value) for the reduction
3808	>	* @param reducer a commutative associative combining function
3809	>	* @return the result of accumulating the given transformation
3810	>	* of all (key, value) pairs
3811	>	*/
3812	>	public long reduceToLong(ObjectByObjectToLong<? super K, ? super V> transformer,
3813	>	long basis,
3814	>	LongByLongToLong reducer) {
3815	>	return ForkJoinTasks.reduceToLong
3816	>	(this, transformer, basis, reducer).invoke();
3817	>	}
3818	>
3819	>	/**
3820	>	* Returns the result of accumulating the given transformation
3821	>	* of all (key, value) pairs using the given reducer to
3822	>	* combine values, and the given basis as an identity value.
3823	>	*
3824	>	* @param transformer a function returning the transformation
3825	>	* for an element
3826	>	* @param basis the identity (initial default value) for the reduction
3827	>	* @param reducer a commutative associative combining function
3828	>	* @return the result of accumulating the given transformation
3829	>	* of all (key, value) pairs
3830	>	*/
3831	>	public int reduceToInt(ObjectByObjectToInt<? super K, ? super V> transformer,
3832	>	int basis,
3833	>	IntByIntToInt reducer) {
3834	>	return ForkJoinTasks.reduceToInt
3835	>	(this, transformer, basis, reducer).invoke();
3836	>	}
3837	>
3838	>	/**
3839	>	* Performs the given action for each key.
3840	>	*
3841	>	* @param action the action
3842	>	*/
3843	>	public void forEachKey(Action<K> action) {
3844	>	ForkJoinTasks.forEachKey
3845	>	(this, action).invoke();
3846	>	}
3847	>
3848	>	/**
3849	>	* Performs the given action for each non-null transformation
3850	>	* of each key.
3851	>	*
3852	>	* @param transformer a function returning the transformation
3853	>	* for an element, or null of there is no transformation (in
3854	>	* which case the action is not applied).
3855	>	* @param action the action
3856	>	*/
3857	>	public <U> void forEachKey(Fun<? super K, ? extends U> transformer,
3858	>	Action<U> action) {
3859	>	ForkJoinTasks.forEachKey
3860	>	(this, transformer, action).invoke();
3861	>	}
3862	>
3863	>	/**
3864	>	* Returns a non-null result from applying the given search
3865	>	* function on each key, or null if none. Upon success,
3866	>	* further element processing is suppressed and the results of
3867	>	* any other parallel invocations of the search function are
3868	>	* ignored.
3869	>	*
3870	>	* @param searchFunction a function returning a non-null
3871	>	* result on success, else null
3872	>	* @return a non-null result from applying the given search
3873	>	* function on each key, or null if none
3874	>	*/
3875	>	public <U> U searchKeys(Fun<? super K, ? extends U> searchFunction) {
3876	>	return ForkJoinTasks.searchKeys
3877	>	(this, searchFunction).invoke();
3878	>	}
3879	>
3880	>	/**
3881	>	* Returns the result of accumulating all keys using the given
3882	>	* reducer to combine values, or null if none.
3883	>	*
3884	>	* @param reducer a commutative associative combining function
3885	>	* @return the result of accumulating all keys using the given
3886	>	* reducer to combine values, or null if none
3887	>	*/
3888	>	public K reduceKeys(BiFun<? super K, ? super K, ? extends K> reducer) {
3889	>	return ForkJoinTasks.reduceKeys
3890	>	(this, reducer).invoke();
3891	>	}
3892	>
3893	>	/**
3894	>	* Returns the result of accumulating the given transformation
3895	>	* of all keys using the given reducer to combine values, or
3896	>	* null if none.
3897	>	*
3898	>	* @param transformer a function returning the transformation
3899	>	* for an element, or null of there is no transformation (in
3900	>	* which case it is not combined).
3901	>	* @param reducer a commutative associative combining function
3902	>	* @return the result of accumulating the given transformation
3903	>	* of all keys
3904	>	*/
3905	>	public <U> U reduceKeys(Fun<? super K, ? extends U> transformer,
3906	>	BiFun<? super U, ? super U, ? extends U> reducer) {
3907	>	return ForkJoinTasks.reduceKeys
3908	>	(this, transformer, reducer).invoke();
3909	>	}
3910	>
3911	>	/**
3912	>	* Returns the result of accumulating the given transformation
3913	>	* of all keys using the given reducer to combine values, and
3914	>	* the given basis as an identity value.
3915	>	*
3916	>	* @param transformer a function returning the transformation
3917	>	* for an element
3918	>	* @param basis the identity (initial default value) for the reduction
3919	>	* @param reducer a commutative associative combining function
3920	>	* @return  the result of accumulating the given transformation
3921	>	* of all keys
3922	>	*/
3923	>	public double reduceKeysToDouble(ObjectToDouble<? super K> transformer,
3924	>	double basis,
3925	>	DoubleByDoubleToDouble reducer) {
3926	>	return ForkJoinTasks.reduceKeysToDouble
3927	>	(this, transformer, basis, reducer).invoke();
3928	>	}
3929	>
3930	>	/**
3931	>	* Returns the result of accumulating the given transformation
3932	>	* of all keys using the given reducer to combine values, and
3933	>	* the given basis as an identity value.
3934	>	*
3935	>	* @param transformer a function returning the transformation
3936	>	* for an element
3937	>	* @param basis the identity (initial default value) for the reduction
3938	>	* @param reducer a commutative associative combining function
3939	>	* @return the result of accumulating the given transformation
3940	>	* of all keys
3941	>	*/
3942	>	public long reduceKeysToLong(ObjectToLong<? super K> transformer,
3943	>	long basis,
3944	>	LongByLongToLong reducer) {
3945	>	return ForkJoinTasks.reduceKeysToLong
3946	>	(this, transformer, basis, reducer).invoke();
3947	>	}
3948	>
3949	>	/**
3950	>	* Returns the result of accumulating the given transformation
3951	>	* of all keys using the given reducer to combine values, and
3952	>	* the given basis as an identity value.
3953	>	*
3954	>	* @param transformer a function returning the transformation
3955	>	* for an element
3956	>	* @param basis the identity (initial default value) for the reduction
3957	>	* @param reducer a commutative associative combining function
3958	>	* @return the result of accumulating the given transformation
3959	>	* of all keys
3960	>	*/
3961	>	public int reduceKeysToInt(ObjectToInt<? super K> transformer,
3962	>	int basis,
3963	>	IntByIntToInt reducer) {
3964	>	return ForkJoinTasks.reduceKeysToInt
3965	>	(this, transformer, basis, reducer).invoke();
3966	>	}
3967	>
3968	>	/**
3969	>	* Performs the given action for each value.
3970	>	*
3971	>	* @param action the action
3972	>	*/
3973	>	public void forEachValue(Action<V> action) {
3974	>	ForkJoinTasks.forEachValue
3975	>	(this, action).invoke();
3976	>	}
3977	>
3978	>	/**
3979	>	* Performs the given action for each non-null transformation
3980	>	* of each value.
3981	>	*
3982	>	* @param transformer a function returning the transformation
3983	>	* for an element, or null of there is no transformation (in
3984	>	* which case the action is not applied).
3985	>	*/
3986	>	public <U> void forEachValue(Fun<? super V, ? extends U> transformer,
3987	>	Action<U> action) {
3988	>	ForkJoinTasks.forEachValue
3989	>	(this, transformer, action).invoke();
3990	>	}
3991	>
3992	>	/**
3993	>	* Returns a non-null result from applying the given search
3994	>	* function on each value, or null if none.  Upon success,
3995	>	* further element processing is suppressed and the results of
3996	>	* any other parallel invocations of the search function are
3997	>	* ignored.
3998	>	*
3999	>	* @param searchFunction a function returning a non-null
4000	>	* result on success, else null
4001	>	* @return a non-null result from applying the given search
4002	>	* function on each value, or null if none
4003	>	*
4004	>	*/
4005	>	public <U> U searchValues(Fun<? super V, ? extends U> searchFunction) {
4006	>	return ForkJoinTasks.searchValues
4007	>	(this, searchFunction).invoke();
4008	>	}
4009	>
4010	>	/**
4011	>	* Returns the result of accumulating all values using the
4012	>	* given reducer to combine values, or null if none.
4013	>	*
4014	>	* @param reducer a commutative associative combining function
4015	>	* @return  the result of accumulating all values
4016	>	*/
4017	>	public V reduceValues(BiFun<? super V, ? super V, ? extends V> reducer) {
4018	>	return ForkJoinTasks.reduceValues
4019	>	(this, reducer).invoke();
4020	>	}
4021	>
4022	>	/**
4023	>	* Returns the result of accumulating the given transformation
4024	>	* of all values using the given reducer to combine values, or
4025	>	* null if none.
4026	>	*
4027	>	* @param transformer a function returning the transformation
4028	>	* for an element, or null of there is no transformation (in
4029	>	* which case it is not combined).
4030	>	* @param reducer a commutative associative combining function
4031	>	* @return the result of accumulating the given transformation
4032	>	* of all values
4033	>	*/
4034	>	public <U> U reduceValues(Fun<? super V, ? extends U> transformer,
4035	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4036	>	return ForkJoinTasks.reduceValues
4037	>	(this, transformer, reducer).invoke();
4038	>	}
4039	>
4040	>	/**
4041	>	* Returns the result of accumulating the given transformation
4042	>	* of all values using the given reducer to combine values,
4043	>	* and the given basis as an identity value.
4044	>	*
4045	>	* @param transformer a function returning the transformation
4046	>	* for an element
4047	>	* @param basis the identity (initial default value) for the reduction
4048	>	* @param reducer a commutative associative combining function
4049	>	* @return the result of accumulating the given transformation
4050	>	* of all values
4051	>	*/
4052	>	public double reduceValuesToDouble(ObjectToDouble<? super V> transformer,
4053	>	double basis,
4054	>	DoubleByDoubleToDouble reducer) {
4055	>	return ForkJoinTasks.reduceValuesToDouble
4056	>	(this, transformer, basis, reducer).invoke();
4057	>	}
4058	>
4059	>	/**
4060	>	* Returns the result of accumulating the given transformation
4061	>	* of all values using the given reducer to combine values,
4062	>	* and the given basis as an identity value.
4063	>	*
4064	>	* @param transformer a function returning the transformation
4065	>	* for an element
4066	>	* @param basis the identity (initial default value) for the reduction
4067	>	* @param reducer a commutative associative combining function
4068	>	* @return the result of accumulating the given transformation
4069	>	* of all values
4070	>	*/
4071	>	public long reduceValuesToLong(ObjectToLong<? super V> transformer,
4072	>	long basis,
4073	>	LongByLongToLong reducer) {
4074	>	return ForkJoinTasks.reduceValuesToLong
4075	>	(this, transformer, basis, reducer).invoke();
4076	>	}
4077	>
4078	>	/**
4079	>	* Returns the result of accumulating the given transformation
4080	>	* of all values using the given reducer to combine values,
4081	>	* and the given basis as an identity value.
4082	>	*
4083	>	* @param transformer a function returning the transformation
4084	>	* for an element
4085	>	* @param basis the identity (initial default value) for the reduction
4086	>	* @param reducer a commutative associative combining function
4087	>	* @return the result of accumulating the given transformation
4088	>	* of all values
4089	>	*/
4090	>	public int reduceValuesToInt(ObjectToInt<? super V> transformer,
4091	>	int basis,
4092	>	IntByIntToInt reducer) {
4093	>	return ForkJoinTasks.reduceValuesToInt
4094	>	(this, transformer, basis, reducer).invoke();
4095	>	}
4096	>
4097	>	/**
4098	>	* Performs the given action for each entry.
4099	>	*
4100	>	* @param action the action
4101	>	*/
4102	>	public void forEachEntry(Action<Map.Entry<K,V>> action) {
4103	>	ForkJoinTasks.forEachEntry
4104	>	(this, action).invoke();
4105	>	}
4106	>
4107	>	/**
4108	>	* Performs the given action for each non-null transformation
4109	>	* of each entry.
4110	>	*
4111	>	* @param transformer a function returning the transformation
4112	>	* for an element, or null of there is no transformation (in
4113	>	* which case the action is not applied).
4114	>	* @param action the action
4115	>	*/
4116	>	public <U> void forEachEntry(Fun<Map.Entry<K,V>, ? extends U> transformer,
4117	>	Action<U> action) {
4118	>	ForkJoinTasks.forEachEntry
4119	>	(this, transformer, action).invoke();
4120	>	}
4121	>
4122	>	/**
4123	>	* Returns a non-null result from applying the given search
4124	>	* function on each entry, or null if none.  Upon success,
4125	>	* further element processing is suppressed and the results of
4126	>	* any other parallel invocations of the search function are
4127	>	* ignored.
4128	>	*
4129	>	* @param searchFunction a function returning a non-null
4130	>	* result on success, else null
4131	>	* @return a non-null result from applying the given search
4132	>	* function on each entry, or null if none
4133	>	*/
4134	>	public <U> U searchEntries(Fun<Map.Entry<K,V>, ? extends U> searchFunction) {
4135	>	return ForkJoinTasks.searchEntries
4136	>	(this, searchFunction).invoke();
4137	>	}
4138	>
4139	>	/**
4140	>	* Returns the result of accumulating all entries using the
4141	>	* given reducer to combine values, or null if none.
4142	>	*
4143	>	* @param reducer a commutative associative combining function
4144	>	* @return the result of accumulating all entries
4145	>	*/
4146	>	public Map.Entry<K,V> reduceEntries(BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer) {
4147	>	return ForkJoinTasks.reduceEntries
4148	>	(this, reducer).invoke();
4149	>	}
4150	>
4151	>	/**
4152	>	* Returns the result of accumulating the given transformation
4153	>	* of all entries using the given reducer to combine values,
4154	>	* or null if none.
4155	>	*
4156	>	* @param transformer a function returning the transformation
4157	>	* for an element, or null of there is no transformation (in
4158	>	* which case it is not combined).
4159	>	* @param reducer a commutative associative combining function
4160	>	* @return the result of accumulating the given transformation
4161	>	* of all entries
4162	>	*/
4163	>	public <U> U reduceEntries(Fun<Map.Entry<K,V>, ? extends U> transformer,
4164	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4165	>	return ForkJoinTasks.reduceEntries
4166	>	(this, transformer, reducer).invoke();
4167	>	}
4168	>
4169	>	/**
4170	>	* Returns the result of accumulating the given transformation
4171	>	* of all entries using the given reducer to combine values,
4172	>	* and the given basis as an identity value.
4173	>	*
4174	>	* @param transformer a function returning the transformation
4175	>	* for an element
4176	>	* @param basis the identity (initial default value) for the reduction
4177	>	* @param reducer a commutative associative combining function
4178	>	* @return the result of accumulating the given transformation
4179	>	* of all entries
4180	>	*/
4181	>	public double reduceEntriesToDouble(ObjectToDouble<Map.Entry<K,V>> transformer,
4182	>	double basis,
4183	>	DoubleByDoubleToDouble reducer) {
4184	>	return ForkJoinTasks.reduceEntriesToDouble
4185	>	(this, transformer, basis, reducer).invoke();
4186	>	}
4187	>
4188	>	/**
4189	>	* Returns the result of accumulating the given transformation
4190	>	* of all entries using the given reducer to combine values,
4191	>	* and the given basis as an identity value.
4192	>	*
4193	>	* @param transformer a function returning the transformation
4194	>	* for an element
4195	>	* @param basis the identity (initial default value) for the reduction
4196	>	* @param reducer a commutative associative combining function
4197	>	* @return  the result of accumulating the given transformation
4198	>	* of all entries
4199	>	*/
4200	>	public long reduceEntriesToLong(ObjectToLong<Map.Entry<K,V>> transformer,
4201	>	long basis,
4202	>	LongByLongToLong reducer) {
4203	>	return ForkJoinTasks.reduceEntriesToLong
4204	>	(this, transformer, basis, reducer).invoke();
4205	>	}
4206	>
4207	>	/**
4208	>	* Returns the result of accumulating the given transformation
4209	>	* of all entries using the given reducer to combine values,
4210	>	* and the given basis as an identity value.
4211	>	*
4212	>	* @param transformer a function returning the transformation
4213	>	* for an element
4214	>	* @param basis the identity (initial default value) for the reduction
4215	>	* @param reducer a commutative associative combining function
4216	>	* @return the result of accumulating the given transformation
4217	>	* of all entries
4218	>	*/
4219	>	public int reduceEntriesToInt(ObjectToInt<Map.Entry<K,V>> transformer,
4220	>	int basis,
4221	>	IntByIntToInt reducer) {
4222	>	return ForkJoinTasks.reduceEntriesToInt
4223	>	(this, transformer, basis, reducer).invoke();
4224	>	}
4225	>
4226	>	// ---------------------------------------------------------------------
4227	>
4228	>	/**
4229	>	* Predefined tasks for performing bulk parallel operations on
4230	>	* ConcurrentHashMaps. These tasks follow the forms and rules used
4231	>	* for bulk operations. Each method has the same name, but returns
4232	>	* a task rather than invoking it. These methods may be useful in
4233	>	* custom applications such as submitting a task without waiting
4234	>	* for completion, using a custom pool, or combining with other
4235	>	* tasks.
