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 1.149 by dl, Thu Dec 13 20:34:05 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.ForkJoinPool;
9	>	import java.util.concurrent.CountedCompleter;
10	>
11	>	import java.util.Comparator;
12	>	import java.util.Arrays;
13	>	import java.util.Map;
14	>	import java.util.Set;
15	>	import java.util.Collection;
16	>	import java.util.AbstractMap;
17	>	import java.util.AbstractSet;
18	>	import java.util.AbstractCollection;
19	>	import java.util.Hashtable;
20	>	import java.util.HashMap;
21	>	import java.util.Iterator;
22	>	import java.util.Enumeration;
23	>	import java.util.ConcurrentModificationException;
24	>	import java.util.NoSuchElementException;
25	>	import java.util.concurrent.ConcurrentMap;
26	>	import java.util.concurrent.locks.AbstractQueuedSynchronizer;
27	>	import java.util.concurrent.atomic.AtomicInteger;
28	>	import java.util.concurrent.atomic.AtomicReference;
29		import java.io.Serializable;
11	–	import java.io.IOException;
12	–	import java.io.ObjectInputStream;
13	–	import java.io.ObjectOutputStream;
30
31		/**
32		* A hash table supporting full concurrency of retrievals and
33	<	* adjustable expected concurrency for updates. This class obeys the
33	>	* high expected concurrency for updates. This class obeys the
34		* same functional specification as {@link java.util.Hashtable}, and
35		* includes versions of methods corresponding to each method of
36	<	* <tt>Hashtable</tt>. However, even though all operations are
36	>	* {@code Hashtable}. However, even though all operations are
37		* thread-safe, retrieval operations do <em>not</em> entail locking,
38		* and there is <em>not</em> any support for locking the entire table
39		* in a way that prevents all access.  This class is fully
40	<	* interoperable with <tt>Hashtable</tt> in programs that rely on its
40	>	* interoperable with {@code Hashtable} in programs that rely on its
41		* thread safety but not on its synchronization details.
42		*
43	<	* <p> Retrieval operations (including <tt>get</tt>) generally do not
44	<	* block, so may overlap with update operations (including
45	<	* <tt>put</tt> and <tt>remove</tt>). Retrievals reflect the results
46	<	* of the most recently <em>completed</em> update operations holding
47	<	* upon their onset.  For aggregate operations such as <tt>putAll</tt>
48	<	* and <tt>clear</tt>, concurrent retrievals may reflect insertion or
49	<	* removal of only some entries.  Similarly, Iterators and
50	<	* Enumerations return elements reflecting the state of the hash table
51	<	* at some point at or since the creation of the iterator/enumeration.
52	<	* They do <em>not</em> throw {@link ConcurrentModificationException}.
53	<	* However, iterators are designed to be used by only one thread at a time.
54	<	*
55	<	* <p> The allowed concurrency among update operations is guided by
56	<	* the optional <tt>concurrencyLevel</tt> constructor argument
57	<	* (default <tt>16</tt>), which is used as a hint for internal sizing.  The
58	<	* table is internally partitioned to try to permit the indicated
59	<	* number of concurrent updates without contention. Because placement
60	<	* in hash tables is essentially random, the actual concurrency will
61	<	* vary.  Ideally, you should choose a value to accommodate as many
62	<	* threads as will ever concurrently modify the table. Using a
63	<	* significantly higher value than you need can waste space and time,
64	<	* and a significantly lower value can lead to thread contention. But
65	<	* overestimates and underestimates within an order of magnitude do
66	<	* not usually have much noticeable impact. A value of one is
67	<	* appropriate when it is known that only one thread will modify and
68	<	* all others will only read. Also, resizing this or any other kind of
69	<	* hash table is a relatively slow operation, so, when possible, it is
70	<	* a good idea to provide estimates of expected table sizes in
71	<	* constructors.
43	>	* <p>Retrieval operations (including {@code get}) generally do not
44	>	* block, so may overlap with update operations (including {@code put}
45	>	* and {@code remove}). Retrievals reflect the results of the most
46	>	* recently <em>completed</em> update operations holding upon their
47	>	* onset. (More formally, an update operation for a given key bears a
48	>	* <em>happens-before</em> relation with any (non-null) retrieval for
49	>	* that key reporting the updated value.)  For aggregate operations
50	>	* such as {@code putAll} and {@code clear}, concurrent retrievals may
51	>	* reflect insertion or removal of only some entries.  Similarly,
52	>	* Iterators and Enumerations return elements reflecting the state of
53	>	* the hash table at some point at or since the creation of the
54	>	* iterator/enumeration.  They do <em>not</em> throw {@link
55	>	* ConcurrentModificationException}.  However, iterators are designed
56	>	* to be used by only one thread at a time.  Bear in mind that the
57	>	* results of aggregate status methods including {@code size}, {@code
58	>	* isEmpty}, and {@code containsValue} are typically useful only when
59	>	* a map is not undergoing concurrent updates in other threads.
60	>	* Otherwise the results of these methods reflect transient states
61	>	* that may be adequate for monitoring or estimation purposes, but not
62	>	* for program control.
63	>	*
64	>	* <p>The table is dynamically expanded when there are too many
65	>	* collisions (i.e., keys that have distinct hash codes but fall into
66	>	* the same slot modulo the table size), with the expected average
67	>	* effect of maintaining roughly two bins per mapping (corresponding
68	>	* to a 0.75 load factor threshold for resizing). There may be much
69	>	* variance around this average as mappings are added and removed, but
70	>	* overall, this maintains a commonly accepted time/space tradeoff for
71	>	* hash tables.  However, resizing this or any other kind of hash
72	>	* table may be a relatively slow operation. When possible, it is a
73	>	* good idea to provide a size estimate as an optional {@code
74	>	* initialCapacity} constructor argument. An additional optional
75	>	* {@code loadFactor} constructor argument provides a further means of
76	>	* customizing initial table capacity by specifying the table density
77	>	* to be used in calculating the amount of space to allocate for the
78	>	* given number of elements.  Also, for compatibility with previous
79	>	* versions of this class, constructors may optionally specify an
80	>	* expected {@code concurrencyLevel} as an additional hint for
81	>	* internal sizing.  Note that using many keys with exactly the same
82	>	* {@code hashCode()} is a sure way to slow down performance of any
83	>	* hash table.
84	>	*
85	>	* <p>A {@link Set} projection of a ConcurrentHashMap may be created
86	>	* (using {@link #newKeySet()} or {@link #newKeySet(int)}), or viewed
87	>	* (using {@link #keySet(Object)} when only keys are of interest, and the
88	>	* mapped values are (perhaps transiently) not used or all take the
89	>	* same mapping value.
90	>	*
91	>	* <p>A ConcurrentHashMap can be used as scalable frequency map (a
92	>	* form of histogram or multiset) by using {@link LongAdder} values
93	>	* and initializing via {@link #computeIfAbsent}. For example, to add
94	>	* a count to a {@code ConcurrentHashMap<String,LongAdder> freqs}, you
95	>	* can use {@code freqs.computeIfAbsent(k -> new
96	>	* LongAdder()).increment();}
97		*
98		* <p>This class and its views and iterators implement all of the
99		* <em>optional</em> methods of the {@link Map} and {@link Iterator}
100		* interfaces.
101		*
102	<	* <p> Like {@link Hashtable} but unlike {@link HashMap}, this class
103	<	* does <em>not</em> allow <tt>null</tt> to be used as a key or value.
102	>	* <p>Like {@link Hashtable} but unlike {@link HashMap}, this class
103	>	* does <em>not</em> allow {@code null} to be used as a key or value.
104	>	*
105	>	* <p>ConcurrentHashMaps support parallel operations using the {@link
106	>	* ForkJoinPool#commonPool}. (Tasks that may be used in other contexts
107	>	* are available in class {@link ForkJoinTasks}). These operations are
108	>	* designed to be safely, and often sensibly, applied even with maps
109	>	* that are being concurrently updated by other threads; for example,
110	>	* when computing a snapshot summary of the values in a shared
111	>	* registry.  There are three kinds of operation, each with four
112	>	* forms, accepting functions with Keys, Values, Entries, and (Key,
113	>	* Value) arguments and/or return values. (The first three forms are
114	>	* also available via the {@link #keySet()}, {@link #values()} and
115	>	* {@link #entrySet()} views). Because the elements of a
116	>	* ConcurrentHashMap are not ordered in any particular way, and may be
117	>	* processed in different orders in different parallel executions, the
118	>	* correctness of supplied functions should not depend on any
119	>	* ordering, or on any other objects or values that may transiently
120	>	* change while computation is in progress; and except for forEach
121	>	* actions, should ideally be side-effect-free.
122	>	*
123	>	* <ul>
124	>	* <li> forEach: Perform a given action on each element.
125	>	* A variant form applies a given transformation on each element
126	>	* before performing the action.</li>
127	>	*
128	>	* <li> search: Return the first available non-null result of
129	>	* applying a given function on each element; skipping further
130	>	* search when a result is found.</li>
131	>	*
132	>	* <li> reduce: Accumulate each element.  The supplied reduction
133	>	* function cannot rely on ordering (more formally, it should be
134	>	* both associative and commutative).  There are five variants:
135	>	*
136	>	* <ul>
137	>	*
138	>	* <li> Plain reductions. (There is not a form of this method for
139	>	* (key, value) function arguments since there is no corresponding
140	>	* return type.)</li>
141	>	*
142	>	* <li> Mapped reductions that accumulate the results of a given
143	>	* function applied to each element.</li>
144	>	*
145	>	* <li> Reductions to scalar doubles, longs, and ints, using a
146	>	* given basis value.</li>
147	>	*
148	>	* </li>
149	>	* </ul>
150	>	* </ul>
151	>	*
152	>	* <p>The concurrency properties of bulk operations follow
153	>	* from those of ConcurrentHashMap: Any non-null result returned
154	>	* from {@code get(key)} and related access methods bears a
155	>	* happens-before relation with the associated insertion or
156	>	* update.  The result of any bulk operation reflects the
157	>	* composition of these per-element relations (but is not
158	>	* necessarily atomic with respect to the map as a whole unless it
159	>	* is somehow known to be quiescent).  Conversely, because keys
160	>	* and values in the map are never null, null serves as a reliable
161	>	* atomic indicator of the current lack of any result.  To
162	>	* maintain this property, null serves as an implicit basis for
163	>	* all non-scalar reduction operations. For the double, long, and
164	>	* int versions, the basis should be one that, when combined with
165	>	* any other value, returns that other value (more formally, it
166	>	* should be the identity element for the reduction). Most common
167	>	* reductions have these properties; for example, computing a sum
168	>	* with basis 0 or a minimum with basis MAX_VALUE.
169	>	*
170	>	* <p>Search and transformation functions provided as arguments
171	>	* should similarly return null to indicate the lack of any result
172	>	* (in which case it is not used). In the case of mapped
173	>	* reductions, this also enables transformations to serve as
174	>	* filters, returning null (or, in the case of primitive
175	>	* specializations, the identity basis) if the element should not
176	>	* be combined. You can create compound transformations and
177	>	* filterings by composing them yourself under this "null means
178	>	* there is nothing there now" rule before using them in search or
179	>	* reduce operations.
180	>	*
181	>	* <p>Methods accepting and/or returning Entry arguments maintain
182	>	* key-value associations. They may be useful for example when
183	>	* finding the key for the greatest value. Note that "plain" Entry
184	>	* arguments can be supplied using {@code new
185	>	* AbstractMap.SimpleEntry(k,v)}.
186	>	*
187	>	* <p>Bulk operations may complete abruptly, throwing an
188	>	* exception encountered in the application of a supplied
189	>	* function. Bear in mind when handling such exceptions that other
190	>	* concurrently executing functions could also have thrown
191	>	* exceptions, or would have done so if the first exception had
192	>	* not occurred.
193	>	*
194	>	* <p>Parallel speedups for bulk operations compared to sequential
195	>	* processing are common but not guaranteed.  Operations involving
196	>	* brief functions on small maps may execute more slowly than
197	>	* sequential loops if the underlying work to parallelize the
198	>	* computation is more expensive than the computation itself.
199	>	* Similarly, parallelization may not lead to much actual parallelism
200	>	* if all processors are busy performing unrelated tasks.
201	>	*
202	>	* <p>All arguments to all task methods must be non-null.
203	>	*
204	>	* <p><em>jsr166e note: During transition, this class
205	>	* uses nested functional interfaces with different names but the
206	>	* same forms as those expected for JDK8.</em>
207		*
208		* <p>This class is a member of the
209	<	* <a href="{@docRoot}/../guide/collections/index.html">
209	>	* <a href="{@docRoot}/../technotes/guides/collections/index.html">
210		* Java Collections Framework</a>.
211		*
212		* @since 1.5
#	Line 70 | Line 214 | import java.io.ObjectOutputStream;
214		* @param <K> the type of keys maintained by this map
215		* @param <V> the type of mapped values
216		*/
217	<	public class ConcurrentHashMap<K, V> extends AbstractMap<K, V>
218	<	implements ConcurrentMap<K, V>, Serializable {
217	>	public class ConcurrentHashMap<K, V>
218	>	implements ConcurrentMap<K, V>, Serializable {
219		private static final long serialVersionUID = 7249069246763182397L;
220
221	+	/**
222	+	* A partitionable iterator. A Spliterator can be traversed
223	+	* directly, but can also be partitioned (before traversal) by
224	+	* creating another Spliterator that covers a non-overlapping
225	+	* portion of the elements, and so may be amenable to parallel
226	+	* execution.
227	+	*
228	+	* <p>This interface exports a subset of expected JDK8
229	+	* functionality.
230	+	*
231	+	* <p>Sample usage: Here is one (of the several) ways to compute
232	+	* the sum of the values held in a map using the ForkJoin
233	+	* framework. As illustrated here, Spliterators are well suited to
234	+	* designs in which a task repeatedly splits off half its work
235	+	* into forked subtasks until small enough to process directly,
236	+	* and then joins these subtasks. Variants of this style can also
237	+	* be used in completion-based designs.
238	+	*
239	+	* <pre>
240	+	* {@code ConcurrentHashMap<String, Long> m = ...
241	+	* // split as if have 8 * parallelism, for load balance
242	+	* int n = m.size();
243	+	* int p = aForkJoinPool.getParallelism() * 8;
244	+	* int split = (n < p)? n : p;
245	+	* long sum = aForkJoinPool.invoke(new SumValues(m.valueSpliterator(), split, null));
246	+	* // ...
247	+	* static class SumValues extends RecursiveTask<Long> {
248	+	*   final Spliterator<Long> s;
249	+	*   final int split;             // split while > 1
250	+	*   final SumValues nextJoin;    // records forked subtasks to join
251	+	*   SumValues(Spliterator<Long> s, int depth, SumValues nextJoin) {
252	+	*     this.s = s; this.depth = depth; this.nextJoin = nextJoin;
253	+	*   }
254	+	*   public Long compute() {
255	+	*     long sum = 0;
256	+	*     SumValues subtasks = null; // fork subtasks
257	+	*     for (int s = split >>> 1; s > 0; s >>>= 1)
258	+	*       (subtasks = new SumValues(s.split(), s, subtasks)).fork();
259	+	*     while (s.hasNext())        // directly process remaining elements
260	+	*       sum += s.next();
261	+	*     for (SumValues t = subtasks; t != null; t = t.nextJoin)
262	+	*       sum += t.join();         // collect subtask results
263	+	*     return sum;
264	+	*   }
265	+	* }
266	+	* }</pre>
267	+	*/
268	+	public static interface Spliterator<T> extends Iterator<T> {
269	+	/**
270	+	* Returns a Spliterator covering approximately half of the
271	+	* elements, guaranteed not to overlap with those subsequently
272	+	* returned by this Spliterator.  After invoking this method,
273	+	* the current Spliterator will <em>not</em> produce any of
274	+	* the elements of the returned Spliterator, but the two
275	+	* Spliterators together will produce all of the elements that
276	+	* would have been produced by this Spliterator had this
277	+	* method not been called. The exact number of elements
278	+	* produced by the returned Spliterator is not guaranteed, and
279	+	* may be zero (i.e., with {@code hasNext()} reporting {@code
280	+	* false}) if this Spliterator cannot be further split.
281	+	*
282	+	* @return a Spliterator covering approximately half of the
283	+	* elements
284	+	* @throws IllegalStateException if this Spliterator has
285	+	* already commenced traversing elements
286	+	*/
287	+	Spliterator<T> split();
288	+	}
289	+
290		/*
291	<	* The basic strategy is to subdivide the table among Segments,
292	<	* each of which itself is a concurrently readable hash table.
291	>	* Overview:
292	>	*
293	>	* The primary design goal of this hash table is to maintain
294	>	* concurrent readability (typically method get(), but also
295	>	* iterators and related methods) while minimizing update
296	>	* contention. Secondary goals are to keep space consumption about
297	>	* the same or better than java.util.HashMap, and to support high
298	>	* initial insertion rates on an empty table by many threads.
299	>	*
300	>	* Each key-value mapping is held in a Node.  Because Node fields
301	>	* can contain special values, they are defined using plain Object
302	>	* types. Similarly in turn, all internal methods that use them
303	>	* work off Object types. And similarly, so do the internal
304	>	* methods of auxiliary iterator and view classes. This also
305	>	* allows many of the public methods to be factored into a smaller
306	>	* number of internal methods (although sadly not so for the five
307	>	* variants of put-related operations). The validation-based
308	>	* approach explained below leads to a lot of code sprawl because
309	>	* retry-control precludes factoring into smaller methods.
310	>	*
311	>	* The table is lazily initialized to a power-of-two size upon the
312	>	* first insertion.  Each bin in the table normally contains a
313	>	* list of Nodes (most often, the list has only zero or one Node).
314	>	* Table accesses require volatile/atomic reads, writes, and
315	>	* CASes.  Because there is no other way to arrange this without
316	>	* adding further indirections, we use intrinsics
317	>	* (sun.misc.Unsafe) operations.  The lists of nodes within bins
318	>	* are always accurately traversable under volatile reads, so long
319	>	* as lookups check hash code and non-nullness of value before
320	>	* checking key equality.
321	>	*
322	>	* We use the top (sign) bit of Node hash fields for control
323	>	* purposes -- it is available anyway because of addressing
324	>	* constraints.  Nodes with negative hash fields are forwarding
325	>	* nodes to either TreeBins or resized tables.  The lower 31 bits
326	>	* of each normal Node's hash field contain a transformation of
327	>	* the key's hash code.