4236	>	*/
4237	>	public static class ForkJoinTasks {
4238	>	private ForkJoinTasks() {}
4239	>
4240	>	/**
4241	>	* Returns a task that when invoked, performs the given
4242	>	* action for each (key, value)
4243	>	*
4244	>	* @param map the map
4245	>	* @param action the action
4246	>	* @return the task
4247	>	*/
4248	>	public static <K,V> ForkJoinTask<Void> forEach
4249	>	(ConcurrentHashMap<K,V> map,
4250	>	BiAction<K,V> action) {
4251	>	if (action == null) throw new NullPointerException();
4252	>	return new ForEachMappingTask<K,V>(map, null, -1, null, action);
4253		}
4254	<	public void clear() {
4255	<	ConcurrentHashMap.this.clear();
4254	>
4255	>	/**
4256	>	* Returns a task that when invoked, performs the given
4257	>	* action for each non-null transformation of each (key, value)
4258	>	*
4259	>	* @param map the map
4260	>	* @param transformer a function returning the transformation
4261	>	* for an element, or null if there is no transformation (in
4262	>	* which case the action is not applied)
4263	>	* @param action the action
4264	>	* @return the task
4265	>	*/
4266	>	public static <K,V,U> ForkJoinTask<Void> forEach
4267	>	(ConcurrentHashMap<K,V> map,
4268	>	BiFun<? super K, ? super V, ? extends U> transformer,
4269	>	Action<U> action) {
4270	>	if (transformer == null || action == null)
4271	>	throw new NullPointerException();
4272	>	return new ForEachTransformedMappingTask<K,V,U>
4273	>	(map, null, -1, null, transformer, action);
4274		}
4275	<	public Object[] toArray() {
4276	<	Collection<V> c = new ArrayList<V>(size());
4277	<	for (V v : this)
4278	<	c.add(v);
4279	<	return c.toArray();
4275	>
4276	>	/**
4277	>	* Returns a task that when invoked, returns a non-null result
4278	>	* from applying the given search function on each (key,
4279	>	* value), or null if none. Upon success, further element
4280	>	* processing is suppressed and the results of any other
4281	>	* parallel invocations of the search function are ignored.
4282	>	*
4283	>	* @param map the map
4284	>	* @param searchFunction a function returning a non-null
4285	>	* result on success, else null
4286	>	* @return the task
4287	>	*/
4288	>	public static <K,V,U> ForkJoinTask<U> search
4289	>	(ConcurrentHashMap<K,V> map,
4290	>	BiFun<? super K, ? super V, ? extends U> searchFunction) {
4291	>	if (searchFunction == null) throw new NullPointerException();
4292	>	return new SearchMappingsTask<K,V,U>
4293	>	(map, null, -1, null, searchFunction,
4294	>	new AtomicReference<U>());
4295		}
4296	<	public <T> T[] toArray(T[] a) {
4297	<	Collection<V> c = new ArrayList<V>(size());
4298	<	for (V v : this)
4299	<	c.add(v);
4300	<	return c.toArray(a);
4296	>
4297	>	/**
4298	>	* Returns a task that when invoked, returns the result of
4299	>	* accumulating the given transformation of all (key, value) pairs
4300	>	* using the given reducer to combine values, or null if none.
4301	>	*
4302	>	* @param map the map
4303	>	* @param transformer a function returning the transformation
4304	>	* for an element, or null if there is no transformation (in
4305	>	* which case it is not combined).
4306	>	* @param reducer a commutative associative combining function
4307	>	* @return the task
4308	>	*/
4309	>	public static <K,V,U> ForkJoinTask<U> reduce
4310	>	(ConcurrentHashMap<K,V> map,
4311	>	BiFun<? super K, ? super V, ? extends U> transformer,
4312	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4313	>	if (transformer == null || reducer == null)
4314	>	throw new NullPointerException();
4315	>	return new MapReduceMappingsTask<K,V,U>
4316	>	(map, null, -1, null, transformer, reducer);
4317		}
1218	–	}
4318
4319	<	final class EntrySet extends AbstractSet<Map.Entry<K,V>> {
4320	<	public Iterator<Map.Entry<K,V>> iterator() {
4321	<	return new EntryIterator();
4319	>	/**
4320	>	* Returns a task that when invoked, returns the result of
4321	>	* accumulating the given transformation of all (key, value) pairs
4322	>	* using the given reducer to combine values, and the given
4323	>	* basis as an identity value.
4324	>	*
4325	>	* @param map the map
4326	>	* @param transformer a function returning the transformation
4327	>	* for an element
4328	>	* @param basis the identity (initial default value) for the reduction
4329	>	* @param reducer a commutative associative combining function
4330	>	* @return the task
4331	>	*/
4332	>	public static <K,V> ForkJoinTask<Double> reduceToDouble
4333	>	(ConcurrentHashMap<K,V> map,
4334	>	ObjectByObjectToDouble<? super K, ? super V> transformer,
4335	>	double basis,
4336	>	DoubleByDoubleToDouble reducer) {
4337	>	if (transformer == null || reducer == null)
4338	>	throw new NullPointerException();
4339	>	return new MapReduceMappingsToDoubleTask<K,V>
4340	>	(map, null, -1, null, transformer, basis, reducer);
4341		}
4342	<	public boolean contains(Object o) {
4343	<	if (!(o instanceof Map.Entry))
4344	<	return false;
4345	<	Map.Entry<?,?> e = (Map.Entry<?,?>)o;
4346	<	V v = ConcurrentHashMap.this.get(e.getKey());
4347	<	return v != null && v.equals(e.getValue());
4342	>
4343	>	/**
4344	>	* Returns a task that when invoked, returns the result of
4345	>	* accumulating the given transformation of all (key, value) pairs
4346	>	* using the given reducer to combine values, and the given
4347	>	* basis as an identity value.
4348	>	*
4349	>	* @param map the map
4350	>	* @param transformer a function returning the transformation
4351	>	* for an element
4352	>	* @param basis the identity (initial default value) for the reduction
4353	>	* @param reducer a commutative associative combining function
4354	>	* @return the task
4355	>	*/
4356	>	public static <K,V> ForkJoinTask<Long> reduceToLong
4357	>	(ConcurrentHashMap<K,V> map,
4358	>	ObjectByObjectToLong<? super K, ? super V> transformer,
4359	>	long basis,
4360	>	LongByLongToLong reducer) {
4361	>	if (transformer == null || reducer == null)
4362	>	throw new NullPointerException();
4363	>	return new MapReduceMappingsToLongTask<K,V>
4364	>	(map, null, -1, null, transformer, basis, reducer);
4365		}
4366	<	public boolean remove(Object o) {
4367	<	if (!(o instanceof Map.Entry))
4368	<	return false;
4369	<	Map.Entry<?,?> e = (Map.Entry<?,?>)o;
4370	<	return ConcurrentHashMap.this.remove(e.getKey(), e.getValue());
4366	>
4367	>	/**
4368	>	* Returns a task that when invoked, returns the result of
4369	>	* accumulating the given transformation of all (key, value) pairs
4370	>	* using the given reducer to combine values, and the given
4371	>	* basis as an identity value.
4372	>	*
4373	>	* @param transformer a function returning the transformation
4374	>	* for an element
4375	>	* @param basis the identity (initial default value) for the reduction
4376	>	* @param reducer a commutative associative combining function
4377	>	* @return the task
4378	>	*/
4379	>	public static <K,V> ForkJoinTask<Integer> reduceToInt
4380	>	(ConcurrentHashMap<K,V> map,
4381	>	ObjectByObjectToInt<? super K, ? super V> transformer,
4382	>	int basis,
4383	>	IntByIntToInt reducer) {
4384	>	if (transformer == null || reducer == null)
4385	>	throw new NullPointerException();
4386	>	return new MapReduceMappingsToIntTask<K,V>
4387	>	(map, null, -1, null, transformer, basis, reducer);
4388	>	}
4389	>
4390	>	/**
4391	>	* Returns a task that when invoked, performs the given action
4392	>	* for each key.
4393	>	*
4394	>	* @param map the map
4395	>	* @param action the action
4396	>	* @return the task
4397	>	*/
4398	>	public static <K,V> ForkJoinTask<Void> forEachKey
4399	>	(ConcurrentHashMap<K,V> map,
4400	>	Action<K> action) {
4401	>	if (action == null) throw new NullPointerException();
4402	>	return new ForEachKeyTask<K,V>(map, null, -1, null, action);
4403	>	}
4404	>
4405	>	/**
4406	>	* Returns a task that when invoked, performs the given action
4407	>	* for each non-null transformation of each key.
4408	>	*
4409	>	* @param map the map
4410	>	* @param transformer a function returning the transformation
4411	>	* for an element, or null if there is no transformation (in
4412	>	* which case the action is not applied)
4413	>	* @param action the action
4414	>	* @return the task
4415	>	*/
4416	>	public static <K,V,U> ForkJoinTask<Void> forEachKey
4417	>	(ConcurrentHashMap<K,V> map,
4418	>	Fun<? super K, ? extends U> transformer,
4419	>	Action<U> action) {
4420	>	if (transformer == null || action == null)
4421	>	throw new NullPointerException();
4422	>	return new ForEachTransformedKeyTask<K,V,U>
4423	>	(map, null, -1, null, transformer, action);
4424	>	}
4425	>
4426	>	/**
4427	>	* Returns a task that when invoked, returns a non-null result
4428	>	* from applying the given search function on each key, or
4429	>	* null if none.  Upon success, further element processing is
4430	>	* suppressed and the results of any other parallel
4431	>	* invocations of the search function are ignored.
4432	>	*
4433	>	* @param map the map
4434	>	* @param searchFunction a function returning a non-null
4435	>	* result on success, else null
4436	>	* @return the task
4437	>	*/
4438	>	public static <K,V,U> ForkJoinTask<U> searchKeys
4439	>	(ConcurrentHashMap<K,V> map,
4440	>	Fun<? super K, ? extends U> searchFunction) {
4441	>	if (searchFunction == null) throw new NullPointerException();
4442	>	return new SearchKeysTask<K,V,U>
4443	>	(map, null, -1, null, searchFunction,
4444	>	new AtomicReference<U>());
4445	>	}
4446	>
4447	>	/**
4448	>	* Returns a task that when invoked, returns the result of
4449	>	* accumulating all keys using the given reducer to combine
4450	>	* values, or null if none.
4451	>	*
4452	>	* @param map the map
4453	>	* @param reducer a commutative associative combining function
4454	>	* @return the task
4455	>	*/
4456	>	public static <K,V> ForkJoinTask<K> reduceKeys
4457	>	(ConcurrentHashMap<K,V> map,
4458	>	BiFun<? super K, ? super K, ? extends K> reducer) {
4459	>	if (reducer == null) throw new NullPointerException();
4460	>	return new ReduceKeysTask<K,V>
4461	>	(map, null, -1, null, reducer);
4462	>	}
4463	>
4464	>	/**
4465	>	* Returns a task that when invoked, returns the result of
4466	>	* accumulating the given transformation of all keys using the given
4467	>	* reducer to combine values, or null if none.
4468	>	*
4469	>	* @param map the map
4470	>	* @param transformer a function returning the transformation
4471	>	* for an element, or null if there is no transformation (in
4472	>	* which case it is not combined).
4473	>	* @param reducer a commutative associative combining function
4474	>	* @return the task
4475	>	*/
4476	>	public static <K,V,U> ForkJoinTask<U> reduceKeys
4477	>	(ConcurrentHashMap<K,V> map,
4478	>	Fun<? super K, ? extends U> transformer,
4479	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4480	>	if (transformer == null || reducer == null)
4481	>	throw new NullPointerException();
4482	>	return new MapReduceKeysTask<K,V,U>
4483	>	(map, null, -1, null, transformer, reducer);
4484	>	}
4485	>
4486	>	/**
4487	>	* Returns a task that when invoked, returns the result of
4488	>	* accumulating the given transformation of all keys using the given
4489	>	* reducer to combine values, and the given basis as an
4490	>	* identity value.
4491	>	*
4492	>	* @param map the map
4493	>	* @param transformer a function returning the transformation
4494	>	* for an element
4495	>	* @param basis the identity (initial default value) for the reduction
4496	>	* @param reducer a commutative associative combining function
4497	>	* @return the task
4498	>	*/
4499	>	public static <K,V> ForkJoinTask<Double> reduceKeysToDouble
4500	>	(ConcurrentHashMap<K,V> map,
4501	>	ObjectToDouble<? super K> transformer,
4502	>	double basis,
4503	>	DoubleByDoubleToDouble reducer) {
4504	>	if (transformer == null || reducer == null)
4505	>	throw new NullPointerException();
4506	>	return new MapReduceKeysToDoubleTask<K,V>
4507	>	(map, null, -1, null, transformer, basis, reducer);
4508	>	}
4509	>
4510	>	/**
4511	>	* Returns a task that when invoked, returns the result of
4512	>	* accumulating the given transformation of all keys using the given
4513	>	* reducer to combine values, and the given basis as an
4514	>	* identity value.
4515	>	*
4516	>	* @param map the map
4517	>	* @param transformer a function returning the transformation
4518	>	* for an element
4519	>	* @param basis the identity (initial default value) for the reduction
4520	>	* @param reducer a commutative associative combining function
4521	>	* @return the task
4522	>	*/
4523	>	public static <K,V> ForkJoinTask<Long> reduceKeysToLong
4524	>	(ConcurrentHashMap<K,V> map,
4525	>	ObjectToLong<? super K> transformer,
4526	>	long basis,
4527	>	LongByLongToLong reducer) {
4528	>	if (transformer == null || reducer == null)
4529	>	throw new NullPointerException();
4530	>	return new MapReduceKeysToLongTask<K,V>
4531	>	(map, null, -1, null, transformer, basis, reducer);
4532	>	}
4533	>
4534	>	/**
4535	>	* Returns a task that when invoked, returns the result of
4536	>	* accumulating the given transformation of all keys using the given
4537	>	* reducer to combine values, and the given basis as an
4538	>	* identity value.
4539	>	*
4540	>	* @param map the map
4541	>	* @param transformer a function returning the transformation
4542	>	* for an element
4543	>	* @param basis the identity (initial default value) for the reduction
4544	>	* @param reducer a commutative associative combining function
4545	>	* @return the task
4546	>	*/
4547	>	public static <K,V> ForkJoinTask<Integer> reduceKeysToInt
4548	>	(ConcurrentHashMap<K,V> map,
4549	>	ObjectToInt<? super K> transformer,
4550	>	int basis,
4551	>	IntByIntToInt reducer) {
4552	>	if (transformer == null || reducer == null)
4553	>	throw new NullPointerException();
4554	>	return new MapReduceKeysToIntTask<K,V>
4555	>	(map, null, -1, null, transformer, basis, reducer);
4556	>	}
4557	>
4558	>	/**
4559	>	* Returns a task that when invoked, performs the given action
4560	>	* for each value.
4561	>	*
4562	>	* @param map the map
4563	>	* @param action the action
4564	>	*/
4565	>	public static <K,V> ForkJoinTask<Void> forEachValue
4566	>	(ConcurrentHashMap<K,V> map,
4567	>	Action<V> action) {
4568	>	if (action == null) throw new NullPointerException();
4569	>	return new ForEachValueTask<K,V>(map, null, -1, null, action);
4570	>	}
4571	>
4572	>	/**
4573	>	* Returns a task that when invoked, performs the given action
4574	>	* for each non-null transformation of each value.
4575	>	*
4576	>	* @param map the map
4577	>	* @param transformer a function returning the transformation
4578	>	* for an element, or null if there is no transformation (in
4579	>	* which case the action is not applied)
4580	>	* @param action the action
4581	>	*/
4582	>	public static <K,V,U> ForkJoinTask<Void> forEachValue
4583	>	(ConcurrentHashMap<K,V> map,
4584	>	Fun<? super V, ? extends U> transformer,
4585	>	Action<U> action) {
4586	>	if (transformer == null || action == null)
4587	>	throw new NullPointerException();
4588	>	return new ForEachTransformedValueTask<K,V,U>
4589	>	(map, null, -1, null, transformer, action);
4590	>	}
4591	>
4592	>	/**
4593	>	* Returns a task that when invoked, returns a non-null result
4594	>	* from applying the given search function on each value, or
4595	>	* null if none.  Upon success, further element processing is
4596	>	* suppressed and the results of any other parallel
4597	>	* invocations of the search function are ignored.