328	>	*
329	>	* Insertion (via put or its variants) of the first node in an
330	>	* empty bin is performed by just CASing it to the bin.  This is
331	>	* by far the most common case for put operations under most
332	>	* key/hash distributions.  Other update operations (insert,
333	>	* delete, and replace) require locks.  We do not want to waste
334	>	* the space required to associate a distinct lock object with
335	>	* each bin, so instead use the first node of a bin list itself as
336	>	* a lock. Locking support for these locks relies on builtin
337	>	* "synchronized" monitors.
338	>	*
339	>	* Using the first node of a list as a lock does not by itself
340	>	* suffice though: When a node is locked, any update must first
341	>	* validate that it is still the first node after locking it, and
342	>	* retry if not. Because new nodes are always appended to lists,
343	>	* once a node is first in a bin, it remains first until deleted
344	>	* or the bin becomes invalidated (upon resizing).  However,
345	>	* operations that only conditionally update may inspect nodes
346	>	* until the point of update. This is a converse of sorts to the
347	>	* lazy locking technique described by Herlihy & Shavit.
348	>	*
349	>	* The main disadvantage of per-bin locks is that other update
350	>	* operations on other nodes in a bin list protected by the same
351	>	* lock can stall, for example when user equals() or mapping
352	>	* functions take a long time.  However, statistically, under
353	>	* random hash codes, this is not a common problem.  Ideally, the
354	>	* frequency of nodes in bins follows a Poisson distribution
355	>	* (http://en.wikipedia.org/wiki/Poisson_distribution) with a
356	>	* parameter of about 0.5 on average, given the resizing threshold
357	>	* of 0.75, although with a large variance because of resizing
358	>	* granularity. Ignoring variance, the expected occurrences of
359	>	* list size k are (exp(-0.5) * pow(0.5, k) / factorial(k)). The
360	>	* first values are:
361	>	*
362	>	* 0:    0.60653066
363	>	* 1:    0.30326533
364	>	* 2:    0.07581633
365	>	* 3:    0.01263606
366	>	* 4:    0.00157952
367	>	* 5:    0.00015795
368	>	* 6:    0.00001316
369	>	* 7:    0.00000094
370	>	* 8:    0.00000006
371	>	* more: less than 1 in ten million
372	>	*
373	>	* Lock contention probability for two threads accessing distinct
374	>	* elements is roughly 1 / (8 * #elements) under random hashes.
375	>	*
376	>	* Actual hash code distributions encountered in practice
377	>	* sometimes deviate significantly from uniform randomness.  This
378	>	* includes the case when N > (1<<30), so some keys MUST collide.
379	>	* Similarly for dumb or hostile usages in which multiple keys are
380	>	* designed to have identical hash codes. Also, although we guard
381	>	* against the worst effects of this (see method spread), sets of
382	>	* hashes may differ only in bits that do not impact their bin
383	>	* index for a given power-of-two mask.  So we use a secondary
384	>	* strategy that applies when the number of nodes in a bin exceeds
385	>	* a threshold, and at least one of the keys implements
386	>	* Comparable.  These TreeBins use a balanced tree to hold nodes
387	>	* (a specialized form of red-black trees), bounding search time
388	>	* to O(log N).  Each search step in a TreeBin is around twice as
389	>	* slow as in a regular list, but given that N cannot exceed
390	>	* (1<<64) (before running out of addresses) this bounds search
391	>	* steps, lock hold times, etc, to reasonable constants (roughly
392	>	* 100 nodes inspected per operation worst case) so long as keys
393	>	* are Comparable (which is very common -- String, Long, etc).
394	>	* TreeBin nodes (TreeNodes) also maintain the same "next"
395	>	* traversal pointers as regular nodes, so can be traversed in
396	>	* iterators in the same way.
397	>	*
398	>	* The table is resized when occupancy exceeds a percentage
399	>	* threshold (nominally, 0.75, but see below).  Any thread
400	>	* noticing an overfull bin may assist in resizing after the
401	>	* initiating thread allocates and sets up the replacement
402	>	* array. However, rather than stalling, these other threads may
403	>	* proceed with insertions etc.  The use of TreeBins shields us
404	>	* from the worst case effects of overfilling while resizes are in
405	>	* progress.  Resizing proceeds by transferring bins, one by one,
406	>	* from the table to the next table. To enable concurrency, the
407	>	* next table must be (incrementally) prefilled with place-holders
408	>	* serving as reverse forwarders to the old table.  Because we are
409	>	* using power-of-two expansion, the elements from each bin must
410	>	* either stay at same index, or move with a power of two
411	>	* offset. We eliminate unnecessary node creation by catching
412	>	* cases where old nodes can be reused because their next fields
413	>	* won't change.  On average, only about one-sixth of them need
414	>	* cloning when a table doubles. The nodes they replace will be
415	>	* garbage collectable as soon as they are no longer referenced by
416	>	* any reader thread that may be in the midst of concurrently
417	>	* traversing table.  Upon transfer, the old table bin contains
418	>	* only a special forwarding node (with hash field "MOVED") that
419	>	* contains the next table as its key. On encountering a
420	>	* forwarding node, access and update operations restart, using
421	>	* the new table.
422	>	*
423	>	* Each bin transfer requires its bin lock, which can stall
424	>	* waiting for locks while resizing. However, because other
425	>	* threads can join in and help resize rather than contend for
426	>	* locks, average aggregate waits become shorter as resizing
427	>	* progresses.  The transfer operation must also ensure that all
428	>	* accessible bins in both the old and new table are usable by any
429	>	* traversal.  This is arranged by proceeding from the last bin
430	>	* (table.length - 1) up towards the first.  Upon seeing a
431	>	* forwarding node, traversals (see class Traverser) arrange to
432	>	* move to the new table without revisiting nodes.  However, to
433	>	* ensure that no intervening nodes are skipped, bin splitting can
434	>	* only begin after the associated reverse-forwarders are in
435	>	* place.
436	>	*
437	>	* The traversal scheme also applies to partial traversals of
438	>	* ranges of bins (via an alternate Traverser constructor)
439	>	* to support partitioned aggregate operations.  Also, read-only
440	>	* operations give up if ever forwarded to a null table, which
441	>	* provides support for shutdown-style clearing, which is also not
442	>	* currently implemented.
443	>	*
444	>	* Lazy table initialization minimizes footprint until first use,
445	>	* and also avoids resizings when the first operation is from a
446	>	* putAll, constructor with map argument, or deserialization.
447	>	* These cases attempt to override the initial capacity settings,
448	>	* but harmlessly fail to take effect in cases of races.
449	>	*
450	>	* The element count is maintained using a specialization of
451	>	* LongAdder. We need to incorporate a specialization rather than
452	>	* just use a LongAdder in order to access implicit
453	>	* contention-sensing that leads to creation of multiple
454	>	* CounterCells.  The counter mechanics avoid contention on
455	>	* updates but can encounter cache thrashing if read too
456	>	* frequently during concurrent access. To avoid reading so often,
457	>	* resizing under contention is attempted only upon adding to a
458	>	* bin already holding two or more nodes. Under uniform hash
459	>	* distributions, the probability of this occurring at threshold
460	>	* is around 13%, meaning that only about 1 in 8 puts check
461	>	* threshold (and after resizing, many fewer do so). The bulk
462	>	* putAll operation further reduces contention by only committing
463	>	* count updates upon these size checks.
464	>	*
465	>	* Maintaining API and serialization compatibility with previous
466	>	* versions of this class introduces several oddities. Mainly: We
467	>	* leave untouched but unused constructor arguments refering to
468	>	* concurrencyLevel. We accept a loadFactor constructor argument,
469	>	* but apply it only to initial table capacity (which is the only
470	>	* time that we can guarantee to honor it.) We also declare an
471	>	* unused "Segment" class that is instantiated in minimal form
472	>	* only when serializing.
473		*/
474
475		/* ---------------- Constants -------------- */
476
477		/**
478	<	* The default initial capacity for this table,
479	<	* used when not otherwise specified in a constructor.
478	>	* The largest possible table capacity.  This value must be
479	>	* exactly 1<<30 to stay within Java array allocation and indexing
480	>	* bounds for power of two table sizes, and is further required
481	>	* because the top two bits of 32bit hash fields are used for
482	>	* control purposes.
483	>	*/
484	>	private static final int MAXIMUM_CAPACITY = 1 << 30;
485	>
486	>	/**
487	>	* The default initial table capacity.  Must be a power of 2
488	>	* (i.e., at least 1) and at most MAXIMUM_CAPACITY.
489	>	*/
490	>	private static final int DEFAULT_CAPACITY = 16;
491	>
492	>	/**
493	>	* The largest possible (non-power of two) array size.
494	>	* Needed by toArray and related methods.
495	>	*/
496	>	static final int MAX_ARRAY_SIZE = Integer.MAX_VALUE - 8;
497	>
498	>	/**
499	>	* The default concurrency level for this table. Unused but
500	>	* defined for compatibility with previous versions of this class.
501		*/
502	<	static final int DEFAULT_INITIAL_CAPACITY = 16;
502	>	private static final int DEFAULT_CONCURRENCY_LEVEL = 16;
503
504		/**
505	<	* The default load factor for this table, used when not
506	<	* otherwise specified in a constructor.
505	>	* The load factor for this table. Overrides of this value in
506	>	* constructors affect only the initial table capacity.  The
507	>	* actual floating point value isn't normally used -- it is
508	>	* simpler to use expressions such as {@code n - (n >>> 2)} for
509	>	* the associated resizing threshold.
510		*/
511	<	static final float DEFAULT_LOAD_FACTOR = 0.75f;
511	>	private static final float LOAD_FACTOR = 0.75f;
512
513		/**
514	<	* The default concurrency level for this table, used when not
515	<	* otherwise specified in a constructor.
514	>	* The bin count threshold for using a tree rather than list for a
515	>	* bin.  The value reflects the approximate break-even point for
516	>	* using tree-based operations.
517		*/
518	<	static final int DEFAULT_CONCURRENCY_LEVEL = 16;
518	>	private static final int TREE_THRESHOLD = 8;
519
520		/**
521	<	* The maximum capacity, used if a higher value is implicitly
522	<	* specified by either of the constructors with arguments.  MUST
523	<	* be a power of two <= 1<<30 to ensure that entries are indexable
524	<	* using ints.
521	>	* Minimum number of rebinnings per transfer step. Ranges are
522	>	* subdivided to allow multiple resizer threads.  This value
523	>	* serves as a lower bound to avoid resizers encountering
524	>	* excessive memory contention.  The value should be at least
525	>	* DEFAULT_CAPACITY.
526		*/
527	<	static final int MAXIMUM_CAPACITY = 1 << 30;
527	>	private static final int MIN_TRANSFER_STRIDE = 16;
528	>
529	>	/*
530	>	* Encodings for Node hash fields. See above for explanation.
531	>	*/
532	>	static final int MOVED     = 0x80000000; // hash field for forwarding nodes
533	>	static final int HASH_BITS = 0x7fffffff; // usable bits of normal node hash
534	>
535	>	/** Number of CPUS, to place bounds on some sizings */
536	>	static final int NCPU = Runtime.getRuntime().availableProcessors();
537	>
538	>	/* ---------------- Counters -------------- */
539	>
540	>	// Adapted from LongAdder and Striped64.
541	>	// See their internal docs for explanation.
542	>
543	>	// A padded cell for distributing counts
544	>	static final class CounterCell {
545	>	volatile long p0, p1, p2, p3, p4, p5, p6;
546	>	volatile long value;
547	>	volatile long q0, q1, q2, q3, q4, q5, q6;
548	>	CounterCell(long x) { value = x; }
549	>	}
550
551		/**
552	<	* The maximum number of segments to allow; used to bound
553	<	* constructor arguments.
552	>	* Holder for the thread-local hash code determining which
553	>	* CounterCell to use. The code is initialized via the
554	>	* counterHashCodeGenerator, but may be moved upon collisions.
555		*/
556	<	static final int MAX_SEGMENTS = 1 << 16; // slightly conservative
556	>	static final class CounterHashCode {
557	>	int code;
558	>	}
559
560		/**
561	<	* Number of unsynchronized retries in size and containsValue
118	<	* methods before resorting to locking. This is used to avoid
119	<	* unbounded retries if tables undergo continuous modification
120	<	* which would make it impossible to obtain an accurate result.
561	>	* Generates initial value for per-thread CounterHashCodes
562		*/
563	<	static final int RETRIES_BEFORE_LOCK = 2;
563	>	static final AtomicInteger counterHashCodeGenerator = new AtomicInteger();
564	>
565	>	/**
566	>	* Increment for counterHashCodeGenerator. See class ThreadLocal
567	>	* for explanation.
568	>	*/
569	>	static final int SEED_INCREMENT = 0x61c88647;
570	>
571	>	/**
572	>	* Per-thread counter hash codes. Shared across all instances
573	>	*/
574	>	static final ThreadLocal<CounterHashCode> threadCounterHashCode =
575	>	new ThreadLocal<CounterHashCode>();
576
577		/* ---------------- Fields -------------- */
578
579		/**
580	<	* Mask value for indexing into segments. The upper bits of a
581	<	* key's hash code are used to choose the segment.
580	>	* The array of bins. Lazily initialized upon first insertion.
581	>	* Size is always a power of two. Accessed directly by iterators.
582		*/
583	<	final int segmentMask;
583	>	transient volatile Node[] table;
584
585		/**
586	<	* Shift value for indexing within segments.
586	>	* The next table to use; non-null only while resizing.
587		*/
588	<	final int segmentShift;
588	>	private transient volatile Node[] nextTable;
589
590		/**
591	<	* The segments, each of which is a specialized hash table
591	>	* Base counter value, used mainly when there is no contention,
592	>	* but also as a fallback during table initialization
593	>	* races. Updated via CAS.
594		*/
595	<	final Segment<K,V>[] segments;
595	>	private transient volatile long baseCount;
596
597	<	transient Set<K> keySet;
598	<	transient Set<Map.Entry<K,V>> entrySet;
599	<	transient Collection<V> values;
597	>	/**
598	>	* Table initialization and resizing control.  When negative, the
599	>	* table is being initialized or resized: -1 for initialization,
600	>	* else -(1 + the number of active resizing threads).  Otherwise,
601	>	* when table is null, holds the initial table size to use upon
602	>	* creation, or 0 for default. After initialization, holds the
603	>	* next element count value upon which to resize the table.
604	>	*/
605	>	private transient volatile int sizeCtl;
606
607	<	/* ---------------- Small Utilities -------------- */
607	>	/**
608	>	* The next table index (plus one) to split while resizing.
609	>	*/
610	>	private transient volatile int transferIndex;
611
612		/**
613	<	* Returns a hash code for non-null Object x.
150	<	* Uses the same hash code spreader as most other java.util hash tables.
151	<	* @param x the object serving as a key
152	<	* @return the hash code
613	>	* The least available table index to split while resizing.
614		*/
615	<	static int hash(Object x) {
155	<	int h = x.hashCode();
156	<	h += ~(h << 9);
157	<	h ^=  (h >>> 14);
158	<	h +=  (h << 4);
159	<	h ^=  (h >>> 10);
160	<	return h;
161	<	}
615	>	private transient volatile int transferOrigin;
616
617		/**
618	<	* 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
618	>	* Spinlock (locked via CAS) used when resizing and/or creating Cells.
619		*/
620	<	final Segment<K,V> segmentFor(int hash) {
621	<	return segments[(hash >>> segmentShift) & segmentMask];
620	>	private transient volatile int counterBusy;
621	>
622	>	/**
623	>	* Table of counter cells. When non-null, size is a power of 2.
624	>	*/
625	>	private transient volatile CounterCell[] counterCells;
626	>
627	>	// views
628	>	private transient KeySetView<K,V> keySet;
629	>	private transient ValuesView<K,V> values;
630	>	private transient EntrySetView<K,V> entrySet;
631	>
632	>	/** For serialization compatibility. Null unless serialized; see below */
633	>	private Segment<K,V>[] segments;
634	>
635	>	/* ---------------- Table element access -------------- */
636	>
637	>	/*
638	>	* Volatile access methods are used for table elements as well as
639	>	* elements of in-progress next table while resizing.  Uses are
640	>	* null checked by callers, and implicitly bounds-checked, relying
641	>	* on the invariants that tab arrays have non-zero size, and all
642	>	* indices are masked with (tab.length - 1) which is never
643	>	* negative and always less than length. Note that, to be correct
644	>	* wrt arbitrary concurrency errors by users, bounds checks must
645	>	* operate on local variables, which accounts for some odd-looking
646	>	* inline assignments below.
647	>	*/
648	>
649	>	static final Node tabAt(Node[] tab, int i) { // used by Traverser
650	>	return (Node)U.getObjectVolatile(tab, ((long)i << ASHIFT) + ABASE);
651		}
652
653	<	/* ---------------- Inner Classes -------------- */
653	>	private static final boolean casTabAt(Node[] tab, int i, Node c, Node v) {
654	>	return U.compareAndSwapObject(tab, ((long)i << ASHIFT) + ABASE, c, v);
655	>	}
656	>
657	>	private static final void setTabAt(Node[] tab, int i, Node v) {
658	>	U.putObjectVolatile(tab, ((long)i << ASHIFT) + ABASE, v);
659	>	}
660	>
661	>	/* ---------------- Nodes -------------- */
662
663		/**
664	<	* ConcurrentHashMap list entry. Note that this is never exported
665	<	* out as a user-visible Map.Entry.
666	<	*
667	<	* Because the value field is volatile, not final, it is legal wrt
668	<	* the Java Memory Model for an unsynchronized reader to see null
669	<	* instead of initial value when read via a data race.  Although a
670	<	* reordering leading to this is not likely to ever actually
671	<	* occur, the Segment.readValueUnderLock method is used as a
183	<	* backup in case a null (pre-initialized) value is ever seen in
184	<	* an unsynchronized access method.
664	>	* Key-value entry. Note that this is never exported out as a
665	>	* user-visible Map.Entry (see MapEntry below). Nodes with a hash
666	>	* field of MOVED are special, and do not contain user keys or
667	>	* values.  Otherwise, keys are never null, and null val fields
668	>	* indicate that a node is in the process of being deleted or
669	>	* created. For purposes of read-only access, a key may be read
670	>	* before a val, but can only be used after checking val to be
671	>	* non-null.