4598	>	*
4599	>	* @param map the map
4600	>	* @param searchFunction a function returning a non-null
4601	>	* result on success, else null
4602	>	* @return the task
4603	>	*/
4604	>	public static <K,V,U> ForkJoinTask<U> searchValues
4605	>	(ConcurrentHashMap<K,V> map,
4606	>	Fun<? super V, ? extends U> searchFunction) {
4607	>	if (searchFunction == null) throw new NullPointerException();
4608	>	return new SearchValuesTask<K,V,U>
4609	>	(map, null, -1, null, searchFunction,
4610	>	new AtomicReference<U>());
4611	>	}
4612	>
4613	>	/**
4614	>	* Returns a task that when invoked, returns the result of
4615	>	* accumulating all values using the given reducer to combine
4616	>	* values, or null if none.
4617	>	*
4618	>	* @param map the map
4619	>	* @param reducer a commutative associative combining function
4620	>	* @return the task
4621	>	*/
4622	>	public static <K,V> ForkJoinTask<V> reduceValues
4623	>	(ConcurrentHashMap<K,V> map,
4624	>	BiFun<? super V, ? super V, ? extends V> reducer) {
4625	>	if (reducer == null) throw new NullPointerException();
4626	>	return new ReduceValuesTask<K,V>
4627	>	(map, null, -1, null, reducer);
4628	>	}
4629	>
4630	>	/**
4631	>	* Returns a task that when invoked, returns the result of
4632	>	* accumulating the given transformation of all values using the
4633	>	* given reducer to combine values, or null if none.
4634	>	*
4635	>	* @param map the map
4636	>	* @param transformer a function returning the transformation
4637	>	* for an element, or null if there is no transformation (in
4638	>	* which case it is not combined).
4639	>	* @param reducer a commutative associative combining function
4640	>	* @return the task
4641	>	*/
4642	>	public static <K,V,U> ForkJoinTask<U> reduceValues
4643	>	(ConcurrentHashMap<K,V> map,
4644	>	Fun<? super V, ? extends U> transformer,
4645	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4646	>	if (transformer == null || reducer == null)
4647	>	throw new NullPointerException();
4648	>	return new MapReduceValuesTask<K,V,U>
4649	>	(map, null, -1, null, transformer, reducer);
4650	>	}
4651	>
4652	>	/**
4653	>	* Returns a task that when invoked, returns the result of
4654	>	* accumulating the given transformation of all values using the
4655	>	* given reducer to combine values, and the given basis as an
4656	>	* identity value.
4657	>	*
4658	>	* @param map the map
4659	>	* @param transformer a function returning the transformation
4660	>	* for an element
4661	>	* @param basis the identity (initial default value) for the reduction
4662	>	* @param reducer a commutative associative combining function
4663	>	* @return the task
4664	>	*/
4665	>	public static <K,V> ForkJoinTask<Double> reduceValuesToDouble
4666	>	(ConcurrentHashMap<K,V> map,
4667	>	ObjectToDouble<? super V> transformer,
4668	>	double basis,
4669	>	DoubleByDoubleToDouble reducer) {
4670	>	if (transformer == null || reducer == null)
4671	>	throw new NullPointerException();
4672	>	return new MapReduceValuesToDoubleTask<K,V>
4673	>	(map, null, -1, null, transformer, basis, reducer);
4674	>	}
4675	>
4676	>	/**
4677	>	* Returns a task that when invoked, returns the result of
4678	>	* accumulating the given transformation of all values using the
4679	>	* given reducer to combine values, and the given basis as an
4680	>	* identity value.
4681	>	*
4682	>	* @param map the map
4683	>	* @param transformer a function returning the transformation
4684	>	* for an element
4685	>	* @param basis the identity (initial default value) for the reduction
4686	>	* @param reducer a commutative associative combining function
4687	>	* @return the task
4688	>	*/
4689	>	public static <K,V> ForkJoinTask<Long> reduceValuesToLong
4690	>	(ConcurrentHashMap<K,V> map,
4691	>	ObjectToLong<? super V> transformer,
4692	>	long basis,
4693	>	LongByLongToLong reducer) {
4694	>	if (transformer == null || reducer == null)
4695	>	throw new NullPointerException();
4696	>	return new MapReduceValuesToLongTask<K,V>
4697	>	(map, null, -1, null, transformer, basis, reducer);
4698	>	}
4699	>
4700	>	/**
4701	>	* Returns a task that when invoked, returns the result of
4702	>	* accumulating the given transformation of all values using the
4703	>	* given reducer to combine values, and the given basis as an
4704	>	* identity value.
4705	>	*
4706	>	* @param map the map
4707	>	* @param transformer a function returning the transformation
4708	>	* for an element
4709	>	* @param basis the identity (initial default value) for the reduction
4710	>	* @param reducer a commutative associative combining function
4711	>	* @return the task
4712	>	*/
4713	>	public static <K,V> ForkJoinTask<Integer> reduceValuesToInt
4714	>	(ConcurrentHashMap<K,V> map,
4715	>	ObjectToInt<? super V> transformer,
4716	>	int basis,
4717	>	IntByIntToInt reducer) {
4718	>	if (transformer == null || reducer == null)
4719	>	throw new NullPointerException();
4720	>	return new MapReduceValuesToIntTask<K,V>
4721	>	(map, null, -1, null, transformer, basis, reducer);
4722	>	}
4723	>
4724	>	/**
4725	>	* Returns a task that when invoked, perform the given action
4726	>	* for each entry.
4727	>	*
4728	>	* @param map the map
4729	>	* @param action the action
4730	>	*/
4731	>	public static <K,V> ForkJoinTask<Void> forEachEntry
4732	>	(ConcurrentHashMap<K,V> map,
4733	>	Action<Map.Entry<K,V>> action) {
4734	>	if (action == null) throw new NullPointerException();
4735	>	return new ForEachEntryTask<K,V>(map, null, -1, null, action);
4736	>	}
4737	>
4738	>	/**
4739	>	* Returns a task that when invoked, perform the given action
4740	>	* for each non-null transformation of each entry.
4741	>	*
4742	>	* @param map the map
4743	>	* @param transformer a function returning the transformation
4744	>	* for an element, or null if there is no transformation (in
4745	>	* which case the action is not applied)
4746	>	* @param action the action
4747	>	*/
4748	>	public static <K,V,U> ForkJoinTask<Void> forEachEntry
4749	>	(ConcurrentHashMap<K,V> map,
4750	>	Fun<Map.Entry<K,V>, ? extends U> transformer,
4751	>	Action<U> action) {
4752	>	if (transformer == null || action == null)
4753	>	throw new NullPointerException();
4754	>	return new ForEachTransformedEntryTask<K,V,U>
4755	>	(map, null, -1, null, transformer, action);
4756	>	}
4757	>
4758	>	/**
4759	>	* Returns a task that when invoked, returns a non-null result
4760	>	* from applying the given search function on each entry, or
4761	>	* null if none.  Upon success, further element processing is
4762	>	* suppressed and the results of any other parallel
4763	>	* invocations of the search function are ignored.
4764	>	*
4765	>	* @param map the map
4766	>	* @param searchFunction a function returning a non-null
4767	>	* result on success, else null
4768	>	* @return the task
4769	>	*/
4770	>	public static <K,V,U> ForkJoinTask<U> searchEntries
4771	>	(ConcurrentHashMap<K,V> map,
4772	>	Fun<Map.Entry<K,V>, ? extends U> searchFunction) {
4773	>	if (searchFunction == null) throw new NullPointerException();
4774	>	return new SearchEntriesTask<K,V,U>
4775	>	(map, null, -1, null, searchFunction,
4776	>	new AtomicReference<U>());
4777		}
4778	<	public int size() {
4779	<	return ConcurrentHashMap.this.size();
4778	>
4779	>	/**
4780	>	* Returns a task that when invoked, returns the result of
4781	>	* accumulating all entries using the given reducer to combine
4782	>	* values, or null if none.
4783	>	*
4784	>	* @param map the map
4785	>	* @param reducer a commutative associative combining function
4786	>	* @return the task
4787	>	*/
4788	>	public static <K,V> ForkJoinTask<Map.Entry<K,V>> reduceEntries
4789	>	(ConcurrentHashMap<K,V> map,
4790	>	BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer) {
4791	>	if (reducer == null) throw new NullPointerException();
4792	>	return new ReduceEntriesTask<K,V>
4793	>	(map, null, -1, null, reducer);
4794		}
4795	<	public void clear() {
4796	<	ConcurrentHashMap.this.clear();
4795	>
4796	>	/**
4797	>	* Returns a task that when invoked, returns the result of
4798	>	* accumulating the given transformation of all entries using the
4799	>	* given reducer to combine values, or null if none.
4800	>	*
4801	>	* @param map the map
4802	>	* @param transformer a function returning the transformation
4803	>	* for an element, or null if there is no transformation (in
4804	>	* which case it is not combined).
4805	>	* @param reducer a commutative associative combining function
4806	>	* @return the task
4807	>	*/
4808	>	public static <K,V,U> ForkJoinTask<U> reduceEntries
4809	>	(ConcurrentHashMap<K,V> map,
4810	>	Fun<Map.Entry<K,V>, ? extends U> transformer,
4811	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4812	>	if (transformer == null || reducer == null)
4813	>	throw new NullPointerException();
4814	>	return new MapReduceEntriesTask<K,V,U>
4815	>	(map, null, -1, null, transformer, reducer);
4816		}
4817	<	public Object[] toArray() {
4818	<	Collection<Map.Entry<K,V>> c = new ArrayList<Map.Entry<K,V>>(size());
4819	<	for (Map.Entry<K,V> e : this)
4820	<	c.add(e);
4821	<	return c.toArray();
4822	<	}
4823	<	public <T> T[] toArray(T[] a) {
4824	<	Collection<Map.Entry<K,V>> c = new ArrayList<Map.Entry<K,V>>(size());
4825	<	for (Map.Entry<K,V> e : this)
4826	<	c.add(e);
4827	<	return c.toArray(a);
4817	>
4818	>	/**
4819	>	* Returns a task that when invoked, returns the result of
4820	>	* accumulating the given transformation of all entries using the
4821	>	* given reducer to combine values, and the given basis as an
4822	>	* identity value.
4823	>	*
4824	>	* @param map the map
4825	>	* @param transformer a function returning the transformation
4826	>	* for an element
4827	>	* @param basis the identity (initial default value) for the reduction
4828	>	* @param reducer a commutative associative combining function
4829	>	* @return the task
4830	>	*/
4831	>	public static <K,V> ForkJoinTask<Double> reduceEntriesToDouble
4832	>	(ConcurrentHashMap<K,V> map,
4833	>	ObjectToDouble<Map.Entry<K,V>> transformer,
4834	>	double basis,
4835	>	DoubleByDoubleToDouble reducer) {
4836	>	if (transformer == null || reducer == null)
4837	>	throw new NullPointerException();
4838	>	return new MapReduceEntriesToDoubleTask<K,V>
4839	>	(map, null, -1, null, transformer, basis, reducer);
4840	>	}
4841	>
4842	>	/**
4843	>	* Returns a task that when invoked, returns the result of
4844	>	* accumulating the given transformation of all entries using the
4845	>	* given reducer to combine values, and the given basis as an
4846	>	* identity value.
4847	>	*
4848	>	* @param map the map
4849	>	* @param transformer a function returning the transformation
4850	>	* for an element
4851	>	* @param basis the identity (initial default value) for the reduction
4852	>	* @param reducer a commutative associative combining function
4853	>	* @return the task
4854	>	*/
4855	>	public static <K,V> ForkJoinTask<Long> reduceEntriesToLong
4856	>	(ConcurrentHashMap<K,V> map,
4857	>	ObjectToLong<Map.Entry<K,V>> transformer,
4858	>	long basis,
4859	>	LongByLongToLong reducer) {
4860	>	if (transformer == null || reducer == null)
4861	>	throw new NullPointerException();
4862	>	return new MapReduceEntriesToLongTask<K,V>
4863	>	(map, null, -1, null, transformer, basis, reducer);
4864		}
4865
4866	+	/**
4867	+	* Returns a task that when invoked, returns the result of
4868	+	* accumulating the given transformation of all entries using the
4869	+	* given reducer to combine values, and the given basis as an
4870	+	* identity value.
4871	+	*
4872	+	* @param map the map
4873	+	* @param transformer a function returning the transformation
4874	+	* for an element
4875	+	* @param basis the identity (initial default value) for the reduction
4876	+	* @param reducer a commutative associative combining function
4877	+	* @return the task
4878	+	*/
4879	+	public static <K,V> ForkJoinTask<Integer> reduceEntriesToInt
4880	+	(ConcurrentHashMap<K,V> map,
4881	+	ObjectToInt<Map.Entry<K,V>> transformer,
4882	+	int basis,
4883	+	IntByIntToInt reducer) {
4884	+	if (transformer == null || reducer == null)
4885	+	throw new NullPointerException();
4886	+	return new MapReduceEntriesToIntTask<K,V>
4887	+	(map, null, -1, null, transformer, basis, reducer);
4888	+	}
4889		}
4890
4891	<	/* ---------------- Serialization Support -------------- */
4891	>	// -------------------------------------------------------
4892
4893		/**
4894	<	* Save the state of the <tt>ConcurrentHashMap</tt> instance to a
4895	<	* stream (i.e., serialize it).
4896	<	* @param s the stream
4897	<	* @serialData
4898	<	* the key (Object) and value (Object)
4899	<	* for each key-value mapping, followed by a null pair.
4900	<	* The key-value mappings are emitted in no particular order.
4901	<	*/
4902	<	private void writeObject(java.io.ObjectOutputStream s) throws IOException  {
4903	<	s.defaultWriteObject();
4894	>	* Base for FJ tasks for bulk operations. This adds a variant of
4895	>	* CountedCompleters and some split and merge bookkeeping to
4896	>	* iterator functionality. The forEach and reduce methods are
4897	>	* similar to those illustrated in CountedCompleter documentation,
4898	>	* except that bottom-up reduction completions perform them within
4899	>	* their compute methods. The search methods are like forEach
4900	>	* except they continually poll for success and exit early.  Also,
4901	>	* exceptions are handled in a simpler manner, by just trying to
4902	>	* complete root task exceptionally.
4903	>	*/
4904	>	@SuppressWarnings("serial") static abstract class BulkTask<K,V,R> extends Traverser<K,V,R> {
4905	>	final BulkTask<K,V,?> parent;  // completion target
4906	>	int batch;                     // split control; -1 for unknown
4907	>	int pending;                   // completion control
4908	>
4909	>	BulkTask(ConcurrentHashMap<K,V> map, BulkTask<K,V,?> parent,
4910	>	int batch) {
4911	>	super(map);
4912	>	this.parent = parent;
4913	>	this.batch = batch;
4914	>	if (parent != null && map != null) { // split parent
4915	>	Node[] t;
4916	>	if ((t = parent.tab) == null &&
4917	>	(t = parent.tab = map.table) != null)
4918	>	parent.baseLimit = parent.baseSize = t.length;
4919	>	this.tab = t;
4920	>	this.baseSize = parent.baseSize;
4921	>	int hi = this.baseLimit = parent.baseLimit;
4922	>	parent.baseLimit = this.index = this.baseIndex =
4923	>	(hi + parent.baseIndex + 1) >>> 1;
4924	>	}
4925	>	}
4926
4927	<	for (int k = 0; k < segments.length; ++k) {
4928	<	Segment<K,V> seg = segments[k];
4929	<	seg.lock();
4930	<	try {
4931	<	HashEntry<K,V>[] tab = seg.table;
4932	<	for (int i = 0; i < tab.length; ++i) {
4933	<	for (HashEntry<K,V> e = tab[i]; e != null; e = e.next) {
4934	<	s.writeObject(e.key);
4935	<	s.writeObject(e.value);
4936	<	}
4927	>	/**
4928	>	* Forces root task to complete.
4929	>	* @param ex if null, complete normally, else exceptionally
4930	>	* @return false to simplify use
4931	>	*/
4932	>	final boolean tryCompleteComputation(Throwable ex) {
4933	>	for (BulkTask<K,V,?> a = this;;) {
4934	>	BulkTask<K,V,?> p = a.parent;
4935	>	if (p == null) {
4936	>	if (ex != null)
4937	>	a.completeExceptionally(ex);
4938	>	else
4939	>	a.quietlyComplete();
4940	>	return false;
4941		}
4942	<	} finally {
4943	<	seg.unlock();
4942	>	a = p;
4943	>	}
4944	>	}
4945	>
4946	>	/**
4947	>	* Version of tryCompleteComputation for function screening checks
4948	>	*/
4949	>	final boolean abortOnNullFunction() {
4950	>	return tryCompleteComputation(new Error("Unexpected null function"));
4951	>	}
4952	>
4953	>	// utilities
4954	>
4955	>	/** CompareAndSet pending count */
4956	>	final boolean casPending(int cmp, int val) {
4957	>	return U.compareAndSwapInt(this, PENDING, cmp, val);
4958	>	}
4959	>
4960	>	/**
4961	>	* Returns approx exp2 of the number of times (minus one) to
4962	>	* split task by two before executing leaf action. This value
4963	>	* is faster to compute and more convenient to use as a guide
4964	>	* to splitting than is the depth, since it is used while
4965	>	* dividing by two anyway.
4966	>	*/
4967	>	final int batch() {
4968	>	ConcurrentHashMap<K, V> m; int b; Node[] t;  ForkJoinPool pool;
4969	>	if ((b = batch) < 0 && (m = map) != null) { // force initialization
4970	>	if ((t = tab) == null && (t = tab = m.table) != null)
4971	>	baseLimit = baseSize = t.length;
4972	>	if (t != null) {
4973	>	long n = m.counter.sum();
4974	>	int par = ((pool = getPool()) == null) ?
4975	>	ForkJoinPool.getCommonPoolParallelism() :
4976	>	pool.getParallelism();
4977	>	int sp = par << 3; // slack of 8
4978	>	b = batch = (n <= 0L) ? 0 : (n < (long)sp) ? (int)n : sp;
4979	>	}
4980	>	}
4981	>	return b;
4982	>	}
4983	>
4984	>	/**
4985	>	* Returns exportable snapshot entry.
4986	>	*/
4987	>	static <K,V> AbstractMap.SimpleEntry<K,V> entryFor(K k, V v) {
4988	>	return new AbstractMap.SimpleEntry<K,V>(k, v);
4989	>	}
4990	>
4991	>	// Unsafe mechanics
4992	>	private static final sun.misc.Unsafe U;
4993	>	private static final long PENDING;
4994	>	static {
4995	>	try {
4996	>	U = sun.misc.Unsafe.getUnsafe();
4997	>	PENDING = U.objectFieldOffset
4998	>	(BulkTask.class.getDeclaredField("pending"));
4999	>	} catch (Exception e) {
5000	>	throw new Error(e);
5001		}
5002		}
1287	–	s.writeObject(null);
1288	–	s.writeObject(null);
5003		}
5004
5005		/**
5006	<	* Reconstitute the <tt>ConcurrentHashMap</tt> instance from a
1293	<	* stream (i.e., deserialize it).
1294	<	* @param s the stream
5006	>	* Base class for non-reductive actions
5007		*/
5008	<	private void readObject(java.io.ObjectInputStream s)
5009	<	throws IOException, ClassNotFoundException  {
5010	<	s.defaultReadObject();
5008	>	@SuppressWarnings("serial") static abstract class BulkAction<K,V,R> extends BulkTask<K,V,R> {
5009	>	BulkAction<K,V,?> nextTask;
5010	>	BulkAction(ConcurrentHashMap<K,V> map, BulkTask<K,V,?> parent,
5011	>	int batch, BulkAction<K,V,?> nextTask) {
5012	>	super(map, parent, batch);
5013	>	this.nextTask = nextTask;
5014	>	}
5015	>
5016	>	/**
5017	>	* Try to complete task and upward parents. Upon hitting
5018	>	* non-completed parent, if a non-FJ task, try to help out the
5019	>	* computation.
5020	>	*/
5021	>	final void tryComplete(BulkAction<K,V,?> subtasks) {
5022	>	BulkTask<K,V,?> a = this, s = a;
5023	>	for (int c;;) {
5024	>	if ((c = a.pending) == 0) {
5025	>	if ((a = (s = a).parent) == null) {
5026	>	s.quietlyComplete();
5027	>	break;
5028	>	}
5029	>	}
5030	>	else if (a.casPending(c, c - 1)) {
5031	>	if (subtasks != null && !inForkJoinPool()) {
5032	>	while ((s = a.parent) != null)
5033	>	a = s;
5034	>	while (!a.isDone()) {
5035	>	BulkAction<K,V,?> next = subtasks.nextTask;
5036	>	if (subtasks.tryUnfork())
5037	>	subtasks.exec();
5038	>	if ((subtasks = next) == null)
5039	>	break;
5040	>	}
5041	>	}
5042	>	break;
5043	>	}
5044	>	}
5045	>	}
5046	>
5047	>	}
5048
5049	<	// Initialize each segment to be minimally sized, and let grow.
5050	<	for (int i = 0; i < segments.length; ++i) {
5051	<	segments[i].setTable(new HashEntry[1]);
5049	>	/*
5050	>	* Task classes. Coded in a regular but ugly format/style to
5051	>	* simplify checks that each variant differs in the right way from
5052	>	* others.
5053	>	*/
5054	>
5055	>	@SuppressWarnings("serial") static final class ForEachKeyTask<K,V>
5056	>	extends BulkAction<K,V,Void> {
5057	>	final Action<K> action;
5058	>	ForEachKeyTask
5059	>	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
5060	>	ForEachKeyTask<K,V> nextTask,
5061	>	Action<K> action) {
5062	>	super(m, p, b, nextTask);
5063	>	this.action = action;
5064	>	}
5065	>	@SuppressWarnings("unchecked") public final boolean exec() {
5066	>	final Action<K> action = this.action;
5067	>	if (action == null)
5068	>	return abortOnNullFunction();
5069	>	ForEachKeyTask<K,V> subtasks = null;
5070	>	try {
5071	>	int b = batch(), c;
5072	>	while (b > 1 && baseIndex != baseLimit) {
5073	>	do {} while (!casPending(c = pending, c+1));
5074	>	(subtasks = new ForEachKeyTask<K,V>
5075	>	(map, this, b >>>= 1, subtasks, action)).fork();
5076	>	}
5077	>	while (advance() != null)
5078	>	action.apply((K)nextKey);
5079	>	} catch (Throwable ex) {
5080	>	return tryCompleteComputation(ex);
5081	>	}
5082	>	tryComplete(subtasks);
5083	>	return false;
5084		}
5085	+	}
5086
5087	<	// Read the keys and values, and put the mappings in the table
5088	<	for (;;) {
5089	<	K key = (K) s.readObject();
5090	<	V value = (V) s.readObject();
5091	<	if (key == null)
5092	<	break;
5093	<	put(key, value);
5087	>	@SuppressWarnings("serial") static final class ForEachValueTask<K,V>
5088	>	extends BulkAction<K,V,Void> {
5089	>	final Action<V> action;
5090	>	ForEachValueTask
5091	>	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
5092	>	ForEachValueTask<K,V> nextTask,
5093	>	Action<V> action) {
5094	>	super(m, p, b, nextTask);
5095	>	this.action = action;
5096	>	}
5097	>	@SuppressWarnings("unchecked") public final boolean exec() {
5098	>	final Action<V> action = this.action;
5099	>	if (action == null)
5100	>	return abortOnNullFunction();
5101	>	ForEachValueTask<K,V> subtasks = null;
5102	>	try {
5103	>	int b = batch(), c;
5104	>	while (b > 1 && baseIndex != baseLimit) {
5105	>	do {} while (!casPending(c = pending, c+1));
5106	>	(subtasks = new ForEachValueTask<K,V>
5107	>	(map, this, b >>>= 1, subtasks, action)).fork();
5108	>	}
5109	>	Object v;
5110	>	while ((v = advance()) != null)
5111	>	action.apply((V)v);
5112	>	} catch (Throwable ex) {
5113	>	return tryCompleteComputation(ex);
5114	>	}
5115	>	tryComplete(subtasks);
5116	>	return false;
5117	>	}
5118	>	}
5119	>
5120	>	@SuppressWarnings("serial") static final class ForEachEntryTask<K,V>
5121	>	extends BulkAction<K,V,Void> {
5122	>	final Action<Entry<K,V>> action;
5123	>	ForEachEntryTask
5124	>	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
5125	>	ForEachEntryTask<K,V> nextTask,
5126	>	Action<Entry<K,V>> action) {
5127	>	super(m, p, b, nextTask);
5128	>	this.action = action;
5129	>	}
5130	>	@SuppressWarnings("unchecked") public final boolean exec() {
5131	>	final Action<Entry<K,V>> action = this.action;
5132	>	if (action == null)
5133	>	return abortOnNullFunction();
5134	>	ForEachEntryTask<K,V> subtasks = null;
5135	>	try {
5136	>	int b = batch(), c;
5137	>	while (b > 1 && baseIndex != baseLimit) {
5138	>	do {} while (!casPending(c = pending, c+1));
5139	>	(subtasks = new ForEachEntryTask<K,V>
5140	>	(map, this, b >>>= 1, subtasks, action)).fork();
5141	>	}
5142	>	Object v;
5143	>	while ((v = advance()) != null)
5144	>	action.apply(entryFor((K)nextKey, (V)v));
5145	>	} catch (Throwable ex) {
5146	>	return tryCompleteComputation(ex);
5147	>	}
5148	>	tryComplete(subtasks);
5149	>	return false;
5150	>	}
5151	>	}
5152	>
5153	>	@SuppressWarnings("serial") static final class ForEachMappingTask<K,V>
5154	>	extends BulkAction<K,V,Void> {
5155	>	final BiAction<K,V> action;
5156	>	ForEachMappingTask
5157	>	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
5158	>	ForEachMappingTask<K,V> nextTask,
5159	>	BiAction<K,V> action) {
5160	>	super(m, p, b, nextTask);
5161	>	this.action = action;
5162	>	}
5163	>	@SuppressWarnings("unchecked") public final boolean exec() {
5164	>	final BiAction<K,V> action = this.action;
5165	>	if (action == null)
5166	>	return abortOnNullFunction();
5167	>	ForEachMappingTask<K,V> subtasks = null;
5168	>	try {
5169	>	int b = batch(), c;
5170	>	while (b > 1 && baseIndex != baseLimit) {
5171	>	do {} while (!casPending(c = pending, c+1));
5172	>	(subtasks = new ForEachMappingTask<K,V>
5173	>	(map, this, b >>>= 1, subtasks, action)).fork();
5174	>	}
5175	>	Object v;
5176	>	while ((v = advance()) != null)
5177	>	action.apply((K)nextKey, (V)v);
5178	>	} catch (Throwable ex) {
5179	>	return tryCompleteComputation(ex);
5180	>	}
5181	>	tryComplete(subtasks);
5182	>	return false;
5183	>	}
5184	>	}
5185	>
5186	>	@SuppressWarnings("serial") static final class ForEachTransformedKeyTask<K,V,U>
5187	>	extends BulkAction<K,V,Void> {
5188	>	final Fun<? super K, ? extends U> transformer;
5189	>	final Action<U> action;
5190	>	ForEachTransformedKeyTask
5191	>	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
5192	>	ForEachTransformedKeyTask<K,V,U> nextTask,
5193	>	Fun<? super K, ? extends U> transformer,
5194	>	Action<U> action) {
5195	>	super(m, p, b, nextTask);
5196	>	this.transformer = transformer;
5197	>	this.action = action;
5198	>
5199	>	}
5200	>	@SuppressWarnings("unchecked") public final boolean exec() {
5201	>	final Fun<? super K, ? extends U> transformer =
5202	>	this.transformer;
5203	>	final Action<U> action = this.action;
5204	>	if (transformer == null || action == null)
5205	>	return abortOnNullFunction();
5206	>	ForEachTransformedKeyTask<K,V,U> subtasks = null;
5207	>	try {
5208	>	int b = batch(), c;
5209	>	while (b > 1 && baseIndex != baseLimit) {
5210	>	do {} while (!casPending(c = pending, c+1));
5211	>	(subtasks = new ForEachTransformedKeyTask<K,V,U>
5212	>	(map, this, b >>>= 1, subtasks, transformer, action)).fork();
5213	>	}
5214	>	U u;
5215	>	while (advance() != null) {
5216	>	if ((u = transformer.apply((K)nextKey)) != null)
5217	>	action.apply(u);
5218	>	}
5219	>	} catch (Throwable ex) {
5220	>	return tryCompleteComputation(ex);
5221	>	}
5222	>	tryComplete(subtasks);
5223	>	return false;
5224	>	}
5225	>	}
5226	>
5227	>	@SuppressWarnings("serial") static final class ForEachTransformedValueTask<K,V,U>
5228	>	extends BulkAction<K,V,Void> {
5229	>	final Fun<? super V, ? extends U> transformer;
5230	>	final Action<U> action;
5231	>	ForEachTransformedValueTask
5232	>	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
5233	>	ForEachTransformedValueTask<K,V,U> nextTask,
5234	>	Fun<? super V, ? extends U> transformer,
5235	>	Action<U> action) {
5236	>	super(m, p, b, nextTask);
5237	>	this.transformer = transformer;
5238	>	this.action = action;
5239	>
5240	>	}
5241	>	@SuppressWarnings("unchecked") public final boolean exec() {
5242	>	final Fun<? super V, ? extends U> transformer =
5243	>	this.transformer;
5244	>	final Action<U> action = this.action;
5245	>	if (transformer == null || action == null)
5246	>	return abortOnNullFunction();
5247	>	ForEachTransformedValueTask<K,V,U> subtasks = null;
5248	>	try {
5249	>	int b = batch(), c;
5250	>	while (b > 1 && baseIndex != baseLimit) {
5251	>	do {} while (!casPending(c = pending, c+1));
5252	>	(subtasks = new ForEachTransformedValueTask<K,V,U>
5253	>	(map, this, b >>>= 1, subtasks, transformer, action)).fork();
5254	>	}
5255	>	Object v; U u;
5256	>	while ((v = advance()) != null) {
5257	>	if ((u = transformer.apply((V)v)) != null)
5258	>	action.apply(u);
5259	>	}
5260	>	} catch (Throwable ex) {
5261	>	return tryCompleteComputation(ex);
5262	>	}
5263	>	tryComplete(subtasks);
5264	>	return false;
5265	>	}
5266	>	}
5267	>
5268	>	@SuppressWarnings("serial") static final class ForEachTransformedEntryTask<K,V,U>
5269	>	extends BulkAction<K,V,Void> {
5270	>	final Fun<Map.Entry<K,V>, ? extends U> transformer;
5271	>	final Action<U> action;
5272	>	ForEachTransformedEntryTask
5273	>	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
5274	>	ForEachTransformedEntryTask<K,V,U> nextTask,
5275	>	Fun<Map.Entry<K,V>, ? extends U> transformer,
5276	>	Action<U> action) {
5277	>	super(m, p, b, nextTask);
5278	>	this.transformer = transformer;
5279	>	this.action = action;
5280	>
5281	>	}
5282	>	@SuppressWarnings("unchecked") public final boolean exec() {
5283	>	final Fun<Map.Entry<K,V>, ? extends U> transformer =
5284	>	this.transformer;
5285	>	final Action<U> action = this.action;
5286	>	if (transformer == null || action == null)
5287	>	return abortOnNullFunction();
5288	>	ForEachTransformedEntryTask<K,V,U> subtasks = null;
5289	>	try {
5290	>	int b = batch(), c;
5291	>	while (b > 1 && baseIndex != baseLimit) {
5292	>	do {} while (!casPending(c = pending, c+1));
5293	>	(subtasks = new ForEachTransformedEntryTask<K,V,U>
5294	>	(map, this, b >>>= 1, subtasks, transformer, action)).fork();
5295	>	}
5296	>	Object v; U u;
5297	>	while ((v = advance()) != null) {
5298	>	if ((u = transformer.apply(entryFor((K)nextKey, (V)v))) != null)
5299	>	action.apply(u);
5300	>	}
5301	>	} catch (Throwable ex) {
5302	>	return tryCompleteComputation(ex);
5303	>	}
5304	>	tryComplete(subtasks);
5305	>	return false;
5306	>	}
5307	>	}
5308	>
5309	>	@SuppressWarnings("serial") static final class ForEachTransformedMappingTask<K,V,U>
5310	>	extends BulkAction<K,V,Void> {
5311	>	final BiFun<? super K, ? super V, ? extends U> transformer;
5312	>	final Action<U> action;
5313	>	ForEachTransformedMappingTask
5314	>	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
5315	>	ForEachTransformedMappingTask<K,V,U> nextTask,
5316	>	BiFun<? super K, ? super V, ? extends U> transformer,
5317	>	Action<U> action) {
5318	>	super(m, p, b, nextTask);
5319	>	this.transformer = transformer;
5320	>	this.action = action;
5321	>
5322	>	}
5323	>	@SuppressWarnings("unchecked") public final boolean exec() {
5324	>	final BiFun<? super K, ? super V, ? extends U> transformer =
5325	>	this.transformer;
5326	>	final Action<U> action = this.action;
5327	>	if (transformer == null || action == null)
5328	>	return abortOnNullFunction();
5329	>	ForEachTransformedMappingTask<K,V,U> subtasks = null;
5330	>	try {
5331	>	int b = batch(), c;
5332	>	while (b > 1 && baseIndex != baseLimit) {