672		*/
673	<	static final class HashEntry<K,V> {
187	<	final K key;
673	>	static class Node {
674		final int hash;
675	<	volatile V value;
676	<	final HashEntry<K,V> next;
675	>	final Object key;
676	>	volatile Object val;
677	>	volatile Node next;
678
679	<	HashEntry(K key, int hash, HashEntry<K,V> next, V value) {
193	<	this.key = key;
679	>	Node(int hash, Object key, Object val, Node next) {
680		this.hash = hash;
681	+	this.key = key;
682	+	this.val = val;
683		this.next = next;
196	–	this.value = value;
684		}
198	–
199	–	@SuppressWarnings("unchecked")
200	–	static final <K,V> HashEntry<K,V>[] newArray(int i) {
201	–	return new HashEntry[i];
202	–	}
685		}
686
687	+	/* ---------------- TreeBins -------------- */
688	+
689		/**
690	<	* Segments are specialized versions of hash tables.  This
691	<	* subclasses from ReentrantLock opportunistically, just to
692	<	* simplify some locking and avoid separate construction.
693	<	*/
694	<	static final class Segment<K,V> extends ReentrantLock implements Serializable {
695	<	/*
696	<	* Segments maintain a table of entry lists that are ALWAYS
697	<	* kept in a consistent state, so can be read without locking.
698	<	* Next fields of nodes are immutable (final).  All list
699	<	* additions are performed at the front of each bin. This
700	<	* makes it easy to check changes, and also fast to traverse.
701	<	* When nodes would otherwise be changed, new nodes are
702	<	* created to replace them. This works well for hash tables
703	<	* 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	<	*/
690	>	* Nodes for use in TreeBins
691	>	*/
692	>	static final class TreeNode extends Node {
693	>	TreeNode parent;  // red-black tree links
694	>	TreeNode left;
695	>	TreeNode right;
696	>	TreeNode prev;    // needed to unlink next upon deletion
697	>	boolean red;
698	>
699	>	TreeNode(int hash, Object key, Object val, Node next, TreeNode parent) {
700	>	super(hash, key, val, next);
701	>	this.parent = parent;
702	>	}
703	>	}
704
705	+	/**
706	+	* A specialized form of red-black tree for use in bins
707	+	* whose size exceeds a threshold.
708	+	*
709	+	* TreeBins use a special form of comparison for search and
710	+	* related operations (which is the main reason we cannot use
711	+	* existing collections such as TreeMaps). TreeBins contain
712	+	* Comparable elements, but may contain others, as well as
713	+	* elements that are Comparable but not necessarily Comparable<T>
714	+	* for the same T, so we cannot invoke compareTo among them. To
715	+	* handle this, the tree is ordered primarily by hash value, then
716	+	* by getClass().getName() order, and then by Comparator order
717	+	* among elements of the same class.  On lookup at a node, if
718	+	* elements are not comparable or compare as 0, both left and
719	+	* right children may need to be searched in the case of tied hash
720	+	* values. (This corresponds to the full list search that would be
721	+	* necessary if all elements were non-Comparable and had tied
722	+	* hashes.)  The red-black balancing code is updated from
723	+	* pre-jdk-collections
724	+	* (http://gee.cs.oswego.edu/dl/classes/collections/RBCell.java)
725	+	* based in turn on Cormen, Leiserson, and Rivest "Introduction to
726	+	* Algorithms" (CLR).
727	+	*
728	+	* TreeBins also maintain a separate locking discipline than
729	+	* regular bins. Because they are forwarded via special MOVED
730	+	* nodes at bin heads (which can never change once established),
731	+	* we cannot use those nodes as locks. Instead, TreeBin
732	+	* extends AbstractQueuedSynchronizer to support a simple form of
733	+	* read-write lock. For update operations and table validation,
734	+	* the exclusive form of lock behaves in the same way as bin-head
735	+	* locks. However, lookups use shared read-lock mechanics to allow
736	+	* multiple readers in the absence of writers.  Additionally,
737	+	* these lookups do not ever block: While the lock is not
738	+	* available, they proceed along the slow traversal path (via
739	+	* next-pointers) until the lock becomes available or the list is
740	+	* exhausted, whichever comes first. (These cases are not fast,
741	+	* but maximize aggregate expected throughput.)  The AQS mechanics
742	+	* for doing this are straightforward.  The lock state is held as
743	+	* AQS getState().  Read counts are negative; the write count (1)
744	+	* is positive.  There are no signalling preferences among readers
745	+	* and writers. Since we don't need to export full Lock API, we
746	+	* just override the minimal AQS methods and use them directly.
747	+	*/
748	+	static final class TreeBin extends AbstractQueuedSynchronizer {
749		private static final long serialVersionUID = 2249069246763182397L;
750	+	transient TreeNode root;  // root of tree
751	+	transient TreeNode first; // head of next-pointer list
752
753	<	/**
754	<	* The number of elements in this segment's region.
755	<	*/
756	<	transient volatile int count;
753	>	/* AQS overrides */
754	>	public final boolean isHeldExclusively() { return getState() > 0; }
755	>	public final boolean tryAcquire(int ignore) {
756	>	if (compareAndSetState(0, 1)) {
757	>	setExclusiveOwnerThread(Thread.currentThread());
758	>	return true;
759	>	}
760	>	return false;
761	>	}
762	>	public final boolean tryRelease(int ignore) {
763	>	setExclusiveOwnerThread(null);
764	>	setState(0);
765	>	return true;
766	>	}
767	>	public final int tryAcquireShared(int ignore) {
768	>	for (int c;;) {
769	>	if ((c = getState()) > 0)
770	>	return -1;
771	>	if (compareAndSetState(c, c -1))
772	>	return 1;
773	>	}
774	>	}
775	>	public final boolean tryReleaseShared(int ignore) {
776	>	int c;
777	>	do {} while (!compareAndSetState(c = getState(), c + 1));
778	>	return c == -1;
779	>	}
780	>
781	>	/** From CLR */
782	>	private void rotateLeft(TreeNode p) {
783	>	if (p != null) {
784	>	TreeNode r = p.right, pp, rl;
785	>	if ((rl = p.right = r.left) != null)
786	>	rl.parent = p;
787	>	if ((pp = r.parent = p.parent) == null)
788	>	root = r;
789	>	else if (pp.left == p)
790	>	pp.left = r;
791	>	else
792	>	pp.right = r;
793	>	r.left = p;
794	>	p.parent = r;
795	>	}
796	>	}
797
798	<	/**
799	<	* Number of updates that alter the size of the table. This is
800	<	* used during bulk-read methods to make sure they see a
801	<	* consistent snapshot: If modCounts change during a traversal
802	<	* of segments computing size or checking containsValue, then
803	<	* we might have an inconsistent view of state so (usually)
804	<	* must retry.
805	<	*/
806	<	transient int modCount;
798	>	/** From CLR */
799	>	private void rotateRight(TreeNode p) {
800	>	if (p != null) {
801	>	TreeNode l = p.left, pp, lr;
802	>	if ((lr = p.left = l.right) != null)
803	>	lr.parent = p;
804	>	if ((pp = l.parent = p.parent) == null)
805	>	root = l;
806	>	else if (pp.right == p)
807	>	pp.right = l;
808	>	else
809	>	pp.left = l;
810	>	l.right = p;
811	>	p.parent = l;
812	>	}
813	>	}
814
815		/**
816	<	* The table is rehashed when its size exceeds this threshold.
817	<	* (The value of this field is always <tt>(int)(capacity *
268	<	* loadFactor)</tt>.)
816	>	* Returns the TreeNode (or null if not found) for the given key
817	>	* starting at given root.
818		*/
819	<	transient int threshold;
819	>	@SuppressWarnings("unchecked") final TreeNode getTreeNode
820	>	(int h, Object k, TreeNode p) {
821	>	Class<?> c = k.getClass();
822	>	while (p != null) {
823	>	int dir, ph;  Object pk; Class<?> pc;
824	>	if ((ph = p.hash) == h) {
825	>	if ((pk = p.key) == k || k.equals(pk))
826	>	return p;
827	>	if (c != (pc = pk.getClass()) ||
828	>	!(k instanceof Comparable) ||
829	>	(dir = ((Comparable)k).compareTo((Comparable)pk)) == 0) {
830	>	if ((dir = (c == pc) ? 0 :
831	>	c.getName().compareTo(pc.getName())) == 0) {
832	>	TreeNode r = null, pl, pr; // check both sides
833	>	if ((pr = p.right) != null && h >= pr.hash &&
834	>	(r = getTreeNode(h, k, pr)) != null)
835	>	return r;
836	>	else if ((pl = p.left) != null && h <= pl.hash)
837	>	dir = -1;
838	>	else // nothing there
839	>	return null;
840	>	}
841	>	}
842	>	}
843	>	else
844	>	dir = (h < ph) ? -1 : 1;
845	>	p = (dir > 0) ? p.right : p.left;
846	>	}
847	>	return null;
848	>	}
849
850		/**
851	<	* The per-segment table.
851	>	* Wrapper for getTreeNode used by CHM.get. Tries to obtain
852	>	* read-lock to call getTreeNode, but during failure to get
853	>	* lock, searches along next links.
854		*/
855	<	transient volatile HashEntry<K,V>[] table;
855	>	final Object getValue(int h, Object k) {
856	>	Node r = null;
857	>	int c = getState(); // Must read lock state first
858	>	for (Node e = first; e != null; e = e.next) {
859	>	if (c <= 0 && compareAndSetState(c, c - 1)) {
860	>	try {
861	>	r = getTreeNode(h, k, root);
862	>	} finally {
863	>	releaseShared(0);
864	>	}
865	>	break;
866	>	}
867	>	else if (e.hash == h && k.equals(e.key)) {
868	>	r = e;
869	>	break;
870	>	}
871	>	else
872	>	c = getState();
873	>	}
874	>	return r == null ? null : r.val;
875	>	}
876
877		/**
878	<	* The load factor for the hash table.  Even though this value
879	<	* is same for all segments, it is replicated to avoid needing
280	<	* links to outer object.
281	<	* @serial
878	>	* Finds or adds a node.
879	>	* @return null if added
880		*/
881	<	final float loadFactor;
882	<
883	<	Segment(int initialCapacity, float lf) {
884	<	loadFactor = lf;
885	<	setTable(HashEntry.<K,V>newArray(initialCapacity));
886	<	}
881	>	@SuppressWarnings("unchecked") final TreeNode putTreeNode
882	>	(int h, Object k, Object v) {
883	>	Class<?> c = k.getClass();
884	>	TreeNode pp = root, p = null;
885	>	int dir = 0;
886	>	while (pp != null) { // find existing node or leaf to insert at
887	>	int ph;  Object pk; Class<?> pc;
888	>	p = pp;
889	>	if ((ph = p.hash) == h) {
890	>	if ((pk = p.key) == k || k.equals(pk))
891	>	return p;
892	>	if (c != (pc = pk.getClass()) ||
893	>	!(k instanceof Comparable) ||
894	>	(dir = ((Comparable)k).compareTo((Comparable)pk)) == 0) {
895	>	TreeNode s = null, r = null, pr;
896	>	if ((dir = (c == pc) ? 0 :
897	>	c.getName().compareTo(pc.getName())) == 0) {
898	>	if ((pr = p.right) != null && h >= pr.hash &&
899	>	(r = getTreeNode(h, k, pr)) != null)
900	>	return r;
901	>	else // continue left
902	>	dir = -1;
903	>	}
904	>	else if ((pr = p.right) != null && h >= pr.hash)
905	>	s = pr;
906	>	if (s != null && (r = getTreeNode(h, k, s)) != null)
907	>	return r;
908	>	}
909	>	}
910	>	else
911	>	dir = (h < ph) ? -1 : 1;
912	>	pp = (dir > 0) ? p.right : p.left;
913	>	}
914
915	<	@SuppressWarnings("unchecked")
916	<	static final <K,V> Segment<K,V>[] newArray(int i) {
917	<	return new Segment[i];
915	>	TreeNode f = first;
916	>	TreeNode x = first = new TreeNode(h, k, v, f, p);
917	>	if (p == null)
918	>	root = x;
919	>	else { // attach and rebalance; adapted from CLR
920	>	TreeNode xp, xpp;
921	>	if (f != null)
922	>	f.prev = x;
923	>	if (dir <= 0)
924	>	p.left = x;
925	>	else
926	>	p.right = x;
927	>	x.red = true;
928	>	while (x != null && (xp = x.parent) != null && xp.red &&
929	>	(xpp = xp.parent) != null) {
930	>	TreeNode xppl = xpp.left;
931	>	if (xp == xppl) {
932	>	TreeNode y = xpp.right;
933	>	if (y != null && y.red) {
934	>	y.red = false;
935	>	xp.red = false;
936	>	xpp.red = true;
937	>	x = xpp;
938	>	}
939	>	else {
940	>	if (x == xp.right) {
941	>	rotateLeft(x = xp);
942	>	xpp = (xp = x.parent) == null ? null : xp.parent;
943	>	}
944	>	if (xp != null) {
945	>	xp.red = false;
946	>	if (xpp != null) {
947	>	xpp.red = true;
948	>	rotateRight(xpp);
949	>	}
950	>	}
951	>	}
952	>	}
953	>	else {
954	>	TreeNode y = xppl;
955	>	if (y != null && y.red) {
956	>	y.red = false;
957	>	xp.red = false;
958	>	xpp.red = true;
959	>	x = xpp;
960	>	}
961	>	else {
962	>	if (x == xp.left) {
963	>	rotateRight(x = xp);
964	>	xpp = (xp = x.parent) == null ? null : xp.parent;
965	>	}
966	>	if (xp != null) {
967	>	xp.red = false;
968	>	if (xpp != null) {
969	>	xpp.red = true;
970	>	rotateLeft(xpp);
971	>	}
972	>	}
973	>	}
974	>	}
975	>	}
976	>	TreeNode r = root;
977	>	if (r != null && r.red)
978	>	r.red = false;
979	>	}
980	>	return null;
981		}
982
983		/**
984	<	* Sets table to new HashEntry array.
985	<	* Call only while holding lock or in constructor.
984	>	* Removes the given node, that must be present before this
985	>	* call.  This is messier than typical red-black deletion code
986	>	* because we cannot swap the contents of an interior node
987	>	* with a leaf successor that is pinned by "next" pointers
988	>	* that are accessible independently of lock. So instead we
989	>	* swap the tree linkages.
990		*/
991	<	void setTable(HashEntry<K,V>[] newTable) {
992	<	threshold = (int)(newTable.length * loadFactor);
993	<	table = newTable;
991	>	final void deleteTreeNode(TreeNode p) {
992	>	TreeNode next = (TreeNode)p.next; // unlink traversal pointers
993	>	TreeNode pred = p.prev;
994	>	if (pred == null)
995	>	first = next;
996	>	else
997	>	pred.next = next;
998	>	if (next != null)
999	>	next.prev = pred;
1000	>	TreeNode replacement;
1001	>	TreeNode pl = p.left;
1002	>	TreeNode pr = p.right;
1003	>	if (pl != null && pr != null) {
1004	>	TreeNode s = pr, sl;
1005	>	while ((sl = s.left) != null) // find successor
1006	>	s = sl;
1007	>	boolean c = s.red; s.red = p.red; p.red = c; // swap colors
1008	>	TreeNode sr = s.right;
1009	>	TreeNode pp = p.parent;
1010	>	if (s == pr) { // p was s's direct parent
1011	>	p.parent = s;
1012	>	s.right = p;
1013	>	}
1014	>	else {
1015	>	TreeNode sp = s.parent;
1016	>	if ((p.parent = sp) != null) {
1017	>	if (s == sp.left)
1018	>	sp.left = p;
1019	>	else
1020	>	sp.right = p;
1021	>	}
1022	>	if ((s.right = pr) != null)
1023	>	pr.parent = s;
1024	>	}
1025	>	p.left = null;
1026	>	if ((p.right = sr) != null)
1027	>	sr.parent = p;
1028	>	if ((s.left = pl) != null)
1029	>	pl.parent = s;
1030	>	if ((s.parent = pp) == null)
1031	>	root = s;
1032	>	else if (p == pp.left)
1033	>	pp.left = s;
1034	>	else
1035	>	pp.right = s;
1036	>	replacement = sr;
1037	>	}
1038	>	else
1039	>	replacement = (pl != null) ? pl : pr;
1040	>	TreeNode pp = p.parent;
1041	>	if (replacement == null) {
1042	>	if (pp == null) {
1043	>	root = null;
1044	>	return;
1045	>	}
1046	>	replacement = p;
1047	>	}
1048	>	else {
1049	>	replacement.parent = pp;
1050	>	if (pp == null)
1051	>	root = replacement;
1052	>	else if (p == pp.left)
1053	>	pp.left = replacement;
1054	>	else
1055	>	pp.right = replacement;
1056	>	p.left = p.right = p.parent = null;
1057	>	}
1058	>	if (!p.red) { // rebalance, from CLR
1059	>	TreeNode x = replacement;
1060	>	while (x != null) {
1061	>	TreeNode xp, xpl;
1062	>	if (x.red || (xp = x.parent) == null) {
1063	>	x.red = false;
1064	>	break;
1065	>	}
1066	>	if (x == (xpl = xp.left)) {
1067	>	TreeNode sib = xp.right;
1068	>	if (sib != null && sib.red) {
1069	>	sib.red = false;
1070	>	xp.red = true;
1071	>	rotateLeft(xp);
1072	>	sib = (xp = x.parent) == null ? null : xp.right;
1073	>	}
1074	>	if (sib == null)
1075	>	x = xp;
1076	>	else {
1077	>	TreeNode sl = sib.left, sr = sib.right;
1078	>	if ((sr == null || !sr.red) &&
1079	>	(sl == null || !sl.red)) {
1080	>	sib.red = true;
1081	>	x = xp;
1082	>	}
1083	>	else {
1084	>	if (sr == null || !sr.red) {
1085	>	if (sl != null)
1086	>	sl.red = false;
1087	>	sib.red = true;
1088	>	rotateRight(sib);
1089	>	sib = (xp = x.parent) == null ?