5333	>	do {} while (!casPending(c = pending, c+1));
5334	>	(subtasks = new ForEachTransformedMappingTask<K,V,U>
5335	>	(map, this, b >>>= 1, subtasks, transformer, action)).fork();
5336	>	}
5337	>	Object v; U u;
5338	>	while ((v = advance()) != null) {
5339	>	if ((u = transformer.apply((K)nextKey, (V)v)) != null)
5340	>	action.apply(u);
5341	>	}
5342	>	} catch (Throwable ex) {
5343	>	return tryCompleteComputation(ex);
5344	>	}
5345	>	tryComplete(subtasks);
5346	>	return false;
5347	>	}
5348	>	}
5349	>
5350	>	@SuppressWarnings("serial") static final class SearchKeysTask<K,V,U>
5351	>	extends BulkAction<K,V,U> {
5352	>	final Fun<? super K, ? extends U> searchFunction;
5353	>	final AtomicReference<U> result;
5354	>	SearchKeysTask
5355	>	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
5356	>	SearchKeysTask<K,V,U> nextTask,
5357	>	Fun<? super K, ? extends U> searchFunction,
5358	>	AtomicReference<U> result) {
5359	>	super(m, p, b, nextTask);
5360	>	this.searchFunction = searchFunction; this.result = result;
5361	>	}
5362	>	@SuppressWarnings("unchecked") public final boolean exec() {
5363	>	AtomicReference<U> result = this.result;
5364	>	final Fun<? super K, ? extends U> searchFunction =
5365	>	this.searchFunction;
5366	>	if (searchFunction == null || result == null)
5367	>	return abortOnNullFunction();
5368	>	SearchKeysTask<K,V,U> subtasks = null;
5369	>	try {
5370	>	int b = batch(), c;
5371	>	while (b > 1 && baseIndex != baseLimit && result.get() == null) {
5372	>	do {} while (!casPending(c = pending, c+1));
5373	>	(subtasks = new SearchKeysTask<K,V,U>
5374	>	(map, this, b >>>= 1, subtasks, searchFunction, result)).fork();
5375	>	}
5376	>	U u;
5377	>	while (result.get() == null && advance() != null) {
5378	>	if ((u = searchFunction.apply((K)nextKey)) != null) {
5379	>	if (result.compareAndSet(null, u))
5380	>	tryCompleteComputation(null);
5381	>	break;
5382	>	}
5383	>	}
5384	>	} catch (Throwable ex) {
5385	>	return tryCompleteComputation(ex);
5386	>	}
5387	>	tryComplete(subtasks);
5388	>	return false;
5389		}
5390	+	public final U getRawResult() { return result.get(); }
5391	+	}
5392	+
5393	+	@SuppressWarnings("serial") static final class SearchValuesTask<K,V,U>
5394	+	extends BulkAction<K,V,U> {
5395	+	final Fun<? super V, ? extends U> searchFunction;
5396	+	final AtomicReference<U> result;
5397	+	SearchValuesTask
5398	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
5399	+	SearchValuesTask<K,V,U> nextTask,
5400	+	Fun<? super V, ? extends U> searchFunction,
5401	+	AtomicReference<U> result) {
5402	+	super(m, p, b, nextTask);
5403	+	this.searchFunction = searchFunction; this.result = result;
5404	+	}
5405	+	@SuppressWarnings("unchecked") public final boolean exec() {
5406	+	AtomicReference<U> result = this.result;
5407	+	final Fun<? super V, ? extends U> searchFunction =
5408	+	this.searchFunction;
5409	+	if (searchFunction == null || result == null)
5410	+	return abortOnNullFunction();
5411	+	SearchValuesTask<K,V,U> subtasks = null;
5412	+	try {
5413	+	int b = batch(), c;
5414	+	while (b > 1 && baseIndex != baseLimit && result.get() == null) {
5415	+	do {} while (!casPending(c = pending, c+1));
5416	+	(subtasks = new SearchValuesTask<K,V,U>
5417	+	(map, this, b >>>= 1, subtasks, searchFunction, result)).fork();
5418	+	}
5419	+	Object v; U u;
5420	+	while (result.get() == null && (v = advance()) != null) {
5421	+	if ((u = searchFunction.apply((V)v)) != null) {
5422	+	if (result.compareAndSet(null, u))
5423	+	tryCompleteComputation(null);
5424	+	break;
5425	+	}
5426	+	}
5427	+	} catch (Throwable ex) {
5428	+	return tryCompleteComputation(ex);
5429	+	}
5430	+	tryComplete(subtasks);
5431	+	return false;
5432	+	}
5433	+	public final U getRawResult() { return result.get(); }
5434	+	}
5435	+
5436	+	@SuppressWarnings("serial") static final class SearchEntriesTask<K,V,U>
5437	+	extends BulkAction<K,V,U> {
5438	+	final Fun<Entry<K,V>, ? extends U> searchFunction;
5439	+	final AtomicReference<U> result;
5440	+	SearchEntriesTask
5441	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
5442	+	SearchEntriesTask<K,V,U> nextTask,
5443	+	Fun<Entry<K,V>, ? extends U> searchFunction,
5444	+	AtomicReference<U> result) {
5445	+	super(m, p, b, nextTask);
5446	+	this.searchFunction = searchFunction; this.result = result;
5447	+	}
5448	+	@SuppressWarnings("unchecked") public final boolean exec() {
5449	+	AtomicReference<U> result = this.result;
5450	+	final Fun<Entry<K,V>, ? extends U> searchFunction =
5451	+	this.searchFunction;
5452	+	if (searchFunction == null || result == null)
5453	+	return abortOnNullFunction();
5454	+	SearchEntriesTask<K,V,U> subtasks = null;
5455	+	try {
5456	+	int b = batch(), c;
5457	+	while (b > 1 && baseIndex != baseLimit && result.get() == null) {
5458	+	do {} while (!casPending(c = pending, c+1));
5459	+	(subtasks = new SearchEntriesTask<K,V,U>
5460	+	(map, this, b >>>= 1, subtasks, searchFunction, result)).fork();
5461	+	}
5462	+	Object v; U u;
5463	+	while (result.get() == null && (v = advance()) != null) {
5464	+	if ((u = searchFunction.apply(entryFor((K)nextKey, (V)v))) != null) {
5465	+	if (result.compareAndSet(null, u))
5466	+	tryCompleteComputation(null);
5467	+	break;
5468	+	}
5469	+	}
5470	+	} catch (Throwable ex) {
5471	+	return tryCompleteComputation(ex);
5472	+	}
5473	+	tryComplete(subtasks);
5474	+	return false;
5475	+	}
5476	+	public final U getRawResult() { return result.get(); }
5477	+	}
5478	+
5479	+	@SuppressWarnings("serial") static final class SearchMappingsTask<K,V,U>
5480	+	extends BulkAction<K,V,U> {
5481	+	final BiFun<? super K, ? super V, ? extends U> searchFunction;
5482	+	final AtomicReference<U> result;
5483	+	SearchMappingsTask
5484	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
5485	+	SearchMappingsTask<K,V,U> nextTask,
5486	+	BiFun<? super K, ? super V, ? extends U> searchFunction,
5487	+	AtomicReference<U> result) {
5488	+	super(m, p, b, nextTask);
5489	+	this.searchFunction = searchFunction; this.result = result;
5490	+	}
5491	+	@SuppressWarnings("unchecked") public final boolean exec() {
5492	+	AtomicReference<U> result = this.result;
5493	+	final BiFun<? super K, ? super V, ? extends U> searchFunction =
5494	+	this.searchFunction;
5495	+	if (searchFunction == null || result == null)
5496	+	return abortOnNullFunction();
5497	+	SearchMappingsTask<K,V,U> subtasks = null;
5498	+	try {
5499	+	int b = batch(), c;
5500	+	while (b > 1 && baseIndex != baseLimit && result.get() == null) {
5501	+	do {} while (!casPending(c = pending, c+1));
5502	+	(subtasks = new SearchMappingsTask<K,V,U>
5503	+	(map, this, b >>>= 1, subtasks, searchFunction, result)).fork();
5504	+	}
5505	+	Object v; U u;
5506	+	while (result.get() == null && (v = advance()) != null) {
5507	+	if ((u = searchFunction.apply((K)nextKey, (V)v)) != null) {
5508	+	if (result.compareAndSet(null, u))
5509	+	tryCompleteComputation(null);
5510	+	break;
5511	+	}
5512	+	}
5513	+	} catch (Throwable ex) {
5514	+	return tryCompleteComputation(ex);
5515	+	}
5516	+	tryComplete(subtasks);
5517	+	return false;
5518	+	}
5519	+	public final U getRawResult() { return result.get(); }
5520	+	}
5521	+
5522	+	@SuppressWarnings("serial") static final class ReduceKeysTask<K,V>
5523	+	extends BulkTask<K,V,K> {
5524	+	final BiFun<? super K, ? super K, ? extends K> reducer;
5525	+	K result;
5526	+	ReduceKeysTask<K,V> rights, nextRight;
5527	+	ReduceKeysTask
5528	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
5529	+	ReduceKeysTask<K,V> nextRight,
5530	+	BiFun<? super K, ? super K, ? extends K> reducer) {
5531	+	super(m, p, b); this.nextRight = nextRight;
5532	+	this.reducer = reducer;
5533	+	}
5534	+	@SuppressWarnings("unchecked") public final boolean exec() {
5535	+	final BiFun<? super K, ? super K, ? extends K> reducer =
5536	+	this.reducer;
5537	+	if (reducer == null)
5538	+	return abortOnNullFunction();
5539	+	try {
5540	+	for (int c, b = batch(); b > 1 && baseIndex != baseLimit;) {
5541	+	do {} while (!casPending(c = pending, c+1));
5542	+	(rights = new ReduceKeysTask<K,V>
5543	+	(map, this, b >>>= 1, rights, reducer)).fork();
5544	+	}
5545	+	K r = null;
5546	+	while (advance() != null) {
5547	+	K u = (K)nextKey;
5548	+	r = (r == null) ? u : reducer.apply(r, u);
5549	+	}
5550	+	result = r;
5551	+	for (ReduceKeysTask<K,V> t = this, s;;) {
5552	+	int c; BulkTask<K,V,?> par; K tr, sr;
5553	+	if ((c = t.pending) == 0) {
5554	+	for (s = t.rights; s != null; s = t.rights = s.nextRight) {
5555	+	if ((sr = s.result) != null)
5556	+	t.result = ((tr = t.result) == null) ? sr : reducer.apply(tr, sr);
5557	+	}
5558	+	if ((par = t.parent) == null ||
5559	+	!(par instanceof ReduceKeysTask)) {
5560	+	t.quietlyComplete();
5561	+	break;
5562	+	}
5563	+	t = (ReduceKeysTask<K,V>)par;
5564	+	}
5565	+	else if (t.casPending(c, c - 1))
5566	+	break;
5567	+	}
5568	+	} catch (Throwable ex) {
5569	+	return tryCompleteComputation(ex);
5570	+	}
5571	+	ReduceKeysTask<K,V> s = rights;
5572	+	if (s != null && !inForkJoinPool()) {
5573	+	do  {
5574	+	if (s.tryUnfork())
5575	+	s.exec();
5576	+	} while ((s = s.nextRight) != null);
5577	+	}
5578	+	return false;
5579	+	}
5580	+	public final K getRawResult() { return result; }
5581	+	}
5582	+
5583	+	@SuppressWarnings("serial") static final class ReduceValuesTask<K,V>
5584	+	extends BulkTask<K,V,V> {
5585	+	final BiFun<? super V, ? super V, ? extends V> reducer;
5586	+	V result;
5587	+	ReduceValuesTask<K,V> rights, nextRight;
5588	+	ReduceValuesTask
5589	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
5590	+	ReduceValuesTask<K,V> nextRight,
5591	+	BiFun<? super V, ? super V, ? extends V> reducer) {
5592	+	super(m, p, b); this.nextRight = nextRight;
5593	+	this.reducer = reducer;
5594	+	}
5595	+	@SuppressWarnings("unchecked") public final boolean exec() {
5596	+	final BiFun<? super V, ? super V, ? extends V> reducer =
5597	+	this.reducer;
5598	+	if (reducer == null)
5599	+	return abortOnNullFunction();
5600	+	try {
5601	+	for (int c, b = batch(); b > 1 && baseIndex != baseLimit;) {
5602	+	do {} while (!casPending(c = pending, c+1));
5603	+	(rights = new ReduceValuesTask<K,V>
5604	+	(map, this, b >>>= 1, rights, reducer)).fork();
5605	+	}
5606	+	V r = null;
5607	+	Object v;
5608	+	while ((v = advance()) != null) {
5609	+	V u = (V)v;
5610	+	r = (r == null) ? u : reducer.apply(r, u);
5611	+	}
5612	+	result = r;
5613	+	for (ReduceValuesTask<K,V> t = this, s;;) {
5614	+	int c; BulkTask<K,V,?> par; V tr, sr;
5615	+	if ((c = t.pending) == 0) {
5616	+	for (s = t.rights; s != null; s = t.rights = s.nextRight) {
5617	+	if ((sr = s.result) != null)
5618	+	t.result = ((tr = t.result) == null) ? sr : reducer.apply(tr, sr);
5619	+	}
5620	+	if ((par = t.parent) == null ||
5621	+	!(par instanceof ReduceValuesTask)) {
5622	+	t.quietlyComplete();
5623	+	break;
5624	+	}
5625	+	t = (ReduceValuesTask<K,V>)par;
5626	+	}
5627	+	else if (t.casPending(c, c - 1))
5628	+	break;
5629	+	}
5630	+	} catch (Throwable ex) {
5631	+	return tryCompleteComputation(ex);
5632	+	}
5633	+	ReduceValuesTask<K,V> s = rights;
5634	+	if (s != null && !inForkJoinPool()) {
5635	+	do  {
5636	+	if (s.tryUnfork())
5637	+	s.exec();
5638	+	} while ((s = s.nextRight) != null);
5639	+	}
5640	+	return false;
5641	+	}
5642	+	public final V getRawResult() { return result; }
5643	+	}
5644	+
5645	+	@SuppressWarnings("serial") static final class ReduceEntriesTask<K,V>
5646	+	extends BulkTask<K,V,Map.Entry<K,V>> {
5647	+	final BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer;
5648	+	Map.Entry<K,V> result;
5649	+	ReduceEntriesTask<K,V> rights, nextRight;
5650	+	ReduceEntriesTask
5651	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
5652	+	ReduceEntriesTask<K,V> nextRight,
5653	+	BiFun<Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer) {
5654	+	super(m, p, b); this.nextRight = nextRight;
5655	+	this.reducer = reducer;
5656	+	}
5657	+	@SuppressWarnings("unchecked") public final boolean exec() {
5658	+	final BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer =
5659	+	this.reducer;
5660	+	if (reducer == null)
5661	+	return abortOnNullFunction();
5662	+	try {
5663	+	for (int c, b = batch(); b > 1 && baseIndex != baseLimit;) {
5664	+	do {} while (!casPending(c = pending, c+1));
5665	+	(rights = new ReduceEntriesTask<K,V>
5666	+	(map, this, b >>>= 1, rights, reducer)).fork();
5667	+	}
5668	+	Map.Entry<K,V> r = null;
5669	+	Object v;
5670	+	while ((v = advance()) != null) {
5671	+	Map.Entry<K,V> u = entryFor((K)nextKey, (V)v);
5672	+	r = (r == null) ? u : reducer.apply(r, u);
5673	+	}
5674	+	result = r;
5675	+	for (ReduceEntriesTask<K,V> t = this, s;;) {
5676	+	int c; BulkTask<K,V,?> par; Map.Entry<K,V> tr, sr;
5677	+	if ((c = t.pending) == 0) {
5678	+	for (s = t.rights; s != null; s = t.rights = s.nextRight) {
5679	+	if ((sr = s.result) != null)
5680	+	t.result = ((tr = t.result) == null) ? sr : reducer.apply(tr, sr);
5681	+	}
5682	+	if ((par = t.parent) == null ||
5683	+	!(par instanceof ReduceEntriesTask)) {
5684	+	t.quietlyComplete();
5685	+	break;
5686	+	}
5687	+	t = (ReduceEntriesTask<K,V>)par;
5688	+	}
5689	+	else if (t.casPending(c, c - 1))
5690	+	break;
5691	+	}
5692	+	} catch (Throwable ex) {
5693	+	return tryCompleteComputation(ex);
5694	+	}
5695	+	ReduceEntriesTask<K,V> s = rights;
5696	+	if (s != null && !inForkJoinPool()) {
5697	+	do  {
5698	+	if (s.tryUnfork())
5699	+	s.exec();
5700	+	} while ((s = s.nextRight) != null);
5701	+	}
5702	+	return false;
5703	+	}
5704	+	public final Map.Entry<K,V> getRawResult() { return result; }
5705	+	}
5706	+
5707	+	@SuppressWarnings("serial") static final class MapReduceKeysTask<K,V,U>
5708	+	extends BulkTask<K,V,U> {
5709	+	final Fun<? super K, ? extends U> transformer;
5710	+	final BiFun<? super U, ? super U, ? extends U> reducer;
5711	+	U result;
5712	+	MapReduceKeysTask<K,V,U> rights, nextRight;
5713	+	MapReduceKeysTask
5714	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
5715	+	MapReduceKeysTask<K,V,U> nextRight,
5716	+	Fun<? super K, ? extends U> transformer,
5717	+	BiFun<? super U, ? super U, ? extends U> reducer) {
5718	+	super(m, p, b); this.nextRight = nextRight;
5719	+	this.transformer = transformer;
5720	+	this.reducer = reducer;
5721	+	}
5722	+	@SuppressWarnings("unchecked") public final boolean exec() {
5723	+	final Fun<? super K, ? extends U> transformer =
5724	+	this.transformer;
5725	+	final BiFun<? super U, ? super U, ? extends U> reducer =
5726	+	this.reducer;
5727	+	if (transformer == null || reducer == null)
5728	+	return abortOnNullFunction();
5729	+	try {
5730	+	for (int c, b = batch(); b > 1 && baseIndex != baseLimit;) {
5731	+	do {} while (!casPending(c = pending, c+1));
5732	+	(rights = new MapReduceKeysTask<K,V,U>
5733	+	(map, this, b >>>= 1, rights, transformer, reducer)).fork();
5734	+	}
5735	+	U r = null, u;
5736	+	while (advance() != null) {
5737	+	if ((u = transformer.apply((K)nextKey)) != null)
5738	+	r = (r == null) ? u : reducer.apply(r, u);
5739	+	}
5740	+	result = r;
5741	+	for (MapReduceKeysTask<K,V,U> t = this, s;;) {
5742	+	int c; BulkTask<K,V,?> par; U tr, sr;
5743	+	if ((c = t.pending) == 0) {
5744	+	for (s = t.rights; s != null; s = t.rights = s.nextRight) {
5745	+	if ((sr = s.result) != null)
5746	+	t.result = ((tr = t.result) == null) ? sr : reducer.apply(tr, sr);
5747	+	}
5748	+	if ((par = t.parent) == null ||
5749	+	!(par instanceof MapReduceKeysTask)) {
5750	+	t.quietlyComplete();
5751	+	break;
5752	+	}
5753	+	t = (MapReduceKeysTask<K,V,U>)par;
5754	+	}
5755	+	else if (t.casPending(c, c - 1))
5756	+	break;
5757	+	}
5758	+	} catch (Throwable ex) {
5759	+	return tryCompleteComputation(ex);
5760	+	}
5761	+	MapReduceKeysTask<K,V,U> s = rights;
5762	+	if (s != null && !inForkJoinPool()) {
5763	+	do  {
5764	+	if (s.tryUnfork())
5765	+	s.exec();
5766	+	} while ((s = s.nextRight) != null);
5767	+	}
5768	+	return false;
5769	+	}
5770	+	public final U getRawResult() { return result; }
5771	+	}
5772	+
5773	+	@SuppressWarnings("serial") static final class MapReduceValuesTask<K,V,U>
5774	+	extends BulkTask<K,V,U> {
5775	+	final Fun<? super V, ? extends U> transformer;
5776	+	final BiFun<? super U, ? super U, ? extends U> reducer;
5777	+	U result;
5778	+	MapReduceValuesTask<K,V,U> rights, nextRight;
5779	+	MapReduceValuesTask
5780	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
5781	+	MapReduceValuesTask<K,V,U> nextRight,
5782	+	Fun<? super V, ? extends U> transformer,
5783	+	BiFun<? super U, ? super U, ? extends U> reducer) {
5784	+	super(m, p, b); this.nextRight = nextRight;
5785	+	this.transformer = transformer;
5786	+	this.reducer = reducer;
5787	+	}
5788	+	@SuppressWarnings("unchecked") public final boolean exec() {
5789	+	final Fun<? super V, ? extends U> transformer =
5790	+	this.transformer;
5791	+	final BiFun<? super U, ? super U, ? extends U> reducer =
5792	+	this.reducer;
5793	+	if (transformer == null || reducer == null)
5794	+	return abortOnNullFunction();
5795	+	try {
5796	+	for (int c, b = batch(); b > 1 && baseIndex != baseLimit;) {
5797	+	do {} while (!casPending(c = pending, c+1));
5798	+	(rights = new MapReduceValuesTask<K,V,U>
5799	+	(map, this, b >>>= 1, rights, transformer, reducer)).fork();
5800	+	}
5801	+	U r = null, u;
5802	+	Object v;
5803	+	while ((v = advance()) != null) {
5804	+	if ((u = transformer.apply((V)v)) != null)
5805	+	r = (r == null) ? u : reducer.apply(r, u);
5806	+	}
5807	+	result = r;
5808	+	for (MapReduceValuesTask<K,V,U> t = this, s;;) {
5809	+	int c; BulkTask<K,V,?> par; U tr, sr;
5810	+	if ((c = t.pending) == 0) {
5811	+	for (s = t.rights; s != null; s = t.rights = s.nextRight) {
5812	+	if ((sr = s.result) != null)
5813	+	t.result = ((tr = t.result) == null) ? sr : reducer.apply(tr, sr);
5814	+	}
5815	+	if ((par = t.parent) == null ||
5816	+	!(par instanceof MapReduceValuesTask)) {
5817	+	t.quietlyComplete();
5818	+	break;
5819	+	}
5820	+	t = (MapReduceValuesTask<K,V,U>)par;
5821	+	}
5822	+	else if (t.casPending(c, c - 1))
5823	+	break;
5824	+	}
5825	+	} catch (Throwable ex) {
5826	+	return tryCompleteComputation(ex);
5827	+	}
5828	+	MapReduceValuesTask<K,V,U> s = rights;
5829	+	if (s != null && !inForkJoinPool()) {
5830	+	do  {
5831	+	if (s.tryUnfork())
5832	+	s.exec();
5833	+	} while ((s = s.nextRight) != null);
5834	+	}
5835	+	return false;
5836	+	}
5837	+	public final U getRawResult() { return result; }
5838	+	}
5839	+
5840	+	@SuppressWarnings("serial") static final class MapReduceEntriesTask<K,V,U>
5841	+	extends BulkTask<K,V,U> {
5842	+	final Fun<Map.Entry<K,V>, ? extends U> transformer;
5843	+	final BiFun<? super U, ? super U, ? extends U> reducer;
5844	+	U result;
5845	+	MapReduceEntriesTask<K,V,U> rights, nextRight;
5846	+	MapReduceEntriesTask
5847	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
5848	+	MapReduceEntriesTask<K,V,U> nextRight,
5849	+	Fun<Map.Entry<K,V>, ? extends U> transformer,
5850	+	BiFun<? super U, ? super U, ? extends U> reducer) {
5851	+	super(m, p, b); this.nextRight = nextRight;
5852	+	this.transformer = transformer;
5853	+	this.reducer = reducer;
5854	+	}
5855	+	@SuppressWarnings("unchecked") public final boolean exec() {
5856	+	final Fun<Map.Entry<K,V>, ? extends U> transformer =
5857	+	this.transformer;
5858	+	final BiFun<? super U, ? super U, ? extends U> reducer =
5859	+	this.reducer;
5860	+	if (transformer == null || reducer == null)
5861	+	return abortOnNullFunction();
5862	+	try {
5863	+	for (int c, b = batch(); b > 1 && baseIndex != baseLimit;) {
5864	+	do {} while (!casPending(c = pending, c+1));
5865	+	(rights = new MapReduceEntriesTask<K,V,U>
5866	+	(map, this, b >>>= 1, rights, transformer, reducer)).fork();
5867	+	}
5868	+	U r = null, u;
5869	+	Object v;
5870	+	while ((v = advance()) != null) {
5871	+	if ((u = transformer.apply(entryFor((K)nextKey, (V)v))) != null)
5872	+	r = (r == null) ? u : reducer.apply(r, u);
5873	+	}
5874	+	result = r;
5875	+	for (MapReduceEntriesTask<K,V,U> t = this, s;;) {
5876	+	int c; BulkTask<K,V,?> par; U tr, sr;
5877	+	if ((c = t.pending) == 0) {
5878	+	for (s = t.rights; s != null; s = t.rights = s.nextRight) {
5879	+	if ((sr = s.result) != null)
5880	+	t.result = ((tr = t.result) == null) ? sr : reducer.apply(tr, sr);
5881	+	}
5882	+	if ((par = t.parent) == null ||
5883	+	!(par instanceof MapReduceEntriesTask)) {
5884	+	t.quietlyComplete();
5885	+	break;
5886	+	}
5887	+	t = (MapReduceEntriesTask<K,V,U>)par;
5888	+	}
5889	+	else if (t.casPending(c, c - 1))
5890	+	break;
5891	+	}
5892	+	} catch (Throwable ex) {
5893	+	return tryCompleteComputation(ex);
5894	+	}
5895	+	MapReduceEntriesTask<K,V,U> s = rights;
5896	+	if (s != null && !inForkJoinPool()) {
5897	+	do  {
5898	+	if (s.tryUnfork())
5899	+	s.exec();
5900	+	} while ((s = s.nextRight) != null);
5901	+	}
5902	+	return false;
5903	+	}
5904	+	public final U getRawResult() { return result; }
5905	+	}
5906	+
5907	+	@SuppressWarnings("serial") static final class MapReduceMappingsTask<K,V,U>
5908	+	extends BulkTask<K,V,U> {
5909	+	final BiFun<? super K, ? super V, ? extends U> transformer;
5910	+	final BiFun<? super U, ? super U, ? extends U> reducer;
5911	+	U result;
5912	+	MapReduceMappingsTask<K,V,U> rights, nextRight;
5913	+	MapReduceMappingsTask
5914	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
5915	+	MapReduceMappingsTask<K,V,U> nextRight,
5916	+	BiFun<? super K, ? super V, ? extends U> transformer,
5917	+	BiFun<? super U, ? super U, ? extends U> reducer) {
5918	+	super(m, p, b); this.nextRight = nextRight;
5919	+	this.transformer = transformer;
5920	+	this.reducer = reducer;
5921	+	}
5922	+	@SuppressWarnings("unchecked") public final boolean exec() {
5923	+	final BiFun<? super K, ? super V, ? extends U> transformer =
5924	+	this.transformer;
5925	+	final BiFun<? super U, ? super U, ? extends U> reducer =
5926	+	this.reducer;
5927	+	if (transformer == null || reducer == null)
5928	+	return abortOnNullFunction();
5929	+	try {
5930	+	for (int c, b = batch(); b > 1 && baseIndex != baseLimit;) {
5931	+	do {} while (!casPending(c = pending, c+1));
5932	+	(rights = new MapReduceMappingsTask<K,V,U>
5933	+	(map, this, b >>>= 1, rights, transformer, reducer)).fork();
5934	+	}
5935	+	U r = null, u;
5936	+	Object v;
5937	+	while ((v = advance()) != null) {
5938	+	if ((u = transformer.apply((K)nextKey, (V)v)) != null)
5939	+	r = (r == null) ? u : reducer.apply(r, u);
5940	+	}
5941	+	result = r;
5942	+	for (MapReduceMappingsTask<K,V,U> t = this, s;;) {
5943	+	int c; BulkTask<K,V,?> par; U tr, sr;
5944	+	if ((c = t.pending) == 0) {
5945	+	for (s = t.rights; s != null; s = t.rights = s.nextRight) {
5946	+	if ((sr = s.result) != null)
5947	+	t.result = ((tr = t.result) == null) ? sr : reducer.apply(tr, sr);
5948	+	}
5949	+	if ((par = t.parent) == null ||
5950	+	!(par instanceof MapReduceMappingsTask)) {
5951	+	t.quietlyComplete();
5952	+	break;
5953	+	}
5954	+	t = (MapReduceMappingsTask<K,V,U>)par;
5955	+	}
5956	+	else if (t.casPending(c, c - 1))
5957	+	break;
5958	+	}
5959	+	} catch (Throwable ex) {
5960	+	return tryCompleteComputation(ex);
5961	+	}
5962	+	MapReduceMappingsTask<K,V,U> s = rights;
5963	+	if (s != null && !inForkJoinPool()) {
5964	+	do  {
5965	+	if (s.tryUnfork())
5966	+	s.exec();
5967	+	} while ((s = s.nextRight) != null);
5968	+	}
5969	+	return false;
5970	+	}
5971	+	public final U getRawResult() { return result; }
5972	+	}
5973	+
5974	+	@SuppressWarnings("serial") static final class MapReduceKeysToDoubleTask<K,V>
5975	+	extends BulkTask<K,V,Double> {
5976	+	final ObjectToDouble<? super K> transformer;
5977	+	final DoubleByDoubleToDouble reducer;
5978	+	final double basis;
5979	+	double result;
5980	+	MapReduceKeysToDoubleTask<K,V> rights, nextRight;
5981	+	MapReduceKeysToDoubleTask
5982	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
5983	+	MapReduceKeysToDoubleTask<K,V> nextRight,
5984	+	ObjectToDouble<? super K> transformer,
5985	+	double basis,
5986	+	DoubleByDoubleToDouble reducer) {
5987	+	super(m, p, b); this.nextRight = nextRight;
5988	+	this.transformer = transformer;
5989	+	this.basis = basis; this.reducer = reducer;
5990	+	}
5991	+	@SuppressWarnings("unchecked") public final boolean exec() {
5992	+	final ObjectToDouble<? super K> transformer =
5993	+	this.transformer;
5994	+	final DoubleByDoubleToDouble reducer = this.reducer;
5995	+	if (transformer == null || reducer == null)
5996	+	return abortOnNullFunction();
5997	+	try {
5998	+	final double id = this.basis;
5999	+	for (int c, b = batch(); b > 1 && baseIndex != baseLimit;) {
6000	+	do {} while (!casPending(c = pending, c+1));
6001	+	(rights = new MapReduceKeysToDoubleTask<K,V>
6002	+	(map, this, b >>>= 1, rights, transformer, id, reducer)).fork();
6003	+	}
6004	+	double r = id;
6005	+	while (advance() != null)
6006	+	r = reducer.apply(r, transformer.apply((K)nextKey));
6007	+	result = r;
6008	+	for (MapReduceKeysToDoubleTask<K,V> t = this, s;;) {
6009	+	int c; BulkTask<K,V,?> par;
6010	+	if ((c = t.pending) == 0) {
6011	+	for (s = t.rights; s != null; s = t.rights = s.nextRight) {
6012	+	t.result = reducer.apply(t.result, s.result);
6013	+	}
6014	+	if ((par = t.parent) == null ||
6015	+	!(par instanceof MapReduceKeysToDoubleTask)) {
6016	+	t.quietlyComplete();
6017	+	break;
6018	+	}
6019	+	t = (MapReduceKeysToDoubleTask<K,V>)par;
6020	+	}
6021	+	else if (t.casPending(c, c - 1))
6022	+	break;
6023	+	}
6024	+	} catch (Throwable ex) {
6025	+	return tryCompleteComputation(ex);
6026	+	}
6027	+	MapReduceKeysToDoubleTask<K,V> s = rights;
6028	+	if (s != null && !inForkJoinPool()) {
6029	+	do  {
6030	+	if (s.tryUnfork())
6031	+	s.exec();
6032	+	} while ((s = s.nextRight) != null);
6033	+	}
6034	+	return false;
6035	+	}
6036	+	public final Double getRawResult() { return result; }
6037	+	}
6038	+
6039	+	@SuppressWarnings("serial") static final class MapReduceValuesToDoubleTask<K,V>
6040	+	extends BulkTask<K,V,Double> {
6041	+	final ObjectToDouble<? super V> transformer;
6042	+	final DoubleByDoubleToDouble reducer;
6043	+	final double basis;
6044	+	double result;
6045	+	MapReduceValuesToDoubleTask<K,V> rights, nextRight;
6046	+	MapReduceValuesToDoubleTask
6047	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
6048	+	MapReduceValuesToDoubleTask<K,V> nextRight,
6049	+	ObjectToDouble<? super V> transformer,
6050	+	double basis,
6051	+	DoubleByDoubleToDouble reducer) {
6052	+	super(m, p, b); this.nextRight = nextRight;
6053	+	this.transformer = transformer;
6054	+	this.basis = basis; this.reducer = reducer;
6055	+	}
6056	+	@SuppressWarnings("unchecked") public final boolean exec() {
6057	+	final ObjectToDouble<? super V> transformer =