1090	>	null : xp.right;
1091	>	}
1092	>	if (sib != null) {
1093	>	sib.red = (xp == null) ? false : xp.red;
1094	>	if ((sr = sib.right) != null)
1095	>	sr.red = false;
1096	>	}
1097	>	if (xp != null) {
1098	>	xp.red = false;
1099	>	rotateLeft(xp);
1100	>	}
1101	>	x = root;
1102	>	}
1103	>	}
1104	>	}
1105	>	else { // symmetric
1106	>	TreeNode sib = xpl;
1107	>	if (sib != null && sib.red) {
1108	>	sib.red = false;
1109	>	xp.red = true;
1110	>	rotateRight(xp);
1111	>	sib = (xp = x.parent) == null ? null : xp.left;
1112	>	}
1113	>	if (sib == null)
1114	>	x = xp;
1115	>	else {
1116	>	TreeNode sl = sib.left, sr = sib.right;
1117	>	if ((sl == null || !sl.red) &&
1118	>	(sr == null || !sr.red)) {
1119	>	sib.red = true;
1120	>	x = xp;
1121	>	}
1122	>	else {
1123	>	if (sl == null || !sl.red) {
1124	>	if (sr != null)
1125	>	sr.red = false;
1126	>	sib.red = true;
1127	>	rotateLeft(sib);
1128	>	sib = (xp = x.parent) == null ?
1129	>	null : xp.left;
1130	>	}
1131	>	if (sib != null) {
1132	>	sib.red = (xp == null) ? false : xp.red;
1133	>	if ((sl = sib.left) != null)
1134	>	sl.red = false;
1135	>	}
1136	>	if (xp != null) {
1137	>	xp.red = false;
1138	>	rotateRight(xp);
1139	>	}
1140	>	x = root;
1141	>	}
1142	>	}
1143	>	}
1144	>	}
1145	>	}
1146	>	if (p == replacement && (pp = p.parent) != null) {
1147	>	if (p == pp.left) // detach pointers
1148	>	pp.left = null;
1149	>	else if (p == pp.right)
1150	>	pp.right = null;
1151	>	p.parent = null;
1152	>	}
1153		}
1154	+	}
1155
1156	<	/**
1157	<	* Returns properly casted first entry of bin for given hash.
1158	<	*/
1159	<	HashEntry<K,V> getFirst(int hash) {
1160	<	HashEntry<K,V>[] tab = table;
1161	<	return tab[hash & (tab.length - 1)];
1156	>	/* ---------------- Collision reduction methods -------------- */
1157	>
1158	>	/**
1159	>	* Spreads higher bits to lower, and also forces top bit to 0.
1160	>	* Because the table uses power-of-two masking, sets of hashes
1161	>	* that vary only in bits above the current mask will always
1162	>	* collide. (Among known examples are sets of Float keys holding
1163	>	* consecutive whole numbers in small tables.)  To counter this,
1164	>	* we apply a transform that spreads the impact of higher bits
1165	>	* downward. There is a tradeoff between speed, utility, and
1166	>	* quality of bit-spreading. Because many common sets of hashes
1167	>	* are already reasonably distributed across bits (so don't benefit
1168	>	* from spreading), and because we use trees to handle large sets
1169	>	* of collisions in bins, we don't need excessively high quality.
1170	>	*/
1171	>	private static final int spread(int h) {
1172	>	h ^= (h >>> 18) ^ (h >>> 12);
1173	>	return (h ^ (h >>> 10)) & HASH_BITS;
1174	>	}
1175	>
1176	>	/**
1177	>	* Replaces a list bin with a tree bin if key is comparable.  Call
1178	>	* only when locked.
1179	>	*/
1180	>	private final void replaceWithTreeBin(Node[] tab, int index, Object key) {
1181	>	if (key instanceof Comparable) {
1182	>	TreeBin t = new TreeBin();
1183	>	for (Node e = tabAt(tab, index); e != null; e = e.next)
1184	>	t.putTreeNode(e.hash, e.key, e.val);
1185	>	setTabAt(tab, index, new Node(MOVED, t, null, null));
1186		}
1187	+	}
1188
1189	<	/**
1190	<	* Reads value field of an entry under lock. Called if value
1191	<	* field ever appears to be null. This is possible only if a
1192	<	* compiler happens to reorder a HashEntry initialization with
1193	<	* its table assignment, which is legal under memory model
1194	<	* but is not known to ever occur.
1195	<	*/
1196	<	V readValueUnderLock(HashEntry<K,V> e) {
1197	<	lock();
1198	<	try {
1199	<	return e.value;
1200	<	} finally {
1201	<	unlock();
1189	>	/* ---------------- Internal access and update methods -------------- */
1190	>
1191	>	/** Implementation for get and containsKey */
1192	>	@SuppressWarnings("unchecked") private final V internalGet(Object k) {
1193	>	int h = spread(k.hashCode());
1194	>	retry: for (Node[] tab = table; tab != null;) {
1195	>	Node e; Object ek, ev; int eh;      // locals to read fields once
1196	>	for (e = tabAt(tab, (tab.length - 1) & h); e != null; e = e.next) {
1197	>	if ((eh = e.hash) < 0) {
1198	>	if ((ek = e.key) instanceof TreeBin)  // search TreeBin
1199	>	return (V)((TreeBin)ek).getValue(h, k);
1200	>	else {                        // restart with new table
1201	>	tab = (Node[])ek;
1202	>	continue retry;
1203	>	}
1204	>	}
1205	>	else if (eh == h && (ev = e.val) != null &&
1206	>	((ek = e.key) == k || k.equals(ek)))
1207	>	return (V)ev;
1208		}
1209	+	break;
1210		}
1211	+	return null;
1212	+	}
1213
1214	<	/* Specialized implementations of map methods */
1215	<
1216	<	V get(Object key, int hash) {
1217	<	if (count != 0) { // read-volatile
1218	<	HashEntry<K,V> e = getFirst(hash);
1219	<	while (e != null) {
1220	<	if (e.hash == hash && key.equals(e.key)) {
1221	<	V v = e.value;
1222	<	if (v != null)
1223	<	return v;
1224	<	return readValueUnderLock(e); // recheck
1214	>	/**
1215	>	* Implementation for the four public remove/replace methods:
1216	>	* Replaces node value with v, conditional upon match of cv if
1217	>	* non-null.  If resulting value is null, delete.
1218	>	*/
1219	>	@SuppressWarnings("unchecked") private final V internalReplace
1220	>	(Object k, V v, Object cv) {
1221	>	int h = spread(k.hashCode());
1222	>	Object oldVal = null;
1223	>	for (Node[] tab = table;;) {
1224	>	Node f; int i, fh; Object fk;
1225	>	if (tab == null ||
1226	>	(f = tabAt(tab, i = (tab.length - 1) & h)) == null)
1227	>	break;
1228	>	else if ((fh = f.hash) < 0) {
1229	>	if ((fk = f.key) instanceof TreeBin) {
1230	>	TreeBin t = (TreeBin)fk;
1231	>	boolean validated = false;
1232	>	boolean deleted = false;
1233	>	t.acquire(0);
1234	>	try {
1235	>	if (tabAt(tab, i) == f) {
1236	>	validated = true;
1237	>	TreeNode p = t.getTreeNode(h, k, t.root);
1238	>	if (p != null) {
1239	>	Object pv = p.val;
1240	>	if (cv == null || cv == pv || cv.equals(pv)) {
1241	>	oldVal = pv;
1242	>	if ((p.val = v) == null) {
1243	>	deleted = true;
1244	>	t.deleteTreeNode(p);
1245	>	}
1246	>	}
1247	>	}
1248	>	}
1249	>	} finally {
1250	>	t.release(0);
1251	>	}
1252	>	if (validated) {
1253	>	if (deleted)
1254	>	addCount(-1L, -1);
1255	>	break;
1256		}
1257	<	e = e.next;
1257	>	}
1258	>	else
1259	>	tab = (Node[])fk;
1260	>	}
1261	>	else if (fh != h && f.next == null) // precheck
1262	>	break;                          // rules out possible existence
1263	>	else {
1264	>	boolean validated = false;
1265	>	boolean deleted = false;
1266	>	synchronized(f) {
1267	>	if (tabAt(tab, i) == f) {
1268	>	validated = true;
1269	>	for (Node e = f, pred = null;;) {
1270	>	Object ek, ev;
1271	>	if (e.hash == h &&
1272	>	((ev = e.val) != null) &&
1273	>	((ek = e.key) == k || k.equals(ek))) {
1274	>	if (cv == null || cv == ev || cv.equals(ev)) {
1275	>	oldVal = ev;
1276	>	if ((e.val = v) == null) {
1277	>	deleted = true;
1278	>	Node en = e.next;
1279	>	if (pred != null)
1280	>	pred.next = en;
1281	>	else
1282	>	setTabAt(tab, i, en);
1283	>	}
1284	>	}
1285	>	break;
1286	>	}
1287	>	pred = e;
1288	>	if ((e = e.next) == null)
1289	>	break;
1290	>	}
1291	>	}
1292	>	}
1293	>	if (validated) {
1294	>	if (deleted)
1295	>	addCount(-1L, -1);
1296	>	break;
1297		}
1298		}
343	–	return null;
1299		}
1300	+	return (V)oldVal;
1301	+	}
1302
1303	<	boolean containsKey(Object key, int hash) {
1304	<	if (count != 0) { // read-volatile
1305	<	HashEntry<K,V> e = getFirst(hash);
1306	<	while (e != null) {
1307	<	if (e.hash == hash && key.equals(e.key))
1308	<	return true;
1309	<	e = e.next;
1303	>	/*
1304	>	* Internal versions of insertion methods
1305	>	* All have the same basic structure as the first (internalPut):
1306	>	*  1. If table uninitialized, create
1307	>	*  2. If bin empty, try to CAS new node
1308	>	*  3. If bin stale, use new table
1309	>	*  4. if bin converted to TreeBin, validate and relay to TreeBin methods
1310	>	*  5. Lock and validate; if valid, scan and add or update
1311	>	*
1312	>	* The putAll method differs mainly in attempting to pre-allocate
1313	>	* enough table space, and also more lazily performs count updates
1314	>	* and checks.
1315	>	*
1316	>	* Most of the function-accepting methods can't be factored nicely
1317	>	* because they require different functional forms, so instead
1318	>	* sprawl out similar mechanics.
1319	>	*/
1320	>
1321	>	/** Implementation for put and putIfAbsent */
1322	>	@SuppressWarnings("unchecked") private final V internalPut
1323	>	(K k, V v, boolean onlyIfAbsent) {
1324	>	if (k == null || v == null) throw new NullPointerException();
1325	>	int h = spread(k.hashCode());
1326	>	int len = 0;
1327	>	for (Node[] tab = table;;) {
1328	>	int i, fh; Node f; Object fk, fv;
1329	>	if (tab == null)
1330	>	tab = initTable();
1331	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1332	>	if (casTabAt(tab, i, null, new Node(h, k, v, null)))
1333	>	break;                   // no lock when adding to empty bin
1334	>	}
1335	>	else if ((fh = f.hash) < 0) {
1336	>	if ((fk = f.key) instanceof TreeBin) {
1337	>	TreeBin t = (TreeBin)fk;
1338	>	Object oldVal = null;
1339	>	t.acquire(0);
1340	>	try {
1341	>	if (tabAt(tab, i) == f) {
1342	>	len = 2;
1343	>	TreeNode p = t.putTreeNode(h, k, v);
1344	>	if (p != null) {
1345	>	oldVal = p.val;
1346	>	if (!onlyIfAbsent)
1347	>	p.val = v;
1348	>	}
1349	>	}
1350	>	} finally {
1351	>	t.release(0);
1352	>	}
1353	>	if (len != 0) {
1354	>	if (oldVal != null)
1355	>	return (V)oldVal;
1356	>	break;
1357	>	}
1358	>	}
1359	>	else
1360	>	tab = (Node[])fk;
1361	>	}
1362	>	else if (onlyIfAbsent && fh == h && (fv = f.val) != null &&
1363	>	((fk = f.key) == k || k.equals(fk))) // peek while nearby
1364	>	return (V)fv;
1365	>	else {
1366	>	Object oldVal = null;
1367	>	synchronized(f) {
1368	>	if (tabAt(tab, i) == f) {
1369	>	len = 1;
1370	>	for (Node e = f;; ++len) {
1371	>	Object ek, ev;
1372	>	if (e.hash == h &&
1373	>	(ev = e.val) != null &&
1374	>	((ek = e.key) == k || k.equals(ek))) {
1375	>	oldVal = ev;
1376	>	if (!onlyIfAbsent)
1377	>	e.val = v;
1378	>	break;
1379	>	}
1380	>	Node last = e;
1381	>	if ((e = e.next) == null) {
1382	>	last.next = new Node(h, k, v, null);
1383	>	if (len >= TREE_THRESHOLD)
1384	>	replaceWithTreeBin(tab, i, k);
1385	>	break;
1386	>	}
1387	>	}
1388	>	}
1389	>	}
1390	>	if (len != 0) {
1391	>	if (oldVal != null)
1392	>	return (V)oldVal;
1393	>	break;
1394		}
1395		}
355	–	return false;
1396		}
1397	+	addCount(1L, len);
1398	+	return null;
1399	+	}
1400
1401	<	boolean containsValue(Object value) {
1402	<	if (count != 0) { // read-volatile
1403	<	HashEntry<K,V>[] tab = table;
1404	<	int len = tab.length;
1405	<	for (int i = 0 ; i < len; i++) {
1406	<	for (HashEntry<K,V> e = tab[i]; e != null; e = e.next) {
1407	<	V v = e.value;
1408	<	if (v == null) // recheck
1409	<	v = readValueUnderLock(e);
1410	<	if (value.equals(v))
1411	<	return true;
1401	>	/** Implementation for computeIfAbsent */
1402	>	@SuppressWarnings("unchecked") private final V internalComputeIfAbsent
1403	>	(K k, Fun<? super K, ?> mf) {
1404	>	if (k == null || mf == null)
1405	>	throw new NullPointerException();
1406	>	int h = spread(k.hashCode());
1407	>	Object val = null;
1408	>	int len = 0;
1409	>	for (Node[] tab = table;;) {
1410	>	Node f; int i; Object fk;
1411	>	if (tab == null)
1412	>	tab = initTable();
1413	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1414	>	Node node = new Node(h, k, null, null);
1415	>	synchronized(node) {
1416	>	if (casTabAt(tab, i, null, node)) {
1417	>	len = 1;
1418	>	try {
1419	>	if ((val = mf.apply(k)) != null)
1420	>	node.val = val;
1421	>	} finally {
1422	>	if (val == null)
1423	>	setTabAt(tab, i, null);
1424	>	}
1425		}
1426		}
1427	+	if (len != 0)
1428	+	break;
1429	+	}
1430	+	else if (f.hash < 0) {
1431	+	if ((fk = f.key) instanceof TreeBin) {
1432	+	TreeBin t = (TreeBin)fk;
1433	+	boolean added = false;
1434	+	t.acquire(0);
1435	+	try {
1436	+	if (tabAt(tab, i) == f) {
1437	+	len = 1;
1438	+	TreeNode p = t.getTreeNode(h, k, t.root);
1439	+	if (p != null)
1440	+	val = p.val;
1441	+	else if ((val = mf.apply(k)) != null) {
1442	+	added = true;
1443	+	len = 2;
1444	+	t.putTreeNode(h, k, val);
1445	+	}
1446	+	}
1447	+	} finally {
1448	+	t.release(0);
1449	+	}
1450	+	if (len != 0) {
1451	+	if (!added)
1452	+	return (V)val;
1453	+	break;
1454	+	}
1455	+	}
1456	+	else
1457	+	tab = (Node[])fk;
1458	+	}
1459	+	else {
1460	+	for (Node e = f; e != null; e = e.next) { // prescan
1461	+	Object ek, ev;
1462	+	if (e.hash == h && (ev = e.val) != null &&
1463	+	((ek = e.key) == k || k.equals(ek)))
1464	+	return (V)ev;
1465	+	}
1466	+	boolean added = false;
1467	+	synchronized(f) {
1468	+	if (tabAt(tab, i) == f) {
1469	+	len = 1;
1470	+	for (Node e = f;; ++len) {
1471	+	Object ek, ev;
1472	+	if (e.hash == h &&
1473	+	(ev = e.val) != null &&
1474	+	((ek = e.key) == k || k.equals(ek))) {
1475	+	val = ev;
1476	+	break;
1477	+	}
1478	+	Node last = e;
1479	+	if ((e = e.next) == null) {
1480	+	if ((val = mf.apply(k)) != null) {
1481	+	added = true;
1482	+	last.next = new Node(h, k, val, null);
1483	+	if (len >= TREE_THRESHOLD)
1484	+	replaceWithTreeBin(tab, i, k);
1485	+	}
1486	+	break;
1487	+	}
1488	+	}
1489	+	}
1490	+	}
1491	+	if (len != 0) {
1492	+	if (!added)
1493	+	return (V)val;
1494	+	break;
1495	+	}
1496		}
372	–	return false;
1497		}
1498	+	if (val != null)
1499	+	addCount(1L, len);
1500	+	return (V)val;
1501	+	}
1502
1503	<	boolean replace(K key, int hash, V oldValue, V newValue) {
1504	<	lock();
1505	<	try {
1506	<	HashEntry<K,V> e = getFirst(hash);
1507	<	while (e != null && (e.hash != hash || !key.equals(e.key)))
1508	<	e = e.next;
1503	>	/** Implementation for compute */
1504	>	@SuppressWarnings("unchecked") private final V internalCompute
1505	>	(K k, boolean onlyIfPresent,
1506	>	BiFun<? super K, ? super V, ? extends V> mf) {
1507	>	if (k == null || mf == null)
1508	>	throw new NullPointerException();
1509	>	int h = spread(k.hashCode());