6058	+	this.transformer;
6059	+	final DoubleByDoubleToDouble reducer = this.reducer;
6060	+	if (transformer == null || reducer == null)
6061	+	return abortOnNullFunction();
6062	+	try {
6063	+	final double id = this.basis;
6064	+	for (int c, b = batch(); b > 1 && baseIndex != baseLimit;) {
6065	+	do {} while (!casPending(c = pending, c+1));
6066	+	(rights = new MapReduceValuesToDoubleTask<K,V>
6067	+	(map, this, b >>>= 1, rights, transformer, id, reducer)).fork();
6068	+	}
6069	+	double r = id;
6070	+	Object v;
6071	+	while ((v = advance()) != null)
6072	+	r = reducer.apply(r, transformer.apply((V)v));
6073	+	result = r;
6074	+	for (MapReduceValuesToDoubleTask<K,V> t = this, s;;) {
6075	+	int c; BulkTask<K,V,?> par;
6076	+	if ((c = t.pending) == 0) {
6077	+	for (s = t.rights; s != null; s = t.rights = s.nextRight) {
6078	+	t.result = reducer.apply(t.result, s.result);
6079	+	}
6080	+	if ((par = t.parent) == null ||
6081	+	!(par instanceof MapReduceValuesToDoubleTask)) {
6082	+	t.quietlyComplete();
6083	+	break;
6084	+	}
6085	+	t = (MapReduceValuesToDoubleTask<K,V>)par;
6086	+	}
6087	+	else if (t.casPending(c, c - 1))
6088	+	break;
6089	+	}
6090	+	} catch (Throwable ex) {
6091	+	return tryCompleteComputation(ex);
6092	+	}
6093	+	MapReduceValuesToDoubleTask<K,V> s = rights;
6094	+	if (s != null && !inForkJoinPool()) {
6095	+	do  {
6096	+	if (s.tryUnfork())
6097	+	s.exec();
6098	+	} while ((s = s.nextRight) != null);
6099	+	}
6100	+	return false;
6101	+	}
6102	+	public final Double getRawResult() { return result; }
6103	+	}
6104	+
6105	+	@SuppressWarnings("serial") static final class MapReduceEntriesToDoubleTask<K,V>
6106	+	extends BulkTask<K,V,Double> {
6107	+	final ObjectToDouble<Map.Entry<K,V>> transformer;
6108	+	final DoubleByDoubleToDouble reducer;
6109	+	final double basis;
6110	+	double result;
6111	+	MapReduceEntriesToDoubleTask<K,V> rights, nextRight;
6112	+	MapReduceEntriesToDoubleTask
6113	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
6114	+	MapReduceEntriesToDoubleTask<K,V> nextRight,
6115	+	ObjectToDouble<Map.Entry<K,V>> transformer,
6116	+	double basis,
6117	+	DoubleByDoubleToDouble reducer) {
6118	+	super(m, p, b); this.nextRight = nextRight;
6119	+	this.transformer = transformer;
6120	+	this.basis = basis; this.reducer = reducer;
6121	+	}
6122	+	@SuppressWarnings("unchecked") public final boolean exec() {
6123	+	final ObjectToDouble<Map.Entry<K,V>> transformer =
6124	+	this.transformer;
6125	+	final DoubleByDoubleToDouble reducer = this.reducer;
6126	+	if (transformer == null || reducer == null)
6127	+	return abortOnNullFunction();
6128	+	try {
6129	+	final double id = this.basis;
6130	+	for (int c, b = batch(); b > 1 && baseIndex != baseLimit;) {
6131	+	do {} while (!casPending(c = pending, c+1));
6132	+	(rights = new MapReduceEntriesToDoubleTask<K,V>
6133	+	(map, this, b >>>= 1, rights, transformer, id, reducer)).fork();
6134	+	}
6135	+	double r = id;
6136	+	Object v;
6137	+	while ((v = advance()) != null)
6138	+	r = reducer.apply(r, transformer.apply(entryFor((K)nextKey, (V)v)));
6139	+	result = r;
6140	+	for (MapReduceEntriesToDoubleTask<K,V> t = this, s;;) {
6141	+	int c; BulkTask<K,V,?> par;
6142	+	if ((c = t.pending) == 0) {
6143	+	for (s = t.rights; s != null; s = t.rights = s.nextRight) {
6144	+	t.result = reducer.apply(t.result, s.result);
6145	+	}
6146	+	if ((par = t.parent) == null ||
6147	+	!(par instanceof MapReduceEntriesToDoubleTask)) {
6148	+	t.quietlyComplete();
6149	+	break;
6150	+	}
6151	+	t = (MapReduceEntriesToDoubleTask<K,V>)par;
6152	+	}
6153	+	else if (t.casPending(c, c - 1))
6154	+	break;
6155	+	}
6156	+	} catch (Throwable ex) {
6157	+	return tryCompleteComputation(ex);
6158	+	}
6159	+	MapReduceEntriesToDoubleTask<K,V> s = rights;
6160	+	if (s != null && !inForkJoinPool()) {
6161	+	do  {
6162	+	if (s.tryUnfork())
6163	+	s.exec();
6164	+	} while ((s = s.nextRight) != null);
6165	+	}
6166	+	return false;
6167	+	}
6168	+	public final Double getRawResult() { return result; }
6169	+	}
6170	+
6171	+	@SuppressWarnings("serial") static final class MapReduceMappingsToDoubleTask<K,V>
6172	+	extends BulkTask<K,V,Double> {
6173	+	final ObjectByObjectToDouble<? super K, ? super V> transformer;
6174	+	final DoubleByDoubleToDouble reducer;
6175	+	final double basis;
6176	+	double result;
6177	+	MapReduceMappingsToDoubleTask<K,V> rights, nextRight;
6178	+	MapReduceMappingsToDoubleTask
6179	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
6180	+	MapReduceMappingsToDoubleTask<K,V> nextRight,
6181	+	ObjectByObjectToDouble<? super K, ? super V> transformer,
6182	+	double basis,
6183	+	DoubleByDoubleToDouble reducer) {
6184	+	super(m, p, b); this.nextRight = nextRight;
6185	+	this.transformer = transformer;
6186	+	this.basis = basis; this.reducer = reducer;
6187	+	}
6188	+	@SuppressWarnings("unchecked") public final boolean exec() {
6189	+	final ObjectByObjectToDouble<? super K, ? super V> transformer =
6190	+	this.transformer;
6191	+	final DoubleByDoubleToDouble reducer = this.reducer;
6192	+	if (transformer == null || reducer == null)
6193	+	return abortOnNullFunction();
6194	+	try {
6195	+	final double id = this.basis;
6196	+	for (int c, b = batch(); b > 1 && baseIndex != baseLimit;) {
6197	+	do {} while (!casPending(c = pending, c+1));
6198	+	(rights = new MapReduceMappingsToDoubleTask<K,V>
6199	+	(map, this, b >>>= 1, rights, transformer, id, reducer)).fork();
6200	+	}
6201	+	double r = id;
6202	+	Object v;
6203	+	while ((v = advance()) != null)
6204	+	r = reducer.apply(r, transformer.apply((K)nextKey, (V)v));
6205	+	result = r;
6206	+	for (MapReduceMappingsToDoubleTask<K,V> t = this, s;;) {
6207	+	int c; BulkTask<K,V,?> par;
6208	+	if ((c = t.pending) == 0) {
6209	+	for (s = t.rights; s != null; s = t.rights = s.nextRight) {
6210	+	t.result = reducer.apply(t.result, s.result);
6211	+	}
6212	+	if ((par = t.parent) == null ||
6213	+	!(par instanceof MapReduceMappingsToDoubleTask)) {
6214	+	t.quietlyComplete();
6215	+	break;
6216	+	}
6217	+	t = (MapReduceMappingsToDoubleTask<K,V>)par;
6218	+	}
6219	+	else if (t.casPending(c, c - 1))
6220	+	break;
6221	+	}
6222	+	} catch (Throwable ex) {
6223	+	return tryCompleteComputation(ex);
6224	+	}
6225	+	MapReduceMappingsToDoubleTask<K,V> s = rights;
6226	+	if (s != null && !inForkJoinPool()) {
6227	+	do  {
6228	+	if (s.tryUnfork())
6229	+	s.exec();
6230	+	} while ((s = s.nextRight) != null);
6231	+	}
6232	+	return false;
6233	+	}
6234	+	public final Double getRawResult() { return result; }
6235	+	}
6236	+
6237	+	@SuppressWarnings("serial") static final class MapReduceKeysToLongTask<K,V>
6238	+	extends BulkTask<K,V,Long> {
6239	+	final ObjectToLong<? super K> transformer;
6240	+	final LongByLongToLong reducer;
6241	+	final long basis;
6242	+	long result;
6243	+	MapReduceKeysToLongTask<K,V> rights, nextRight;
6244	+	MapReduceKeysToLongTask
6245	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
6246	+	MapReduceKeysToLongTask<K,V> nextRight,
6247	+	ObjectToLong<? super K> transformer,
6248	+	long basis,
6249	+	LongByLongToLong reducer) {
6250	+	super(m, p, b); this.nextRight = nextRight;
6251	+	this.transformer = transformer;
6252	+	this.basis = basis; this.reducer = reducer;
6253	+	}
6254	+	@SuppressWarnings("unchecked") public final boolean exec() {
6255	+	final ObjectToLong<? super K> transformer =
6256	+	this.transformer;
6257	+	final LongByLongToLong reducer = this.reducer;
6258	+	if (transformer == null || reducer == null)
6259	+	return abortOnNullFunction();
6260	+	try {
6261	+	final long id = this.basis;
6262	+	for (int c, b = batch(); b > 1 && baseIndex != baseLimit;) {
6263	+	do {} while (!casPending(c = pending, c+1));
6264	+	(rights = new MapReduceKeysToLongTask<K,V>
6265	+	(map, this, b >>>= 1, rights, transformer, id, reducer)).fork();
6266	+	}
6267	+	long r = id;
6268	+	while (advance() != null)
6269	+	r = reducer.apply(r, transformer.apply((K)nextKey));
6270	+	result = r;
6271	+	for (MapReduceKeysToLongTask<K,V> t = this, s;;) {
6272	+	int c; BulkTask<K,V,?> par;
6273	+	if ((c = t.pending) == 0) {
6274	+	for (s = t.rights; s != null; s = t.rights = s.nextRight) {
6275	+	t.result = reducer.apply(t.result, s.result);
6276	+	}
6277	+	if ((par = t.parent) == null ||
6278	+	!(par instanceof MapReduceKeysToLongTask)) {
6279	+	t.quietlyComplete();
6280	+	break;
6281	+	}
6282	+	t = (MapReduceKeysToLongTask<K,V>)par;
6283	+	}
6284	+	else if (t.casPending(c, c - 1))
6285	+	break;
6286	+	}
6287	+	} catch (Throwable ex) {
6288	+	return tryCompleteComputation(ex);
6289	+	}
6290	+	MapReduceKeysToLongTask<K,V> s = rights;
6291	+	if (s != null && !inForkJoinPool()) {
6292	+	do  {
6293	+	if (s.tryUnfork())
6294	+	s.exec();
6295	+	} while ((s = s.nextRight) != null);
6296	+	}
6297	+	return false;
6298	+	}
6299	+	public final Long getRawResult() { return result; }
6300	+	}
6301	+
6302	+	@SuppressWarnings("serial") static final class MapReduceValuesToLongTask<K,V>
6303	+	extends BulkTask<K,V,Long> {
6304	+	final ObjectToLong<? super V> transformer;
6305	+	final LongByLongToLong reducer;
6306	+	final long basis;
6307	+	long result;
6308	+	MapReduceValuesToLongTask<K,V> rights, nextRight;
6309	+	MapReduceValuesToLongTask
6310	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
6311	+	MapReduceValuesToLongTask<K,V> nextRight,
6312	+	ObjectToLong<? super V> transformer,
6313	+	long basis,
6314	+	LongByLongToLong reducer) {
6315	+	super(m, p, b); this.nextRight = nextRight;
6316	+	this.transformer = transformer;
6317	+	this.basis = basis; this.reducer = reducer;
6318	+	}
6319	+	@SuppressWarnings("unchecked") public final boolean exec() {
6320	+	final ObjectToLong<? super V> transformer =
6321	+	this.transformer;
6322	+	final LongByLongToLong reducer = this.reducer;
6323	+	if (transformer == null || reducer == null)
6324	+	return abortOnNullFunction();
6325	+	try {
6326	+	final long id = this.basis;
6327	+	for (int c, b = batch(); b > 1 && baseIndex != baseLimit;) {
6328	+	do {} while (!casPending(c = pending, c+1));
6329	+	(rights = new MapReduceValuesToLongTask<K,V>
6330	+	(map, this, b >>>= 1, rights, transformer, id, reducer)).fork();
6331	+	}
6332	+	long r = id;
6333	+	Object v;
6334	+	while ((v = advance()) != null)
6335	+	r = reducer.apply(r, transformer.apply((V)v));
6336	+	result = r;
6337	+	for (MapReduceValuesToLongTask<K,V> t = this, s;;) {
6338	+	int c; BulkTask<K,V,?> par;
6339	+	if ((c = t.pending) == 0) {
6340	+	for (s = t.rights; s != null; s = t.rights = s.nextRight) {
6341	+	t.result = reducer.apply(t.result, s.result);
6342	+	}
6343	+	if ((par = t.parent) == null ||
6344	+	!(par instanceof MapReduceValuesToLongTask)) {
6345	+	t.quietlyComplete();
6346	+	break;
6347	+	}
6348	+	t = (MapReduceValuesToLongTask<K,V>)par;
6349	+	}
6350	+	else if (t.casPending(c, c - 1))
6351	+	break;
6352	+	}
6353	+	} catch (Throwable ex) {
6354	+	return tryCompleteComputation(ex);
6355	+	}
6356	+	MapReduceValuesToLongTask<K,V> s = rights;
6357	+	if (s != null && !inForkJoinPool()) {
6358	+	do  {
6359	+	if (s.tryUnfork())
6360	+	s.exec();
6361	+	} while ((s = s.nextRight) != null);
6362	+	}
6363	+	return false;
6364	+	}
6365	+	public final Long getRawResult() { return result; }
6366	+	}
6367	+
6368	+	@SuppressWarnings("serial") static final class MapReduceEntriesToLongTask<K,V>
6369	+	extends BulkTask<K,V,Long> {
6370	+	final ObjectToLong<Map.Entry<K,V>> transformer;
6371	+	final LongByLongToLong reducer;
6372	+	final long basis;
6373	+	long result;
6374	+	MapReduceEntriesToLongTask<K,V> rights, nextRight;
6375	+	MapReduceEntriesToLongTask
6376	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
6377	+	MapReduceEntriesToLongTask<K,V> nextRight,
6378	+	ObjectToLong<Map.Entry<K,V>> transformer,
6379	+	long basis,
6380	+	LongByLongToLong reducer) {
6381	+	super(m, p, b); this.nextRight = nextRight;
6382	+	this.transformer = transformer;
6383	+	this.basis = basis; this.reducer = reducer;
6384	+	}
6385	+	@SuppressWarnings("unchecked") public final boolean exec() {
6386	+	final ObjectToLong<Map.Entry<K,V>> transformer =
6387	+	this.transformer;
6388	+	final LongByLongToLong reducer = this.reducer;
6389	+	if (transformer == null || reducer == null)
6390	+	return abortOnNullFunction();
6391	+	try {
6392	+	final long id = this.basis;
6393	+	for (int c, b = batch(); b > 1 && baseIndex != baseLimit;) {
6394	+	do {} while (!casPending(c = pending, c+1));
6395	+	(rights = new MapReduceEntriesToLongTask<K,V>
6396	+	(map, this, b >>>= 1, rights, transformer, id, reducer)).fork();
6397	+	}
6398	+	long r = id;
6399	+	Object v;
6400	+	while ((v = advance()) != null)
6401	+	r = reducer.apply(r, transformer.apply(entryFor((K)nextKey, (V)v)));
6402	+	result = r;
6403	+	for (MapReduceEntriesToLongTask<K,V> t = this, s;;) {
6404	+	int c; BulkTask<K,V,?> par;
6405	+	if ((c = t.pending) == 0) {
6406	+	for (s = t.rights; s != null; s = t.rights = s.nextRight) {
6407	+	t.result = reducer.apply(t.result, s.result);
6408	+	}
6409	+	if ((par = t.parent) == null ||
6410	+	!(par instanceof MapReduceEntriesToLongTask)) {
6411	+	t.quietlyComplete();
6412	+	break;
6413	+	}
6414	+	t = (MapReduceEntriesToLongTask<K,V>)par;
6415	+	}
6416	+	else if (t.casPending(c, c - 1))
6417	+	break;
6418	+	}
6419	+	} catch (Throwable ex) {
6420	+	return tryCompleteComputation(ex);
6421	+	}
6422	+	MapReduceEntriesToLongTask<K,V> s = rights;
6423	+	if (s != null && !inForkJoinPool()) {
6424	+	do  {
6425	+	if (s.tryUnfork())
6426	+	s.exec();
6427	+	} while ((s = s.nextRight) != null);
6428	+	}
6429	+	return false;
6430	+	}
6431	+	public final Long getRawResult() { return result; }
6432	+	}
6433	+
6434	+	@SuppressWarnings("serial") static final class MapReduceMappingsToLongTask<K,V>
6435	+	extends BulkTask<K,V,Long> {
6436	+	final ObjectByObjectToLong<? super K, ? super V> transformer;
6437	+	final LongByLongToLong reducer;
6438	+	final long basis;
6439	+	long result;
6440	+	MapReduceMappingsToLongTask<K,V> rights, nextRight;
6441	+	MapReduceMappingsToLongTask
6442	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
6443	+	MapReduceMappingsToLongTask<K,V> nextRight,
6444	+	ObjectByObjectToLong<? super K, ? super V> transformer,
6445	+	long basis,
6446	+	LongByLongToLong reducer) {
6447	+	super(m, p, b); this.nextRight = nextRight;
6448	+	this.transformer = transformer;
6449	+	this.basis = basis; this.reducer = reducer;
6450	+	}
6451	+	@SuppressWarnings("unchecked") public final boolean exec() {
6452	+	final ObjectByObjectToLong<? super K, ? super V> transformer =
6453	+	this.transformer;
6454	+	final LongByLongToLong reducer = this.reducer;
6455	+	if (transformer == null || reducer == null)
6456	+	return abortOnNullFunction();
6457	+	try {
6458	+	final long id = this.basis;
6459	+	for (int c, b = batch(); b > 1 && baseIndex != baseLimit;) {
6460	+	do {} while (!casPending(c = pending, c+1));
6461	+	(rights = new MapReduceMappingsToLongTask<K,V>
6462	+	(map, this, b >>>= 1, rights, transformer, id, reducer)).fork();
6463	+	}
6464	+	long r = id;
6465	+	Object v;
6466	+	while ((v = advance()) != null)
6467	+	r = reducer.apply(r, transformer.apply((K)nextKey, (V)v));
6468	+	result = r;
6469	+	for (MapReduceMappingsToLongTask<K,V> t = this, s;;) {
6470	+	int c; BulkTask<K,V,?> par;
6471	+	if ((c = t.pending) == 0) {
6472	+	for (s = t.rights; s != null; s = t.rights = s.nextRight) {
6473	+	t.result = reducer.apply(t.result, s.result);
6474	+	}
6475	+	if ((par = t.parent) == null ||
6476	+	!(par instanceof MapReduceMappingsToLongTask)) {
6477	+	t.quietlyComplete();
6478	+	break;
6479	+	}
6480	+	t = (MapReduceMappingsToLongTask<K,V>)par;
6481	+	}
6482	+	else if (t.casPending(c, c - 1))
6483	+	break;
6484	+	}
6485	+	} catch (Throwable ex) {
6486	+	return tryCompleteComputation(ex);
6487	+	}
6488	+	MapReduceMappingsToLongTask<K,V> s = rights;
6489	+	if (s != null && !inForkJoinPool()) {
6490	+	do  {
6491	+	if (s.tryUnfork())
6492	+	s.exec();
6493	+	} while ((s = s.nextRight) != null);
6494	+	}
6495	+	return false;
6496	+	}
6497	+	public final Long getRawResult() { return result; }
6498	+	}
6499	+
6500	+	@SuppressWarnings("serial") static final class MapReduceKeysToIntTask<K,V>
6501	+	extends BulkTask<K,V,Integer> {
6502	+	final ObjectToInt<? super K> transformer;
6503	+	final IntByIntToInt reducer;
6504	+	final int basis;
6505	+	int result;
6506	+	MapReduceKeysToIntTask<K,V> rights, nextRight;
6507	+	MapReduceKeysToIntTask
6508	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
6509	+	MapReduceKeysToIntTask<K,V> nextRight,
6510	+	ObjectToInt<? super K> transformer,
6511	+	int basis,
6512	+	IntByIntToInt reducer) {
6513	+	super(m, p, b); this.nextRight = nextRight;
6514	+	this.transformer = transformer;
6515	+	this.basis = basis; this.reducer = reducer;
6516	+	}
6517	+	@SuppressWarnings("unchecked") public final boolean exec() {
6518	+	final ObjectToInt<? super K> transformer =
6519	+	this.transformer;
6520	+	final IntByIntToInt reducer = this.reducer;
6521	+	if (transformer == null || reducer == null)
6522	+	return abortOnNullFunction();
6523	+	try {
6524	+	final int id = this.basis;
6525	+	for (int c, b = batch(); b > 1 && baseIndex != baseLimit;) {
6526	+	do {} while (!casPending(c = pending, c+1));
6527	+	(rights = new MapReduceKeysToIntTask<K,V>
6528	+	(map, this, b >>>= 1, rights, transformer, id, reducer)).fork();
6529	+	}
6530	+	int r = id;
6531	+	while (advance() != null)
6532	+	r = reducer.apply(r, transformer.apply((K)nextKey));
6533	+	result = r;
6534	+	for (MapReduceKeysToIntTask<K,V> t = this, s;;) {
6535	+	int c; BulkTask<K,V,?> par;
6536	+	if ((c = t.pending) == 0) {
6537	+	for (s = t.rights; s != null; s = t.rights = s.nextRight) {
6538	+	t.result = reducer.apply(t.result, s.result);
6539	+	}
6540	+	if ((par = t.parent) == null ||
6541	+	!(par instanceof MapReduceKeysToIntTask)) {
6542	+	t.quietlyComplete();
6543	+	break;
6544	+	}
6545	+	t = (MapReduceKeysToIntTask<K,V>)par;
6546	+	}
6547	+	else if (t.casPending(c, c - 1))
6548	+	break;
6549	+	}
6550	+	} catch (Throwable ex) {
6551	+	return tryCompleteComputation(ex);
6552	+	}
6553	+	MapReduceKeysToIntTask<K,V> s = rights;
6554	+	if (s != null && !inForkJoinPool()) {
6555	+	do  {
6556	+	if (s.tryUnfork())
6557	+	s.exec();
6558	+	} while ((s = s.nextRight) != null);
6559	+	}
6560	+	return false;
6561	+	}
6562	+	public final Integer getRawResult() { return result; }
6563	+	}
6564	+
6565	+	@SuppressWarnings("serial") static final class MapReduceValuesToIntTask<K,V>
6566	+	extends BulkTask<K,V,Integer> {
6567	+	final ObjectToInt<? super V> transformer;
6568	+	final IntByIntToInt reducer;
6569	+	final int basis;
6570	+	int result;
6571	+	MapReduceValuesToIntTask<K,V> rights, nextRight;
6572	+	MapReduceValuesToIntTask
6573	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
6574	+	MapReduceValuesToIntTask<K,V> nextRight,
6575	+	ObjectToInt<? super V> transformer,
6576	+	int basis,
6577	+	IntByIntToInt reducer) {
6578	+	super(m, p, b); this.nextRight = nextRight;
6579	+	this.transformer = transformer;
6580	+	this.basis = basis; this.reducer = reducer;
6581	+	}
6582	+	@SuppressWarnings("unchecked") public final boolean exec() {
6583	+	final ObjectToInt<? super V> transformer =
6584	+	this.transformer;
6585	+	final IntByIntToInt reducer = this.reducer;
6586	+	if (transformer == null || reducer == null)
6587	+	return abortOnNullFunction();
6588	+	try {
6589	+	final int id = this.basis;
6590	+	for (int c, b = batch(); b > 1 && baseIndex != baseLimit;) {
6591	+	do {} while (!casPending(c = pending, c+1));
6592	+	(rights = new MapReduceValuesToIntTask<K,V>
6593	+	(map, this, b >>>= 1, rights, transformer, id, reducer)).fork();
6594	+	}
6595	+	int r = id;
6596	+	Object v;
6597	+	while ((v = advance()) != null)
6598	+	r = reducer.apply(r, transformer.apply((V)v));
6599	+	result = r;
6600	+	for (MapReduceValuesToIntTask<K,V> t = this, s;;) {
6601	+	int c; BulkTask<K,V,?> par;
6602	+	if ((c = t.pending) == 0) {
6603	+	for (s = t.rights; s != null; s = t.rights = s.nextRight) {
6604	+	t.result = reducer.apply(t.result, s.result);
6605	+	}
6606	+	if ((par = t.parent) == null ||
6607	+	!(par instanceof MapReduceValuesToIntTask)) {
6608	+	t.quietlyComplete();
6609	+	break;
6610	+	}
6611	+	t = (MapReduceValuesToIntTask<K,V>)par;
6612	+	}
6613	+	else if (t.casPending(c, c - 1))
6614	+	break;
6615	+	}
6616	+	} catch (Throwable ex) {
6617	+	return tryCompleteComputation(ex);
6618	+	}
6619	+	MapReduceValuesToIntTask<K,V> s = rights;
6620	+	if (s != null && !inForkJoinPool()) {
6621	+	do  {
6622	+	if (s.tryUnfork())
6623	+	s.exec();
6624	+	} while ((s = s.nextRight) != null);
6625	+	}
6626	+	return false;
6627	+	}
6628	+	public final Integer getRawResult() { return result; }
6629	+	}
6630	+
6631	+	@SuppressWarnings("serial") static final class MapReduceEntriesToIntTask<K,V>
6632	+	extends BulkTask<K,V,Integer> {
6633	+	final ObjectToInt<Map.Entry<K,V>> transformer;
6634	+	final IntByIntToInt reducer;
6635	+	final int basis;
6636	+	int result;
6637	+	MapReduceEntriesToIntTask<K,V> rights, nextRight;
6638	+	MapReduceEntriesToIntTask
6639	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
6640	+	MapReduceEntriesToIntTask<K,V> nextRight,
6641	+	ObjectToInt<Map.Entry<K,V>> transformer,
6642	+	int basis,
6643	+	IntByIntToInt reducer) {
6644	+	super(m, p, b); this.nextRight = nextRight;
6645	+	this.transformer = transformer;
6646	+	this.basis = basis; this.reducer = reducer;
6647	+	}
6648	+	@SuppressWarnings("unchecked") public final boolean exec() {
6649	+	final ObjectToInt<Map.Entry<K,V>> transformer =
6650	+	this.transformer;
6651	+	final IntByIntToInt reducer = this.reducer;
6652	+	if (transformer == null || reducer == null)
6653	+	return abortOnNullFunction();
6654	+	try {
6655	+	final int id = this.basis;
6656	+	for (int c, b = batch(); b > 1 && baseIndex != baseLimit;) {
6657	+	do {} while (!casPending(c = pending, c+1));
6658	+	(rights = new MapReduceEntriesToIntTask<K,V>
6659	+	(map, this, b >>>= 1, rights, transformer, id, reducer)).fork();
6660	+	}
6661	+	int r = id;
6662	+	Object v;
6663	+	while ((v = advance()) != null)
6664	+	r = reducer.apply(r, transformer.apply(entryFor((K)nextKey, (V)v)));
6665	+	result = r;
6666	+	for (MapReduceEntriesToIntTask<K,V> t = this, s;;) {
6667	+	int c; BulkTask<K,V,?> par;
6668	+	if ((c = t.pending) == 0) {
6669	+	for (s = t.rights; s != null; s = t.rights = s.nextRight) {
6670	+	t.result = reducer.apply(t.result, s.result);
6671	+	}
6672	+	if ((par = t.parent) == null ||
6673	+	!(par instanceof MapReduceEntriesToIntTask)) {
6674	+	t.quietlyComplete();
6675	+	break;
6676	+	}
6677	+	t = (MapReduceEntriesToIntTask<K,V>)par;
6678	+	}
6679	+	else if (t.casPending(c, c - 1))
6680	+	break;
6681	+	}
6682	+	} catch (Throwable ex) {
6683	+	return tryCompleteComputation(ex);
6684	+	}
6685	+	MapReduceEntriesToIntTask<K,V> s = rights;
6686	+	if (s != null && !inForkJoinPool()) {
6687	+	do  {
6688	+	if (s.tryUnfork())
6689	+	s.exec();
6690	+	} while ((s = s.nextRight) != null);
6691	+	}
6692	+	return false;
6693	+	}
6694	+	public final Integer getRawResult() { return result; }
6695	+	}
6696	+
6697	+	@SuppressWarnings("serial") static final class MapReduceMappingsToIntTask<K,V>
6698	+	extends BulkTask<K,V,Integer> {
6699	+	final ObjectByObjectToInt<? super K, ? super V> transformer;
6700	+	final IntByIntToInt reducer;
6701	+	final int basis;
6702	+	int result;
6703	+	MapReduceMappingsToIntTask<K,V> rights, nextRight;
6704	+	MapReduceMappingsToIntTask
6705	+	(ConcurrentHashMap<K,V> m, BulkTask<K,V,?> p, int b,
6706	+	MapReduceMappingsToIntTask<K,V> rights,
6707	+	ObjectByObjectToInt<? super K, ? super V> transformer,
6708	+	int basis,
6709	+	IntByIntToInt reducer) {
6710	+	super(m, p, b); this.nextRight = nextRight;
6711	+	this.transformer = transformer;
6712	+	this.basis = basis; this.reducer = reducer;
6713	+	}
6714	+	@SuppressWarnings("unchecked") public final boolean exec() {
6715	+	final ObjectByObjectToInt<? super K, ? super V> transformer =
6716	+	this.transformer;
6717	+	final IntByIntToInt reducer = this.reducer;
6718	+	if (transformer == null || reducer == null)
6719	+	return abortOnNullFunction();
6720	+	try {
6721	+	final int id = this.basis;
6722	+	for (int c, b = batch(); b > 1 && baseIndex != baseLimit;) {
6723	+	do {} while (!casPending(c = pending, c+1));
6724	+	(rights = new MapReduceMappingsToIntTask<K,V>
6725	+	(map, this, b >>>= 1, rights, transformer, id, reducer)).fork();
6726	+	}
6727	+	int r = id;
6728	+	Object v;
6729	+	while ((v = advance()) != null)
6730	+	r = reducer.apply(r, transformer.apply((K)nextKey, (V)v));
6731	+	result = r;
6732	+	for (MapReduceMappingsToIntTask<K,V> t = this, s;;) {
6733	+	int c; BulkTask<K,V,?> par;
6734	+	if ((c = t.pending) == 0) {
6735	+	for (s = t.rights; s != null; s = t.rights = s.nextRight) {
6736	+	t.result = reducer.apply(t.result, s.result);
6737	+	}
6738	+	if ((par = t.parent) == null ||
6739	+	!(par instanceof MapReduceMappingsToIntTask)) {
6740	+	t.quietlyComplete();
6741	+	break;
6742	+	}
6743	+	t = (MapReduceMappingsToIntTask<K,V>)par;
6744	+	}
6745	+	else if (t.casPending(c, c - 1))
6746	+	break;
6747	+	}
6748	+	} catch (Throwable ex) {
6749	+	return tryCompleteComputation(ex);
6750	+	}
6751	+	MapReduceMappingsToIntTask<K,V> s = rights;
6752	+	if (s != null && !inForkJoinPool()) {
6753	+	do  {
6754	+	if (s.tryUnfork())
6755	+	s.exec();
6756	+	} while ((s = s.nextRight) != null);
6757	+	}
6758	+	return false;
6759	+	}
6760	+	public final Integer getRawResult() { return result; }
6761	+	}
6762	+
6763	+	// Unsafe mechanics
6764	+	private static final sun.misc.Unsafe UNSAFE;
6765	+	private static final long counterOffset;
6766	+	private static final long sizeCtlOffset;
6767	+	private static final long ABASE;
6768	+	private static final int ASHIFT;
6769	+
6770	+	static {
6771	+	int ss;
6772	+	try {
6773	+	UNSAFE = sun.misc.Unsafe.getUnsafe();
6774	+	Class<?> k = ConcurrentHashMap.class;
6775	+	counterOffset = UNSAFE.objectFieldOffset
6776	+	(k.getDeclaredField("counter"));
6777	+	sizeCtlOffset = UNSAFE.objectFieldOffset
6778	+	(k.getDeclaredField("sizeCtl"));
6779	+	Class<?> sc = Node[].class;
6780	+	ABASE = UNSAFE.arrayBaseOffset(sc);
6781	+	ss = UNSAFE.arrayIndexScale(sc);
6782	+	} catch (Exception e) {
6783	+	throw new Error(e);
6784	+	}
6785	+	if ((ss & (ss-1)) != 0)
6786	+	throw new Error("data type scale not a power of two");
6787	+	ASHIFT = 31 - Integer.numberOfLeadingZeros(ss);
6788		}
6789		}

Diff Legend

–	Removed lines
+	Added lines
<	Changed lines
>	Changed lines