1510	>	Object val = null;
1511	>	int delta = 0;
1512	>	int len = 0;
1513	>	for (Node[] tab = table;;) {
1514	>	Node f; int i, fh; Object fk;
1515	>	if (tab == null)
1516	>	tab = initTable();
1517	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1518	>	if (onlyIfPresent)
1519	>	break;
1520	>	Node node = new Node(h, k, null, null);
1521	>	synchronized(node) {
1522	>	if (casTabAt(tab, i, null, node)) {
1523	>	try {
1524	>	len = 1;
1525	>	if ((val = mf.apply(k, null)) != null) {
1526	>	node.val = val;
1527	>	delta = 1;
1528	>	}
1529	>	} finally {
1530	>	if (delta == 0)
1531	>	setTabAt(tab, i, null);
1532	>	}
1533	>	}
1534	>	}
1535	>	if (len != 0)
1536	>	break;
1537	>	}
1538	>	else if ((fh = f.hash) < 0) {
1539	>	if ((fk = f.key) instanceof TreeBin) {
1540	>	TreeBin t = (TreeBin)fk;
1541	>	t.acquire(0);
1542	>	try {
1543	>	if (tabAt(tab, i) == f) {
1544	>	len = 1;
1545	>	TreeNode p = t.getTreeNode(h, k, t.root);
1546	>	if (p == null && onlyIfPresent)
1547	>	break;
1548	>	Object pv = (p == null) ? null : p.val;
1549	>	if ((val = mf.apply(k, (V)pv)) != null) {
1550	>	if (p != null)
1551	>	p.val = val;
1552	>	else {
1553	>	len = 2;
1554	>	delta = 1;
1555	>	t.putTreeNode(h, k, val);
1556	>	}
1557	>	}
1558	>	else if (p != null) {
1559	>	delta = -1;
1560	>	t.deleteTreeNode(p);
1561	>	}
1562	>	}
1563	>	} finally {
1564	>	t.release(0);
1565	>	}
1566	>	if (len != 0)
1567	>	break;
1568	>	}
1569	>	else
1570	>	tab = (Node[])fk;
1571	>	}
1572	>	else {
1573	>	synchronized(f) {
1574	>	if (tabAt(tab, i) == f) {
1575	>	len = 1;
1576	>	for (Node e = f, pred = null;; ++len) {
1577	>	Object ek, ev;
1578	>	if (e.hash == h &&
1579	>	(ev = e.val) != null &&
1580	>	((ek = e.key) == k || k.equals(ek))) {
1581	>	val = mf.apply(k, (V)ev);
1582	>	if (val != null)
1583	>	e.val = val;
1584	>	else {
1585	>	delta = -1;
1586	>	Node en = e.next;
1587	>	if (pred != null)
1588	>	pred.next = en;
1589	>	else
1590	>	setTabAt(tab, i, en);
1591	>	}
1592	>	break;
1593	>	}
1594	>	pred = e;
1595	>	if ((e = e.next) == null) {
1596	>	if (!onlyIfPresent &&
1597	>	(val = mf.apply(k, null)) != null) {
1598	>	pred.next = new Node(h, k, val, null);
1599	>	delta = 1;
1600	>	if (len >= TREE_THRESHOLD)
1601	>	replaceWithTreeBin(tab, i, k);
1602	>	}
1603	>	break;
1604	>	}
1605	>	}
1606	>	}
1607	>	}
1608	>	if (len != 0)
1609	>	break;
1610	>	}
1611	>	}
1612	>	if (delta != 0)
1613	>	addCount((long)delta, len);
1614	>	return (V)val;
1615	>	}
1616
1617	<	boolean replaced = false;
1618	<	if (e != null && oldValue.equals(e.value)) {
1619	<	replaced = true;
1620	<	e.value = newValue;
1617	>	/** Implementation for merge */
1618	>	@SuppressWarnings("unchecked") private final V internalMerge
1619	>	(K k, V v, BiFun<? super V, ? super V, ? extends V> mf) {
1620	>	if (k == null || v == null || mf == null)
1621	>	throw new NullPointerException();
1622	>	int h = spread(k.hashCode());
1623	>	Object val = null;
1624	>	int delta = 0;
1625	>	int len = 0;
1626	>	for (Node[] tab = table;;) {
1627	>	int i; Node f; Object fk, fv;
1628	>	if (tab == null)
1629	>	tab = initTable();
1630	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1631	>	if (casTabAt(tab, i, null, new Node(h, k, v, null))) {
1632	>	delta = 1;
1633	>	val = v;
1634	>	break;
1635		}
1636	<	return replaced;
1637	<	} finally {
1638	<	unlock();
1636	>	}
1637	>	else if (f.hash < 0) {
1638	>	if ((fk = f.key) instanceof TreeBin) {
1639	>	TreeBin t = (TreeBin)fk;
1640	>	t.acquire(0);
1641	>	try {
1642	>	if (tabAt(tab, i) == f) {
1643	>	len = 1;
1644	>	TreeNode p = t.getTreeNode(h, k, t.root);
1645	>	val = (p == null) ? v : mf.apply((V)p.val, v);
1646	>	if (val != null) {
1647	>	if (p != null)
1648	>	p.val = val;
1649	>	else {
1650	>	len = 2;
1651	>	delta = 1;
1652	>	t.putTreeNode(h, k, val);
1653	>	}
1654	>	}
1655	>	else if (p != null) {
1656	>	delta = -1;
1657	>	t.deleteTreeNode(p);
1658	>	}
1659	>	}
1660	>	} finally {
1661	>	t.release(0);
1662	>	}
1663	>	if (len != 0)
1664	>	break;
1665	>	}
1666	>	else
1667	>	tab = (Node[])fk;
1668	>	}
1669	>	else {
1670	>	synchronized(f) {
1671	>	if (tabAt(tab, i) == f) {
1672	>	len = 1;
1673	>	for (Node e = f, pred = null;; ++len) {
1674	>	Object ek, ev;
1675	>	if (e.hash == h &&
1676	>	(ev = e.val) != null &&
1677	>	((ek = e.key) == k || k.equals(ek))) {
1678	>	val = mf.apply((V)ev, v);
1679	>	if (val != null)
1680	>	e.val = val;
1681	>	else {
1682	>	delta = -1;
1683	>	Node en = e.next;
1684	>	if (pred != null)
1685	>	pred.next = en;
1686	>	else
1687	>	setTabAt(tab, i, en);
1688	>	}
1689	>	break;
1690	>	}
1691	>	pred = e;
1692	>	if ((e = e.next) == null) {
1693	>	val = v;
1694	>	pred.next = new Node(h, k, val, null);
1695	>	delta = 1;
1696	>	if (len >= TREE_THRESHOLD)
1697	>	replaceWithTreeBin(tab, i, k);
1698	>	break;
1699	>	}
1700	>	}
1701	>	}
1702	>	}
1703	>	if (len != 0)
1704	>	break;
1705		}
1706		}
1707	+	if (delta != 0)
1708	+	addCount((long)delta, len);
1709	+	return (V)val;
1710	+	}
1711
1712	<	V replace(K key, int hash, V newValue) {
1713	<	lock();
1714	<	try {
1715	<	HashEntry<K,V> e = getFirst(hash);
1716	<	while (e != null && (e.hash != hash || !key.equals(e.key)))
1717	<	e = e.next;
1712	>	/** Implementation for putAll */
1713	>	private final void internalPutAll(Map<?, ?> m) {
1714	>	tryPresize(m.size());
1715	>	long delta = 0L;     // number of uncommitted additions
1716	>	boolean npe = false; // to throw exception on exit for nulls
1717	>	try {                // to clean up counts on other exceptions
1718	>	for (Map.Entry<?, ?> entry : m.entrySet()) {
1719	>	Object k, v;
1720	>	if (entry == null || (k = entry.getKey()) == null ||
1721	>	(v = entry.getValue()) == null) {
1722	>	npe = true;
1723	>	break;
1724	>	}
1725	>	int h = spread(k.hashCode());
1726	>	for (Node[] tab = table;;) {
1727	>	int i; Node f; int fh; Object fk;
1728	>	if (tab == null)
1729	>	tab = initTable();
1730	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null){
1731	>	if (casTabAt(tab, i, null, new Node(h, k, v, null))) {
1732	>	++delta;
1733	>	break;
1734	>	}
1735	>	}
1736	>	else if ((fh = f.hash) < 0) {
1737	>	if ((fk = f.key) instanceof TreeBin) {
1738	>	TreeBin t = (TreeBin)fk;
1739	>	boolean validated = false;
1740	>	t.acquire(0);
1741	>	try {
1742	>	if (tabAt(tab, i) == f) {
1743	>	validated = true;
1744	>	TreeNode p = t.getTreeNode(h, k, t.root);
1745	>	if (p != null)
1746	>	p.val = v;
1747	>	else {
1748	>	t.putTreeNode(h, k, v);
1749	>	++delta;
1750	>	}
1751	>	}
1752	>	} finally {
1753	>	t.release(0);
1754	>	}
1755	>	if (validated)
1756	>	break;
1757	>	}
1758	>	else
1759	>	tab = (Node[])fk;
1760	>	}
1761	>	else {
1762	>	int len = 0;
1763	>	synchronized(f) {
1764	>	if (tabAt(tab, i) == f) {
1765	>	len = 1;
1766	>	for (Node e = f;; ++len) {
1767	>	Object ek, ev;
1768	>	if (e.hash == h &&
1769	>	(ev = e.val) != null &&
1770	>	((ek = e.key) == k || k.equals(ek))) {
1771	>	e.val = v;
1772	>	break;
1773	>	}
1774	>	Node last = e;
1775	>	if ((e = e.next) == null) {
1776	>	++delta;
1777	>	last.next = new Node(h, k, v, null);
1778	>	if (len >= TREE_THRESHOLD)
1779	>	replaceWithTreeBin(tab, i, k);
1780	>	break;
1781	>	}
1782	>	}
1783	>	}
1784	>	}
1785	>	if (len != 0) {
1786	>	if (len > 1)
1787	>	addCount(delta, len);
1788	>	break;
1789	>	}
1790	>	}
1791	>	}
1792	>	}
1793	>	} finally {
1794	>	if (delta != 0L)
1795	>	addCount(delta, 2);
1796	>	}
1797	>	if (npe)
1798	>	throw new NullPointerException();
1799	>	}
1800
1801	<	V oldValue = null;
1802	<	if (e != null) {
1803	<	oldValue = e.value;
1804	<	e.value = newValue;
1801	>	/**
1802	>	* Implementation for clear. Steps through each bin, removing all
1803	>	* nodes.
1804	>	*/
1805	>	private final void internalClear() {
1806	>	long delta = 0L; // negative number of deletions
1807	>	int i = 0;
1808	>	Node[] tab = table;
1809	>	while (tab != null && i < tab.length) {
1810	>	Node f = tabAt(tab, i);
1811	>	if (f == null)
1812	>	++i;
1813	>	else if (f.hash < 0) {
1814	>	Object fk;
1815	>	if ((fk = f.key) instanceof TreeBin) {
1816	>	TreeBin t = (TreeBin)fk;
1817	>	t.acquire(0);
1818	>	try {
1819	>	if (tabAt(tab, i) == f) {
1820	>	for (Node p = t.first; p != null; p = p.next) {
1821	>	if (p.val != null) { // (currently always true)
1822	>	p.val = null;
1823	>	--delta;
1824	>	}
1825	>	}
1826	>	t.first = null;
1827	>	t.root = null;
1828	>	++i;
1829	>	}
1830	>	} finally {
1831	>	t.release(0);
1832	>	}
1833	>	}
1834	>	else
1835	>	tab = (Node[])fk;
1836	>	}
1837	>	else {
1838	>	synchronized(f) {
1839	>	if (tabAt(tab, i) == f) {
1840	>	for (Node e = f; e != null; e = e.next) {
1841	>	if (e.val != null) {  // (currently always true)
1842	>	e.val = null;
1843	>	--delta;
1844	>	}
1845	>	}
1846	>	setTabAt(tab, i, null);
1847	>	++i;
1848	>	}
1849		}
405	–	return oldValue;
406	–	} finally {
407	–	unlock();
1850		}
1851		}
1852	+	if (delta != 0L)
1853	+	addCount(delta, -1);
1854	+	}
1855
1856	+	/* ---------------- Table Initialization and Resizing -------------- */
1857
1858	<	V put(K key, int hash, V value, boolean onlyIfAbsent) {
1859	<	lock();
1860	<	try {
1861	<	int c = count;
1862	<	if (c++ > threshold) // ensure capacity
1863	<	rehash();
1864	<	HashEntry<K,V>[] tab = table;
1865	<	int index = hash & (tab.length - 1);
1866	<	HashEntry<K,V> first = tab[index];
1867	<	HashEntry<K,V> e = first;
1868	<	while (e != null && (e.hash != hash || !key.equals(e.key)))
1869	<	e = e.next;
1858	>	/**
1859	>	* Returns a power of two table size for the given desired capacity.
1860	>	* See Hackers Delight, sec 3.2
1861	>	*/
1862	>	private static final int tableSizeFor(int c) {
1863	>	int n = c - 1;
1864	>	n |= n >>> 1;
1865	>	n |= n >>> 2;
1866	>	n |= n >>> 4;
1867	>	n |= n >>> 8;
1868	>	n |= n >>> 16;
1869	>	return (n < 0) ? 1 : (n >= MAXIMUM_CAPACITY) ? MAXIMUM_CAPACITY : n + 1;
1870	>	}
1871
1872	<	V oldValue;
1873	<	if (e != null) {
1874	<	oldValue = e.value;
1875	<	if (!onlyIfAbsent)
1876	<	e.value = value;
1877	<	}
1878	<	else {
1879	<	oldValue = null;
1880	<	++modCount;
1881	<	tab[index] = new HashEntry<K,V>(key, hash, first, value);
1882	<	count = c; // write-volatile
1872	>	/**
1873	>	* Initializes table, using the size recorded in sizeCtl.
1874	>	*/
1875	>	private final Node[] initTable() {
1876	>	Node[] tab; int sc;
1877	>	while ((tab = table) == null) {
1878	>	if ((sc = sizeCtl) < 0)
1879	>	Thread.yield(); // lost initialization race; just spin
1880	>	else if (U.compareAndSwapInt(this, SIZECTL, sc, -1)) {
1881	>	try {
1882	>	if ((tab = table) == null) {
1883	>	int n = (sc > 0) ? sc : DEFAULT_CAPACITY;
1884	>	tab = table = new Node[n];
1885	>	sc = n - (n >>> 2);
1886	>	}
1887	>	} finally {
1888	>	sizeCtl = sc;
1889		}
1890	<	return oldValue;
438	<	} finally {
439	<	unlock();
1890	>	break;
1891		}
1892		}
1893	+	return tab;
1894	+	}
1895
1896	<	void rehash() {
1897	<	HashEntry<K,V>[] oldTable = table;
1898	<	int oldCapacity = oldTable.length;
1899	<	if (oldCapacity >= MAXIMUM_CAPACITY)
1896	>	/**
1897	>	* Adds to count, and if table is too small and not already
1898	>	* resizing, initiates transfer. If already resizing, helps
1899	>	* perform transfer if work is available.  Rechecks occupancy
1900	>	* after a transfer to see if another resize is already needed
1901	>	* because resizings are lagging additions.
1902	>	*
1903	>	* @param x the count to add
1904	>	* @param check if <0, don't check resize, if <= 1 only check if uncontended
1905	>	*/
1906	>	private final void addCount(long x, int check) {
1907	>	CounterCell[] as; long b, s;
1908	>	if ((as = counterCells) != null ||
1909	>	!U.compareAndSwapLong(this, BASECOUNT, b = baseCount, s = b + x)) {
1910	>	CounterHashCode hc; CounterCell a; long v; int m;
1911	>	boolean uncontended = true;
1912	>	if ((hc = threadCounterHashCode.get()) == null ||
1913	>	as == null || (m = as.length - 1) < 0 ||
1914	>	(a = as[m & hc.code]) == null ||
1915	>	!(uncontended =
1916	>	U.compareAndSwapLong(a, CELLVALUE, v = a.value, v + x))) {
1917	>	fullAddCount(x, hc, uncontended);
1918		return;
1919	+	}
1920	+	if (check <= 1)
1921	+	return;
1922	+	s = sumCount();
1923	+	}
1924	+	if (check >= 0) {
1925	+	Node[] tab, nt; int sc;
1926	+	while (s >= (long)(sc = sizeCtl) && (tab = table) != null &&
1927	+	tab.length < MAXIMUM_CAPACITY) {
1928	+	if (sc < 0) {
1929	+	if (sc == -1 || transferIndex <= transferOrigin ||
1930	+	(nt = nextTable) == null)
1931	+	break;
1932	+	if (U.compareAndSwapInt(this, SIZECTL, sc, sc - 1))
1933	+	transfer(tab, nt);
1934	+	}
1935	+	else if (U.compareAndSwapInt(this, SIZECTL, sc, -2))
1936	+	transfer(tab, null);
1937	+	s = sumCount();
1938	+	}
1939	+	}
1940	+	}
1941
1942	<	/*
1943	<	* Reclassify nodes in each list to new Map.  Because we are
1944	<	* using power-of-two expansion, the elements from each bin
1945	<	* must either stay at same index, or move with a power of two
1946	<	* offset. We eliminate unnecessary node creation by catching
1947	<	* cases where old nodes can be reused because their next
1948	<	* fields won't change. Statistically, at the default
1949	<	* threshold, only about one-sixth of them need cloning when
1950	<	* a table doubles. The nodes they replace will be garbage
1951	<	* collectable as soon as they are no longer referenced by any
1952	<	* reader thread that may be in the midst of traversing table
1953	<	* right now.
1954	<	*/
1955	<
1956	<	HashEntry<K,V>[] newTable = HashEntry.newArray(oldCapacity<<1);
1957	<	threshold = (int)(newTable.length * loadFactor);
1958	<	int sizeMask = newTable.length - 1;
1959	<	for (int i = 0; i < oldCapacity ; i++) {
1960	<	// We need to guarantee that any existing reads of old Map can
1961	<	//  proceed. So we cannot yet null out each bin.
1962	<	HashEntry<K,V> e = oldTable[i];
1963	<
1964	<	if (e != null) {
1965	<	HashEntry<K,V> next = e.next;
1966	<	int idx = e.hash & sizeMask;
1967	<
1968	<	//  Single node on list
1969	<	if (next == null)
1970	<	newTable[idx] = e;
1942	>	/**
1943	>	* Tries to presize table to accommodate the given number of elements.
1944	>	*
1945	>	* @param size number of elements (doesn't need to be perfectly accurate)
1946	>	*/
1947	>	private final void tryPresize(int size) {
1948	>	int c = (size >= (MAXIMUM_CAPACITY >>> 1)) ? MAXIMUM_CAPACITY :
1949	>	tableSizeFor(size + (size >>> 1) + 1);
1950	>	int sc;
1951	>	while ((sc = sizeCtl) >= 0) {
1952	>	Node[] tab = table; int n;
1953	>	if (tab == null || (n = tab.length) == 0) {
1954	>	n = (sc > c) ? sc : c;
1955	>	if (U.compareAndSwapInt(this, SIZECTL, sc, -1)) {
1956	>	try {
1957	>	if (table == tab) {
1958	>	table = new Node[n];
1959	>	sc = n - (n >>> 2);
1960	>	}
1961	>	} finally {
1962	>	sizeCtl = sc;
1963	>	}
1964	>	}
1965	>	}
1966	>	else if (c <= sc || n >= MAXIMUM_CAPACITY)
1967	>	break;
1968	>	else if (tab == table &&
1969	>	U.compareAndSwapInt(this, SIZECTL, sc, -2))
1970	>	transfer(tab, null);
1971	>	}
1972	>	}
1973
1974	<	else {
1975	<	// Reuse trailing consecutive sequence at same slot
1976	<	HashEntry<K,V> lastRun = e;
1977	<	int lastIdx = idx;
1978	<	for (HashEntry<K,V> last = next;
1979	<	last != null;
1980	<	last = last.next) {
1981	<	int k = last.hash & sizeMask;
1982	<	if (k != lastIdx) {
1983	<	lastIdx = k;
1984	<	lastRun = last;
1974	>	/*
1975	>	* Moves and/or copies the nodes in each bin to new table. See
1976	>	* above for explanation.
1977	>	*/
1978	>	private final void transfer(Node[] tab, Node[] nextTab) {
1979	>	int n = tab.length, stride;
1980	>	if ((stride = (NCPU > 1) ? (n >>> 3) / NCPU : n) < MIN_TRANSFER_STRIDE)
1981	>	stride = MIN_TRANSFER_STRIDE; // subdivide range
1982	>	if (nextTab == null) {            // initiating
1983	>	try {
1984	>	nextTab = new Node[n << 1];
1985	>	} catch(Throwable ex) {       // try to cope with OOME
1986	>	sizeCtl = Integer.MAX_VALUE;
1987	>	return;
1988	>	}
1989	>	nextTable = nextTab;
1990	>	transferOrigin = n;
1991	>	transferIndex = n;
1992	>	Node rev = new Node(MOVED, tab, null, null);
1993	>	for (int k = n; k > 0;) {    // progressively reveal ready slots
1994	>	int nextk = k > stride? k - stride : 0;
1995	>	for (int m = nextk; m < k; ++m)
1996	>	nextTab[m] = rev;
1997	>	for (int m = n + nextk; m < n + k; ++m)
1998	>	nextTab[m] = rev;
1999	>	U.putOrderedInt(this, TRANSFERORIGIN, k = nextk);
2000	>	}
2001	>	}
2002	>	int nextn = nextTab.length;
2003	>	Node fwd = new Node(MOVED, nextTab, null, null);
2004	>	boolean advance = true;
2005	>	for (int i = 0, bound = 0;;) {
2006	>	int nextIndex, nextBound; Node f; Object fk;
2007	>	while (advance) {
2008	>	if (--i >= bound)
2009	>	advance = false;
2010	>	else if ((nextIndex = transferIndex) <= transferOrigin) {
2011	>	i = -1;
2012	>	advance = false;
2013	>	}
2014	>	else if (U.compareAndSwapInt
2015	>	(this, TRANSFERINDEX, nextIndex,
2016	>	nextBound = (nextIndex > stride?
2017	>	nextIndex - stride : 0))) {
2018	>	bound = nextBound;
2019	>	i = nextIndex - 1;
2020	>	advance = false;
2021	>	}
2022	>	}
2023	>	if (i < 0 || i >= n || i + n >= nextn) {
2024	>	for (int sc;;) {
2025	>	if (U.compareAndSwapInt(this, SIZECTL, sc = sizeCtl, ++sc)) {
2026	>	if (sc == -1) {
2027	>	nextTable = null;
2028	>	table = nextTab;
2029	>	sizeCtl = (n << 1) - (n >>> 1);
2030	>	}
2031	>	return;
2032	>	}
2033	>	}
2034	>	}
2035	>	else if ((f = tabAt(tab, i)) == null) {
2036	>	if (casTabAt(tab, i, null, fwd)) {
2037	>	setTabAt(nextTab, i, null);
2038	>	setTabAt(nextTab, i + n, null);
2039	>	advance = true;
2040	>	}
2041	>	}
2042	>	else if (f.hash >= 0) {
2043	>	synchronized(f) {
2044	>	if (tabAt(tab, i) == f) {
2045	>	int runBit = f.hash & n;
2046	>	Node lastRun = f, lo = null, hi = null;
2047	>	for (Node p = f.next; p != null; p = p.next) {
2048	>	int b = p.hash & n;
2049	>	if (b != runBit) {
2050	>	runBit = b;
2051	>	lastRun = p;
2052		}
2053		}
2054	<	newTable[lastIdx] = lastRun;
2055	<
2056	<	// Clone all remaining nodes
2057	<	for (HashEntry<K,V> p = e; p != lastRun; p = p.next) {
2058	<	int k = p.hash & sizeMask;
2059	<	HashEntry<K,V> n = newTable[k];
2060	<	newTable[k] = new HashEntry<K,V>(p.key, p.hash,
2061	<	n, p.value);
2054	>	if (runBit == 0)
2055	>	lo = lastRun;
2056	>	else
2057	>	hi = lastRun;
2058	>	for (Node p = f; p != lastRun; p = p.next) {
2059	>	int ph = p.hash;
2060	>	Object pk = p.key, pv = p.val;
2061	>	if ((ph & n) == 0)
2062	>	lo = new Node(ph, pk, pv, lo);
2063	>	else
2064	>	hi = new Node(ph, pk, pv, hi);
2065	>	}
2066	>	setTabAt(nextTab, i, lo);
2067	>	setTabAt(nextTab, i + n, hi);
2068	>	setTabAt(tab, i, fwd);
2069	>	advance = true;
2070	>	}
2071	>	}
2072	>	}
2073	>	else if ((fk = f.key) instanceof TreeBin) {
2074	>	TreeBin t = (TreeBin)fk;
2075	>	t.acquire(0);
2076	>	try {
2077	>	if (tabAt(tab, i) == f) {
2078	>	TreeBin lt = new TreeBin();
2079	>	TreeBin ht = new TreeBin();
2080	>	int lc = 0, hc = 0;
2081	>	for (Node e = t.first; e != null; e = e.next) {
2082	>	int h = e.hash;
2083	>	Object k = e.key, v = e.val;
2084	>	if ((h & n) == 0) {
2085	>	++lc;
2086	>	lt.putTreeNode(h, k, v);
2087	>	}
2088	>	else {
2089	>	++hc;
2090	>	ht.putTreeNode(h, k, v);
2091	>	}
2092	>	}
2093	>	Node ln, hn; // throw away trees if too small
2094	>	if (lc < TREE_THRESHOLD) {
2095	>	ln = null;
2096	>	for (Node p = lt.first; p != null; p = p.next)
2097	>	ln = new Node(p.hash, p.key, p.val, ln);
2098		}
2099	+	else
2100	+	ln = new Node(MOVED, lt, null, null);
2101	+	setTabAt(nextTab, i, ln);
2102	+	if (hc < TREE_THRESHOLD) {
2103	+	hn = null;
2104	+	for (Node p = ht.first; p != null; p = p.next)
2105	+	hn = new Node(p.hash, p.key, p.val, hn);
2106	+	}
2107	+	else
2108	+	hn = new Node(MOVED, ht, null, null);
2109	+	setTabAt(nextTab, i + n, hn);
2110	+	setTabAt(tab, i, fwd);
2111	+	advance = true;
2112		}
2113	+	} finally {
2114	+	t.release(0);
2115		}
2116		}
2117	<	table = newTable;
2117	>	else
2118	>	advance = true; // already processed
2119		}
2120	+	}
2121
2122	<	/**
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;
2122	>	/* ---------------- Counter support -------------- */
2123
2124	<	V oldValue = null;
2125	<	if (e != null) {
2126	<	V v = e.value;
2127	<	if (value == null || value.equals(v)) {
2128	<	oldValue = v;
2129	<	// All entries following removed node can stay
2130	<	// in list, but all preceding ones need to be
528	<	// cloned.
529	<	++modCount;
530	<	HashEntry<K,V> newFirst = e.next;
531	<	for (HashEntry<K,V> p = first; p != e; p = p.next)
532	<	newFirst = new HashEntry<K,V>(p.key, p.hash,
533	<	newFirst, p.value);
534	<	tab[index] = newFirst;
535	<	count = c; // write-volatile
536	<	}
537	<	}
538	<	return oldValue;
539	<	} finally {
540	<	unlock();
2124	>	final long sumCount() {
2125	>	CounterCell[] as = counterCells; CounterCell a;
2126	>	long sum = baseCount;
2127	>	if (as != null) {
2128	>	for (int i = 0; i < as.length; ++i) {
2129	>	if ((a = as[i]) != null)
2130	>	sum += a.value;
2131		}
2132		}
2133	+	return sum;
2134	+	}
2135
2136	<	void clear() {
2137	<	if (count != 0) {
2138	<	lock();
2139	<	try {
2140	<	HashEntry<K,V>[] tab = table;
2141	<	for (int i = 0; i < tab.length ; i++)
2142	<	tab[i] = null;
2143	<	++modCount;
2144	<	count = 0; // write-volatile
2136	>	// See LongAdder version for explanation
2137	>	private final void fullAddCount(long x, CounterHashCode hc,
2138	>	boolean wasUncontended) {
2139	>	int h;
2140	>	if (hc == null) {
2141	>	hc = new CounterHashCode();
2142	>	int s = counterHashCodeGenerator.addAndGet(SEED_INCREMENT);
2143	>	h = hc.code = (s == 0) ? 1 : s; // Avoid zero
2144	>	threadCounterHashCode.set(hc);
2145	>	}
2146	>	else
2147	>	h = hc.code;
2148	>	boolean collide = false;                // True if last slot nonempty
2149	>	for (;;) {
2150	>	CounterCell[] as; CounterCell a; int n; long v;
2151	>	if ((as = counterCells) != null && (n = as.length) > 0) {
2152	>	if ((a = as[(n - 1) & h]) == null) {
2153	>	if (counterBusy == 0) {            // Try to attach new Cell
2154	>	CounterCell r = new CounterCell(x); // Optimistic create
2155	>	if (counterBusy == 0 &&
2156	>	U.compareAndSwapInt(this, COUNTERBUSY, 0, 1)) {
2157	>	boolean created = false;
2158	>	try {               // Recheck under lock
2159	>	CounterCell[] rs; int m, j;
2160	>	if ((rs = counterCells) != null &&
2161	>	(m = rs.length) > 0 &&
2162	>	rs[j = (m - 1) & h] == null) {
2163	>	rs[j] = r;
2164	>	created = true;
2165	>	}
2166	>	} finally {
2167	>	counterBusy = 0;
2168	>	}
2169	>	if (created)
2170	>	break;
2171	>	continue;           // Slot is now non-empty
2172	>	}
2173	>	}
2174	>	collide = false;
2175	>	}
2176	>	else if (!wasUncontended)       // CAS already known to fail
2177	>	wasUncontended = true;      // Continue after rehash
2178	>	else if (U.compareAndSwapLong(a, CELLVALUE, v = a.value, v + x))
2179	>	break;
2180	>	else if (counterCells != as || n >= NCPU)
2181	>	collide = false;            // At max size or stale
2182	>	else if (!collide)
2183	>	collide = true;
2184	>	else if (counterBusy == 0 &&
2185	>	U.compareAndSwapInt(this, COUNTERBUSY, 0, 1)) {
2186	>	try {
2187	>	if (counterCells == as) {// Expand table unless stale
2188	>	CounterCell[] rs = new CounterCell[n << 1];
2189	>	for (int i = 0; i < n; ++i)
2190	>	rs[i] = as[i];
2191	>	counterCells = rs;
2192	>	}
2193	>	} finally {
2194	>	counterBusy = 0;
2195	>	}
2196	>	collide = false;
2197	>	continue;                   // Retry with expanded table
2198	>	}
2199	>	h ^= h << 13;                   // Rehash
2200	>	h ^= h >>> 17;
2201	>	h ^= h << 5;
2202	>	}
2203	>	else if (counterBusy == 0 && counterCells == as &&
2204	>	U.compareAndSwapInt(this, COUNTERBUSY, 0, 1)) {
2205	>	boolean init = false;
2206	>	try {                           // Initialize table
2207	>	if (counterCells == as) {
2208	>	CounterCell[] rs = new CounterCell[2];
2209	>	rs[h & 1] = new CounterCell(x);
2210	>	counterCells = rs;
2211	>	init = true;
2212	>	}
2213		} finally {
2214	<	unlock();
2214	>	counterBusy = 0;
2215		}
2216	+	if (init)
2217	+	break;
2218		}
2219	+	else if (U.compareAndSwapLong(this, BASECOUNT, v = baseCount, v + x))
2220	+	break;                          // Fall back on using base
2221		}
2222	+	hc.code = h;                            // Record index for next time
2223		}
2224
2225	+	/* ----------------Table Traversal -------------- */
2226	+
2227	+	/**
2228	+	* Encapsulates traversal for methods such as containsValue; also
2229	+	* serves as a base class for other iterators and bulk tasks.
2230	+	*
2231	+	* At each step, the iterator snapshots the key ("nextKey") and
2232	+	* value ("nextVal") of a valid node (i.e., one that, at point of
2233	+	* snapshot, has a non-null user value). Because val fields can
2234	+	* change (including to null, indicating deletion), field nextVal
2235	+	* might not be accurate at point of use, but still maintains the
2236	+	* weak consistency property of holding a value that was once
2237	+	* valid. To support iterator.remove, the nextKey field is not
2238	+	* updated (nulled out) when the iterator cannot advance.
2239	+	*
2240	+	* Internal traversals directly access these fields, as in:
2241	+	* {@code while (it.advance() != null) { process(it.nextKey); }}
2242	+	*
2243	+	* Exported iterators must track whether the iterator has advanced
2244	+	* (in hasNext vs next) (by setting/checking/nulling field
2245	+	* nextVal), and then extract key, value, or key-value pairs as
2246	+	* return values of next().
2247	+	*
2248	+	* The iterator visits once each still-valid node that was
2249	+	* reachable upon iterator construction. It might miss some that
2250	+	* were added to a bin after the bin was visited, which is OK wrt
2251	+	* consistency guarantees. Maintaining this property in the face
2252	+	* of possible ongoing resizes requires a fair amount of
2253	+	* bookkeeping state that is difficult to optimize away amidst
2254	+	* volatile accesses.  Even so, traversal maintains reasonable
2255	+	* throughput.
2256	+	*
2257	+	* Normally, iteration proceeds bin-by-bin traversing lists.
2258	+	* However, if the table has been resized, then all future steps
2259	+	* must traverse both the bin at the current index as well as at
2260	+	* (index + baseSize); and so on for further resizings. To
2261	+	* paranoically cope with potential sharing by users of iterators
2262	+	* across threads, iteration terminates if a bounds checks fails
2263	+	* for a table read.
2264	+	*
2265	+	* This class extends CountedCompleter to streamline parallel
2266	+	* iteration in bulk operations. This adds only a few fields of
2267	+	* space overhead, which is small enough in cases where it is not
2268	+	* needed to not worry about it.  Because CountedCompleter is
2269	+	* Serializable, but iterators need not be, we need to add warning
2270	+	* suppressions.
2271	+	*/
2272	+	@SuppressWarnings("serial") static class Traverser<K,V,R>
2273	+	extends CountedCompleter<R> {
2274	+	final ConcurrentHashMap<K, V> map;
2275	+	Node next;           // the next entry to use
2276	+	Object nextKey;      // cached key field of next
2277	+	Object nextVal;      // cached val field of next
2278	+	Node[] tab;          // current table; updated if resized
2279	+	int index;           // index of bin to use next
2280	+	int baseIndex;       // current index of initial table
2281	+	int baseLimit;       // index bound for initial table
2282	+	int baseSize;        // initial table size
2283	+	int batch;           // split control
2284	+
2285	+	/** Creates iterator for all entries in the table. */
2286	+	Traverser(ConcurrentHashMap<K, V> map) {
2287	+	this.map = map;
2288	+	}
2289	+
2290	+	/** Creates iterator for split() methods and task constructors */
2291	+	Traverser(ConcurrentHashMap<K,V> map, Traverser<K,V,?> it, int batch) {
2292	+	super(it);
2293	+	this.batch = batch;
2294	+	if ((this.map = map) != null && it != null) { // split parent
2295	+	Node[] t;
2296	+	if ((t = it.tab) == null &&
2297	+	(t = it.tab = map.table) != null)
2298	+	it.baseLimit = it.baseSize = t.length;
2299	+	this.tab = t;
2300	+	this.baseSize = it.baseSize;
2301	+	int hi = this.baseLimit = it.baseLimit;
2302	+	it.baseLimit = this.index = this.baseIndex =
2303	+	(hi + it.baseIndex + 1) >>> 1;
2304	+	}
2305	+	}
2306	+
2307	+	/**
2308	+	* Advances next; returns nextVal or null if terminated.
2309	+	* See above for explanation.
2310	+	*/
2311	+	final Object advance() {
2312	+	Node e = next;
2313	+	Object ev = null;
2314	+	outer: do {
2315	+	if (e != null)                  // advance past used/skipped node
2316	+	e = e.next;
2317	+	while (e == null) {             // get to next non-null bin
2318	+	ConcurrentHashMap<K, V> m;
2319	+	Node[] t; int b, i, n; Object ek; // checks must use locals
2320	+	if ((t = tab) != null)
2321	+	n = t.length;
2322	+	else if ((m = map) != null && (t = tab = m.table) != null)
2323	+	n = baseLimit = baseSize = t.length;
2324	+	else
2325	+	break outer;
2326	+	if ((b = baseIndex) >= baseLimit ||
2327	+	(i = index) < 0 || i >= n)
2328	+	break outer;
2329	+	if ((e = tabAt(t, i)) != null && e.hash < 0) {
2330	+	if ((ek = e.key) instanceof TreeBin)
2331	+	e = ((TreeBin)ek).first;
2332	+	else {
2333	+	tab = (Node[])ek;
2334	+	continue;           // restarts due to null val
2335	+	}
2336	+	}                           // visit upper slots if present
2337	+	index = (i += baseSize) < n ? i : (baseIndex = b + 1);
2338	+	}
2339	+	nextKey = e.key;
2340	+	} while ((ev = e.val) == null);    // skip deleted or special nodes
2341	+	next = e;
2342	+	return nextVal = ev;
2343	+	}
2344	+
2345	+	public final void remove() {
2346	+	Object k = nextKey;
2347	+	if (k == null && (advance() == null || (k = nextKey) == null))
2348	+	throw new IllegalStateException();
2349	+	map.internalReplace(k, null, null);
2350	+	}
2351	+
2352	+	public final boolean hasNext() {
2353	+	return nextVal != null || advance() != null;
2354	+	}
2355	+
2356	+	public final boolean hasMoreElements() { return hasNext(); }
2357	+
2358	+	public void compute() { } // default no-op CountedCompleter body
2359	+
2360	+	/**
2361	+	* Returns a batch value > 0 if this task should (and must) be
2362	+	* split, if so, adding to pending count, and in any case
2363	+	* updating batch value. The initial batch value is approx
2364	+	* exp2 of the number of times (minus one) to split task by
2365	+	* two before executing leaf action. This value is faster to
2366	+	* compute and more convenient to use as a guide to splitting
2367	+	* than is the depth, since it is used while dividing by two
2368	+	* anyway.
2369	+	*/
2370	+	final int preSplit() {
2371	+	ConcurrentHashMap<K, V> m; int b; Node[] t;  ForkJoinPool pool;
2372	+	if ((b = batch) < 0 && (m = map) != null) { // force initialization
2373	+	if ((t = tab) == null && (t = tab = m.table) != null)
2374	+	baseLimit = baseSize = t.length;
2375	+	if (t != null) {
2376	+	long n = m.sumCount();
2377	+	int par = ((pool = getPool()) == null) ?
2378	+	ForkJoinPool.getCommonPoolParallelism() :
2379	+	pool.getParallelism();
2380	+	int sp = par << 3; // slack of 8
2381	+	b = (n <= 0L) ? 0 : (n < (long)sp) ? (int)n : sp;
2382	+	}
2383	+	}
2384	+	b = (b <= 1 || baseIndex == baseLimit) ? 0 : (b >>> 1);
2385	+	if ((batch = b) > 0)
2386	+	addToPendingCount(1);
2387	+	return b;
2388	+	}
2389
2390	+	}
2391
2392		/* ---------------- Public operations -------------- */
2393
2394		/**
2395	<	* Creates a new, empty map with the specified initial
2396	<	* capacity, load factor and concurrency level.
2395	>	* Creates a new, empty map with the default initial table size (16).
2396	>	*/
2397	>	public ConcurrentHashMap() {
2398	>	}
2399	>
2400	>	/**
2401	>	* Creates a new, empty map with an initial table size
2402	>	* accommodating the specified number of elements without the need
2403	>	* to dynamically resize.
2404		*
2405	<	* @param initialCapacity the initial capacity. The implementation
2406	<	* performs internal sizing to accommodate this many elements.
2407	<	* @param loadFactor  the load factor threshold, used to control resizing.
2408	<	* 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.
2405	>	* @param initialCapacity The implementation performs internal
2406	>	* sizing to accommodate this many elements.
2407	>	* @throws IllegalArgumentException if the initial capacity of
2408	>	* elements is negative
2409		*/
2410	<	public ConcurrentHashMap(int initialCapacity,
2411	<	float loadFactor, int concurrencyLevel) {
582	<	if (!(loadFactor > 0) || initialCapacity < 0 || concurrencyLevel <= 0)
2410	>	public ConcurrentHashMap(int initialCapacity) {
2411	>	if (initialCapacity < 0)
2412		throw new IllegalArgumentException();
2413	+	int cap = ((initialCapacity >= (MAXIMUM_CAPACITY >>> 1)) ?
2414	+	MAXIMUM_CAPACITY :
2415	+	tableSizeFor(initialCapacity + (initialCapacity >>> 1) + 1));
2416	+	this.sizeCtl = cap;
2417	+	}
2418
2419	<	if (concurrencyLevel > MAX_SEGMENTS)
2420	<	concurrencyLevel = MAX_SEGMENTS;
2421	<
2422	<	// Find power-of-two sizes best matching arguments
2423	<	int sshift = 0;
2424	<	int ssize = 1;
2425	<	while (ssize < concurrencyLevel) {
2426	<	++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;
607	<
608	<	for (int i = 0; i < this.segments.length; ++i)
609	<	this.segments[i] = new Segment<K,V>(cap, loadFactor);
2419	>	/**
2420	>	* Creates a new map with the same mappings as the given map.
2421	>	*
2422	>	* @param m the map
2423	>	*/
2424	>	public ConcurrentHashMap(Map<? extends K, ? extends V> m) {
2425	>	this.sizeCtl = DEFAULT_CAPACITY;
2426	>	internalPutAll(m);
2427		}
2428
2429		/**
2430	<	* Creates a new, empty map with the specified initial capacity
2431	<	* and load factor and with the default concurrencyLevel (16).
2430	>	* Creates a new, empty map with an initial table size based on
2431	>	* the given number of elements ({@code initialCapacity}) and
2432	>	* initial table density ({@code loadFactor}).
2433		*
2434	<	* @param initialCapacity The implementation performs internal
2435	<	* sizing to accommodate this many elements.
2436	<	* @param loadFactor  the load factor threshold, used to control resizing.
2437	<	* Resizing may be performed when the average number of elements per
2438	<	* bin exceeds this threshold.
2434	>	* @param initialCapacity the initial capacity. The implementation
2435	>	* performs internal sizing to accommodate this many elements,
2436	>	* given the specified load factor.
2437	>	* @param loadFactor the load factor (table density) for
2438	>	* establishing the initial table size
2439		* @throws IllegalArgumentException if the initial capacity of
2440		* elements is negative or the load factor is nonpositive
2441		*
2442		* @since 1.6
2443		*/
2444		public ConcurrentHashMap(int initialCapacity, float loadFactor) {
2445	<	this(initialCapacity, loadFactor, DEFAULT_CONCURRENCY_LEVEL);
2445	>	this(initialCapacity, loadFactor, 1);
2446		}
2447
2448		/**
2449	<	* Creates a new, empty map with the specified initial capacity,
2450	<	* and with default load factor (0.75) and concurrencyLevel (16).
2449	>	* Creates a new, empty map with an initial table size based on
2450	>	* the given number of elements ({@code initialCapacity}), table
2451	>	* density ({@code loadFactor}), and number of concurrently
2452	>	* updating threads ({@code concurrencyLevel}).
2453		*
2454		* @param initialCapacity the initial capacity. The implementation
2455	<	* performs internal sizing to accommodate this many elements.
2456	<	* @throws IllegalArgumentException if the initial capacity of
2457	<	* elements is negative.
2455	>	* performs internal sizing to accommodate this many elements,
2456	>	* given the specified load factor.
2457	>	* @param loadFactor the load factor (table density) for
2458	>	* establishing the initial table size
2459	>	* @param concurrencyLevel the estimated number of concurrently
2460	>	* updating threads. The implementation may use this value as
2461	>	* a sizing hint.
2462	>	* @throws IllegalArgumentException if the initial capacity is
2463	>	* negative or the load factor or concurrencyLevel are
2464	>	* nonpositive
2465		*/
2466	<	public ConcurrentHashMap(int initialCapacity) {
2467	<	this(initialCapacity, DEFAULT_LOAD_FACTOR, DEFAULT_CONCURRENCY_LEVEL);
2466	>	public ConcurrentHashMap(int initialCapacity,
2467	>	float loadFactor, int concurrencyLevel) {
2468	>	if (!(loadFactor > 0.0f) || initialCapacity < 0 || concurrencyLevel <= 0)
2469	>	throw new IllegalArgumentException();
2470	>	if (initialCapacity < concurrencyLevel)   // Use at least as many bins
2471	>	initialCapacity = concurrencyLevel;   // as estimated threads
2472	>	long size = (long)(1.0 + (long)initialCapacity / loadFactor);
2473	>	int cap = (size >= (long)MAXIMUM_CAPACITY) ?
2474	>	MAXIMUM_CAPACITY : tableSizeFor((int)size);
2475	>	this.sizeCtl = cap;
2476		}
2477
2478		/**
2479	<	* Creates a new, empty map with a default initial capacity (16),
2480	<	* load factor (0.75) and concurrencyLevel (16).
2479	>	* Creates a new {@link Set} backed by a ConcurrentHashMap
2480	>	* from the given type to {@code Boolean.TRUE}.
2481	>	*
2482	>	* @return the new set
2483		*/
2484	<	public ConcurrentHashMap() {
2485	<	this(DEFAULT_INITIAL_CAPACITY, DEFAULT_LOAD_FACTOR, DEFAULT_CONCURRENCY_LEVEL);
2484	>	public static <K> KeySetView<K,Boolean> newKeySet() {
2485	>	return new KeySetView<K,Boolean>(new ConcurrentHashMap<K,Boolean>(),
2486	>	Boolean.TRUE);
2487		}
2488
2489		/**
2490	<	* Creates a new map with the same mappings as the given map.
2491	<	* 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).
2490	>	* Creates a new {@link Set} backed by a ConcurrentHashMap
2491	>	* from the given type to {@code Boolean.TRUE}.
2492		*
2493	<	* @param m the map
2493	>	* @param initialCapacity The implementation performs internal
2494	>	* sizing to accommodate this many elements.
2495	>	* @throws IllegalArgumentException if the initial capacity of
2496	>	* elements is negative
2497	>	* @return the new set
2498		*/
2499	<	public ConcurrentHashMap(Map<? extends K, ? extends V> m) {
2500	<	this(Math.max((int) (m.size() / DEFAULT_LOAD_FACTOR) + 1,
2501	<	DEFAULT_INITIAL_CAPACITY),
662	<	DEFAULT_LOAD_FACTOR, DEFAULT_CONCURRENCY_LEVEL);
663	<	putAll(m);
2499	>	public static <K> KeySetView<K,Boolean> newKeySet(int initialCapacity) {
2500	>	return new KeySetView<K,Boolean>
2501	>	(new ConcurrentHashMap<K,Boolean>(initialCapacity), Boolean.TRUE);
2502		}
2503
2504		/**
2505	<	* 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
2505	>	* {@inheritDoc}
2506		*/
2507		public boolean isEmpty() {
2508	<	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;
2508	>	return sumCount() <= 0L; // ignore transient negative values
2509		}
2510
2511		/**
2512	<	* 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
2512	>	* {@inheritDoc}
2513		*/
2514		public int size() {
2515	<	final Segment<K,V>[] segments = this.segments;
2516	<	long sum = 0;
2517	<	long check = 0;
2518	<	int[] mc = new int[segments.length];
715	<	// Try a few times to get accurate count. On failure due to
716	<	// continuous async changes in table, resort to locking.
717	<	for (int k = 0; k < RETRIES_BEFORE_LOCK; ++k) {
718	<	check = 0;
719	<	sum = 0;
720	<	int mcsum = 0;
721	<	for (int i = 0; i < segments.length; ++i) {
722	<	sum += segments[i].count;
723	<	mcsum += mc[i] = segments[i].modCount;
724	<	}
725	<	if (mcsum != 0) {
726	<	for (int i = 0; i < segments.length; ++i) {
727	<	check += segments[i].count;
728	<	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;
2515	>	long n = sumCount();
2516	>	return ((n < 0L) ? 0 :
2517	>	(n > (long)Integer.MAX_VALUE) ? Integer.MAX_VALUE :
2518	>	(int)n);
2519		}
2520
2521		/**
2522	<	* Returns the value to which this map maps the specified key, or
2523	<	* <tt>null</tt> if the map contains no mapping for the key.
2522	>	* Returns the number of mappings. This method should be used
2523	>	* instead of {@link #size} because a ConcurrentHashMap may
2524	>	* contain more mappings than can be represented as an int. The
2525	>	* value returned is an estimate; the actual count may differ if
2526	>	* there are concurrent insertions or removals.
2527	>	*
2528	>	* @return the number of mappings
2529	>	*/
2530	>	public long mappingCount() {
2531	>	long n = sumCount();
2532	>	return (n < 0L) ? 0L : n; // ignore transient negative values
2533	>	}
2534	>
2535	>	/**
2536	>	* Returns the value to which the specified key is mapped,
2537	>	* or {@code null} if this map contains no mapping for the key.
2538	>	*
2539	>	* <p>More formally, if this map contains a mapping from a key
2540	>	* {@code k} to a value {@code v} such that {@code key.equals(k)},
2541	>	* then this method returns {@code v}; otherwise it returns
2542	>	* {@code null}.  (There can be at most one such mapping.)
2543		*
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
2544		* @throws NullPointerException if the specified key is null
2545		*/
2546		public V get(Object key) {
2547	<	int hash = hash(key); // throws NullPointerException if key null
2548	<	return segmentFor(hash).get(key, hash);
2547	>	return internalGet(key);
2548	>	}
2549	>
2550	>	/**
2551	>	* Returns the value to which the specified key is mapped,
2552	>	* or the given defaultValue if this map contains no mapping for the key.
2553	>	*
2554	>	* @param key the key
2555	>	* @param defaultValue the value to return if this map contains
2556	>	* no mapping for the given key
2557	>	* @return the mapping for the key, if present; else the defaultValue
2558	>	* @throws NullPointerException if the specified key is null
2559	>	*/
2560	>	public V getValueOrDefault(Object key, V defaultValue) {
2561	>	V v;
2562	>	return (v = internalGet(key)) == null ? defaultValue : v;
2563		}
2564
2565		/**
2566		* Tests if the specified object is a key in this table.
2567		*
2568		* @param  key   possible key
2569	<	* @return <tt>true</tt> if and only if the specified object
2569	>	* @return {@code true} if and only if the specified object
2570		*         is a key in this table, as determined by the
2571	<	*         <tt>equals</tt> method; <tt>false</tt> otherwise.
2571	>	*         {@code equals} method; {@code false} otherwise
2572		* @throws NullPointerException if the specified key is null
2573		*/
2574		public boolean containsKey(Object key) {
2575	<	int hash = hash(key); // throws NullPointerException if key null
777	<	return segmentFor(hash).containsKey(key, hash);
2575	>	return internalGet(key) != null;
2576		}
2577
2578		/**
2579	<	* Returns <tt>true</tt> if this map maps one or more keys to the
2580	<	* specified value. Note: This method requires a full internal
2581	<	* traversal of the hash table, and so is much slower than
784	<	* method <tt>containsKey</tt>.
2579	>	* Returns {@code true} if this map maps one or more keys to the
2580	>	* specified value. Note: This method may require a full traversal
2581	>	* of the map, and is much slower than method {@code containsKey}.
2582		*
2583		* @param value value whose presence in this map is to be tested
2584	<	* @return <tt>true</tt> if this map maps one or more keys to the
2584	>	* @return {@code true} if this map maps one or more keys to the
2585		*         specified value
2586		* @throws NullPointerException if the specified value is null
2587		*/
2588		public boolean containsValue(Object value) {
2589		if (value == null)
2590		throw new NullPointerException();
2591	<
2592	<	// See explanation of modCount use above
2593	<
2594	<	final Segment<K,V>[] segments = this.segments;
2595	<	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();
2591	>	Object v;
2592	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
2593	>	while ((v = it.advance()) != null) {
2594	>	if (v == value || value.equals(v))
2595	>	return true;
2596		}
2597	<	return found;
2597	>	return false;
2598		}
2599
2600		/**
#	Line 845 | Line 2604 | public class ConcurrentHashMap<K, V> ext
2604		* full compatibility with class {@link java.util.Hashtable},
2605		* which supported this method prior to introduction of the
2606		* Java Collections framework.
2607	<
2607	>	*
2608		* @param  value a value to search for
2609	<	* @return <tt>true</tt> if and only if some key maps to the
2610	<	*         <tt>value</tt> argument in this table as
2611	<	*         determined by the <tt>equals</tt> method;
2612	<	*         <tt>false</tt> otherwise
2609	>	* @return {@code true} if and only if some key maps to the
2610	>	*         {@code value} argument in this table as
2611	>	*         determined by the {@code equals} method;
2612	>	*         {@code false} otherwise
2613		* @throws NullPointerException if the specified value is null
2614		*/
2615		public boolean contains(Object value) {
#	Line 861 | Line 2620 | public class ConcurrentHashMap<K, V> ext
2620		* Maps the specified key to the specified value in this table.
2621		* Neither the key nor the value can be null.
2622		*
2623	<	* <p> The value can be retrieved by calling the <tt>get</tt> method
2623	>	* <p>The value can be retrieved by calling the {@code get} method
2624		* with a key that is equal to the original key.
2625		*
2626		* @param key key with which the specified value is to be associated
2627		* @param value value to be associated with the specified key
2628	<	* @return the previous value associated with <tt>key</tt>, or
2629	<	*         <tt>null</tt> if there was no mapping for <tt>key</tt>
2628	>	* @return the previous value associated with {@code key}, or
2629	>	*         {@code null} if there was no mapping for {@code key}
2630		* @throws NullPointerException if the specified key or value is null
2631		*/
2632		public V put(K key, V value) {
2633	<	if (value == null)
875	<	throw new NullPointerException();
876	<	int hash = hash(key);
877	<	return segmentFor(hash).put(key, hash, value, false);
2633	>	return internalPut(key, value, false);
2634		}
2635
2636		/**
2637		* {@inheritDoc}
2638		*
2639		* @return the previous value associated with the specified key,
2640	<	*         or <tt>null</tt> if there was no mapping for the key
2640	>	*         or {@code null} if there was no mapping for the key
2641		* @throws NullPointerException if the specified key or value is null
2642		*/
2643		public V putIfAbsent(K key, V value) {
2644	<	if (value == null)
889	<	throw new NullPointerException();
890	<	int hash = hash(key);
891	<	return segmentFor(hash).put(key, hash, value, true);
2644	>	return internalPut(key, value, true);
2645		}
2646
2647		/**
#	Line 899 | Line 2652 | public class ConcurrentHashMap<K, V> ext
2652		* @param m mappings to be stored in this map
2653		*/
2654		public void putAll(Map<? extends K, ? extends V> m) {
2655	<	for (Iterator<? extends Map.Entry<? extends K, ? extends V>> it = (Iterator<? extends Map.Entry<? extends K, ? extends V>>) m.entrySet().iterator(); it.hasNext(); ) {
2656	<	Entry<? extends K, ? extends V> e = it.next();
2657	<	put(e.getKey(), e.getValue());
2658	<	}
2655	>	internalPutAll(m);
2656	>	}
2657	>
2658	>	/**
2659	>	* If the specified key is not already associated with a value,
2660	>	* computes its value using the given mappingFunction and enters
2661	>	* it into the map unless null.  This is equivalent to
2662	>	* <pre> {@code
2663	>	* if (map.containsKey(key))
2664	>	*   return map.get(key);
2665	>	* value = mappingFunction.apply(key);
2666	>	* if (value != null)
2667	>	*   map.put(key, value);
2668	>	* return value;}</pre>
2669	>	*
2670	>	* except that the action is performed atomically.  If the
2671	>	* function returns {@code null} no mapping is recorded. If the
2672	>	* function itself throws an (unchecked) exception, the exception
2673	>	* is rethrown to its caller, and no mapping is recorded.  Some
2674	>	* attempted update operations on this map by other threads may be
2675	>	* blocked while computation is in progress, so the computation
2676	>	* should be short and simple, and must not attempt to update any
2677	>	* other mappings of this Map. The most appropriate usage is to
2678	>	* construct a new object serving as an initial mapped value, or
2679	>	* memoized result, as in:
2680	>	*
2681	>	*  <pre> {@code
2682	>	* map.computeIfAbsent(key, new Fun<K, V>() {
2683	>	*   public V map(K k) { return new Value(f(k)); }});}</pre>
2684	>	*
2685	>	* @param key key with which the specified value is to be associated
2686	>	* @param mappingFunction the function to compute a value
2687	>	* @return the current (existing or computed) value associated with
2688	>	*         the specified key, or null if the computed value is null
2689	>	* @throws NullPointerException if the specified key or mappingFunction
2690	>	*         is null
2691	>	* @throws IllegalStateException if the computation detectably
2692	>	*         attempts a recursive update to this map that would
2693	>	*         otherwise never complete
2694	>	* @throws RuntimeException or Error if the mappingFunction does so,
2695	>	*         in which case the mapping is left unestablished
2696	>	*/
2697	>	public V computeIfAbsent
2698	>	(K key, Fun<? super K, ? extends V> mappingFunction) {
2699	>	return internalComputeIfAbsent(key, mappingFunction);
2700	>	}
2701	>
2702	>	/**
2703	>	* If the given key is present, computes a new mapping value given a key and
2704	>	* its current mapped value. This is equivalent to
2705	>	*  <pre> {@code
2706	>	*   if (map.containsKey(key)) {
2707	>	*     value = remappingFunction.apply(key, map.get(key));
2708	>	*     if (value != null)
2709	>	*       map.put(key, value);
2710	>	*     else
2711	>	*       map.remove(key);
2712	>	*   }
2713	>	* }</pre>
2714	>	*
2715	>	* except that the action is performed atomically.  If the
2716	>	* function returns {@code null}, the mapping is removed.  If the
2717	>	* function itself throws an (unchecked) exception, the exception
2718	>	* is rethrown to its caller, and the current mapping is left
2719	>	* unchanged.  Some attempted update operations on this map by
2720	>	* other threads may be blocked while computation is in progress,
2721	>	* so the computation should be short and simple, and must not
2722	>	* attempt to update any other mappings of this Map. For example,
2723	>	* to either create or append new messages to a value mapping:
2724	>	*
2725	>	* @param key key with which the specified value is to be associated
2726	>	* @param remappingFunction the function to compute a value
2727	>	* @return the new value associated with the specified key, or null if none
2728	>	* @throws NullPointerException if the specified key or remappingFunction
2729	>	*         is null
2730	>	* @throws IllegalStateException if the computation detectably
2731	>	*         attempts a recursive update to this map that would
2732	>	*         otherwise never complete
2733	>	* @throws RuntimeException or Error if the remappingFunction does so,
2734	>	*         in which case the mapping is unchanged
2735	>	*/
2736	>	public V computeIfPresent
2737	>	(K key, BiFun<? super K, ? super V, ? extends V> remappingFunction) {
2738	>	return internalCompute(key, true, remappingFunction);
2739	>	}
2740	>
2741	>	/**
2742	>	* Computes a new mapping value given a key and
2743	>	* its current mapped value (or {@code null} if there is no current
2744	>	* mapping). This is equivalent to
2745	>	*  <pre> {@code
2746	>	*   value = remappingFunction.apply(key, map.get(key));
2747	>	*   if (value != null)
2748	>	*     map.put(key, value);
2749	>	*   else
2750	>	*     map.remove(key);
2751	>	* }</pre>
2752	>	*
2753	>	* except that the action is performed atomically.  If the
2754	>	* function returns {@code null}, the mapping is removed.  If the
2755	>	* function itself throws an (unchecked) exception, the exception
2756	>	* is rethrown to its caller, and the current mapping is left
2757	>	* unchanged.  Some attempted update operations on this map by
2758	>	* other threads may be blocked while computation is in progress,
2759	>	* so the computation should be short and simple, and must not
2760	>	* attempt to update any other mappings of this Map. For example,
2761	>	* to either create or append new messages to a value mapping:
2762	>	*
2763	>	* <pre> {@code
2764	>	* Map<Key, String> map = ...;
2765	>	* final String msg = ...;
2766	>	* map.compute(key, new BiFun<Key, String, String>() {
2767	>	*   public String apply(Key k, String v) {
2768	>	*    return (v == null) ? msg : v + msg;});}}</pre>
2769	>	*
2770	>	* @param key key with which the specified value is to be associated
2771	>	* @param remappingFunction the function to compute a value
2772	>	* @return the new value associated with the specified key, or null if none
2773	>	* @throws NullPointerException if the specified key or remappingFunction
2774	>	*         is null
2775	>	* @throws IllegalStateException if the computation detectably
2776	>	*         attempts a recursive update to this map that would
2777	>	*         otherwise never complete
2778	>	* @throws RuntimeException or Error if the remappingFunction does so,
2779	>	*         in which case the mapping is unchanged
2780	>	*/
2781	>	public V compute
2782	>	(K key, BiFun<? super K, ? super V, ? extends V> remappingFunction) {
2783	>	return internalCompute(key, false, remappingFunction);
2784	>	}
2785	>
2786	>	/**
2787	>	* If the specified key is not already associated
2788	>	* with a value, associate it with the given value.
2789	>	* Otherwise, replace the value with the results of
2790	>	* the given remapping function. This is equivalent to:
2791	>	*  <pre> {@code
2792	>	*   if (!map.containsKey(key))
2793	>	*     map.put(value);
2794	>	*   else {
2795	>	*     newValue = remappingFunction.apply(map.get(key), value);
2796	>	*     if (value != null)
2797	>	*       map.put(key, value);
2798	>	*     else
2799	>	*       map.remove(key);
2800	>	*   }
2801	>	* }</pre>
2802	>	* except that the action is performed atomically.  If the
2803	>	* function returns {@code null}, the mapping is removed.  If the
2804	>	* function itself throws an (unchecked) exception, the exception
2805	>	* is rethrown to its caller, and the current mapping is left
2806	>	* unchanged.  Some attempted update operations on this map by
2807	>	* other threads may be blocked while computation is in progress,
2808	>	* so the computation should be short and simple, and must not
2809	>	* attempt to update any other mappings of this Map.
2810	>	*/
2811	>	public V merge
2812	>	(K key, V value,
2813	>	BiFun<? super V, ? super V, ? extends V> remappingFunction) {
2814	>	return internalMerge(key, value, remappingFunction);
2815		}
2816
2817		/**
#	Line 910 | Line 2819 | public class ConcurrentHashMap<K, V> ext
2819		* This method does nothing if the key is not in the map.
2820		*
2821		* @param  key the key that needs to be removed
2822	<	* @return the previous value associated with <tt>key</tt>, or
2823	<	*         <tt>null</tt> if there was no mapping for <tt>key</tt>.
2822	>	* @return the previous value associated with {@code key}, or
2823	>	*         {@code null} if there was no mapping for {@code key}
2824		* @throws NullPointerException if the specified key is null
2825		*/
2826		public V remove(Object key) {
2827	<	int hash = hash(key);
919	<	return segmentFor(hash).remove(key, hash, null);
2827	>	return internalReplace(key, null, null);
2828		}
2829
2830		/**
#	Line 925 | Line 2833 | public class ConcurrentHashMap<K, V> ext
2833		* @throws NullPointerException if the specified key is null
2834		*/
2835		public boolean remove(Object key, Object value) {
2836	<	if (value == null)
929	<	return false;
930	<	int hash = hash(key);
931	<	return segmentFor(hash).remove(key, hash, value) != null;
2836	>	return value != null && internalReplace(key, null, value) != null;
2837		}
2838
2839		/**
#	Line 937 | Line 2842 | public class ConcurrentHashMap<K, V> ext
2842		* @throws NullPointerException if any of the arguments are null
2843		*/
2844		public boolean replace(K key, V oldValue, V newValue) {
2845	<	if (oldValue == null || newValue == null)
2845	>	if (key == null || oldValue == null || newValue == null)
2846		throw new NullPointerException();
2847	<	int hash = hash(key);
943	<	return segmentFor(hash).replace(key, hash, oldValue, newValue);
2847	>	return internalReplace(key, newValue, oldValue) != null;
2848		}
2849
2850		/**
2851		* {@inheritDoc}
2852		*
2853		* @return the previous value associated with the specified key,
2854	<	*         or <tt>null</tt> if there was no mapping for the key
2854	>	*         or {@code null} if there was no mapping for the key
2855		* @throws NullPointerException if the specified key or value is null
2856		*/
2857		public V replace(K key, V value) {
2858	<	if (value == null)
2858	>	if (key == null || value == null)
2859		throw new NullPointerException();
2860	<	int hash = hash(key);
957	<	return segmentFor(hash).replace(key, hash, value);
2860	>	return internalReplace(key, value, null);
2861		}
2862
2863		/**
2864		* Removes all of the mappings from this map.
2865		*/
2866		public void clear() {
2867	<	for (int i = 0; i < segments.length; ++i)
965	<	segments[i].clear();
2867	>	internalClear();
2868		}
2869
2870		/**
2871		* Returns a {@link Set} view of the keys contained in this map.
2872		* The set is backed by the map, so changes to the map are
2873	<	* 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.
2873	>	* reflected in the set, and vice-versa.
2874		*
2875	<	* <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.
2875	>	* @return the set view
2876		*/
2877	<	public Set<K> keySet() {
2878	<	Set<K> ks = keySet;
2879	<	return (ks != null) ? ks : (keySet = new KeySet());
2877	>	public KeySetView<K,V> keySet() {
2878	>	KeySetView<K,V> ks = keySet;
2879	>	return (ks != null) ? ks : (keySet = new KeySetView<K,V>(this, null));
2880	>	}
2881	>
2882	>	/**
2883	>	* Returns a {@link Set} view of the keys in this map, using the
2884	>	* given common mapped value for any additions (i.e., {@link
2885	>	* Collection#add} and {@link Collection#addAll}). This is of
2886	>	* course only appropriate if it is acceptable to use the same
2887	>	* value for all additions from this view.
2888	>	*
2889	>	* @param mappedValue the mapped value to use for any
2890	>	* additions.
2891	>	* @return the set view
2892	>	* @throws NullPointerException if the mappedValue is null
2893	>	*/
2894	>	public KeySetView<K,V> keySet(V mappedValue) {
2895	>	if (mappedValue == null)
2896	>	throw new NullPointerException();
2897	>	return new KeySetView<K,V>(this, mappedValue);
2898		}
2899
2900		/**
2901		* Returns a {@link Collection} view of the values contained in this map.
2902		* The collection is backed by the map, so changes to the map are
2903	<	* reflected in the collection, and vice-versa.  The collection
993	<	* supports element removal, which removes the corresponding
994	<	* mapping from this map, via the <tt>Iterator.remove</tt>,
995	<	* <tt>Collection.remove</tt>, <tt>removeAll</tt>,
996	<	* <tt>retainAll</tt>, and <tt>clear</tt> operations.  It does not
997	<	* support the <tt>add</tt> or <tt>addAll</tt> operations.
998	<	*
999	<	* <p>The view's <tt>iterator</tt> is a "weakly consistent" iterator
1000	<	* that will never throw {@link ConcurrentModificationException},
1001	<	* and guarantees to traverse elements as they existed upon
1002	<	* construction of the iterator, and may (but is not guaranteed to)
1003	<	* reflect any modifications subsequent to construction.
2903	>	* reflected in the collection, and vice-versa.
2904		*/
2905	<	public Collection<V> values() {
2906	<	Collection<V> vs = values;
2907	<	return (vs != null) ? vs : (values = new Values());
2905	>	public ValuesView<K,V> values() {
2906	>	ValuesView<K,V> vs = values;
2907	>	return (vs != null) ? vs : (values = new ValuesView<K,V>(this));
2908		}
2909
2910		/**
#	Line 1012 | Line 2912 | public class ConcurrentHashMap<K, V> ext
2912		* The set is backed by the map, so changes to the map are
2913		* reflected in the set, and vice-versa.  The set supports element
2914		* removal, which removes the corresponding mapping from the map,
2915	<	* via the <tt>Iterator.remove</tt>, <tt>Set.remove</tt>,
2916	<	* <tt>removeAll</tt>, <tt>retainAll</tt>, and <tt>clear</tt>
2917	<	* operations.  It does not support the <tt>add</tt> or
2918	<	* <tt>addAll</tt> operations.
2915	>	* via the {@code Iterator.remove}, {@code Set.remove},
2916	>	* {@code removeAll}, {@code retainAll}, and {@code clear}
2917	>	* operations.  It does not support the {@code add} or
2918	>	* {@code addAll} operations.
2919		*
2920	<	* <p>The view's <tt>iterator</tt> is a "weakly consistent" iterator
2920	>	* <p>The view's {@code iterator} is a "weakly consistent" iterator
2921		* that will never throw {@link ConcurrentModificationException},
2922		* and guarantees to traverse elements as they existed upon
2923		* construction of the iterator, and may (but is not guaranteed to)
2924		* reflect any modifications subsequent to construction.
2925		*/
2926		public Set<Map.Entry<K,V>> entrySet() {
2927	<	Set<Map.Entry<K,V>> es = entrySet;
2928	<	return (es != null) ? es : (entrySet = new EntrySet());
2927	>	EntrySetView<K,V> es = entrySet;
2928	>	return (es != null) ? es : (entrySet = new EntrySetView<K,V>(this));
2929		}
2930
2931		/**
2932		* Returns an enumeration of the keys in this table.
2933		*
2934		* @return an enumeration of the keys in this table
2935	<	* @see #keySet
2935	>	* @see #keySet()
2936		*/
2937		public Enumeration<K> keys() {
2938	<	return new KeyIterator();
2938	>	return new KeyIterator<K,V>(this);
2939		}
2940
2941		/**
2942		* Returns an enumeration of the values in this table.
2943		*
2944		* @return an enumeration of the values in this table
2945	<	* @see #values
2945	>	* @see #values()
2946		*/
2947		public Enumeration<V> elements() {
2948	<	return new ValueIterator();
2948	>	return new ValueIterator<K,V>(this);
2949		}
2950
2951	<	/* ---------------- Iterator Support -------------- */
2952	<
2953	<	abstract class HashIterator {
2954	<	int nextSegmentIndex;
2955	<	int nextTableIndex;
2956	<	HashEntry<K,V>[] currentTable;
2957	<	HashEntry<K, V> nextEntry;
2958	<	HashEntry<K, V> lastReturned;
2951	>	/**
2952	>	* Returns a partitionable iterator of the keys in this map.
2953	>	*
2954	>	* @return a partitionable iterator of the keys in this map
2955	>	*/
2956	>	public Spliterator<K> keySpliterator() {
2957	>	return new KeyIterator<K,V>(this);
2958	>	}
2959
2960	<	HashIterator() {
2961	<	nextSegmentIndex = segments.length - 1;
2962	<	nextTableIndex = -1;
2963	<	advance();
2964	<	}
2960	>	/**
2961	>	* Returns a partitionable iterator of the values in this map.
2962	>	*
2963	>	* @return a partitionable iterator of the values in this map
2964	>	*/
2965	>	public Spliterator<V> valueSpliterator() {
2966	>	return new ValueIterator<K,V>(this);
2967	>	}
2968
2969	<	public boolean hasMoreElements() { return hasNext(); }
2969	>	/**
2970	>	* Returns a partitionable iterator of the entries in this map.
2971	>	*
2972	>	* @return a partitionable iterator of the entries in this map
2973	>	*/
2974	>	public Spliterator<Map.Entry<K,V>> entrySpliterator() {
2975	>	return new EntryIterator<K,V>(this);
2976	>	}
2977
2978	<	final void advance() {
2979	<	if (nextEntry != null && (nextEntry = nextEntry.next) != null)
2980	<	return;
2978	>	/**
2979	>	* Returns the hash code value for this {@link Map}, i.e.,
2980	>	* the sum of, for each key-value pair in the map,
2981	>	* {@code key.hashCode() ^ value.hashCode()}.
2982	>	*
2983	>	* @return the hash code value for this map
2984	>	*/
2985	>	public int hashCode() {
2986	>	int h = 0;
2987	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
2988	>	Object v;
2989	>	while ((v = it.advance()) != null) {
2990	>	h += it.nextKey.hashCode() ^ v.hashCode();
2991	>	}
2992	>	return h;
2993	>	}
2994
2995	<	while (nextTableIndex >= 0) {
2996	<	if ( (nextEntry = currentTable[nextTableIndex--]) != null)
2997	<	return;
2995	>	/**
2996	>	* Returns a string representation of this map.  The string
2997	>	* representation consists of a list of key-value mappings (in no
2998	>	* particular order) enclosed in braces ("{@code {}}").  Adjacent
2999	>	* mappings are separated by the characters {@code ", "} (comma
3000	>	* and space).  Each key-value mapping is rendered as the key
3001	>	* followed by an equals sign ("{@code =}") followed by the
3002	>	* associated value.
3003	>	*
3004	>	* @return a string representation of this map
3005	>	*/
3006	>	public String toString() {
3007	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
3008	>	StringBuilder sb = new StringBuilder();
3009	>	sb.append('{');
3010	>	Object v;
3011	>	if ((v = it.advance()) != null) {
3012	>	for (;;) {
3013	>	Object k = it.nextKey;
3014	>	sb.append(k == this ? "(this Map)" : k);
3015	>	sb.append('=');
3016	>	sb.append(v == this ? "(this Map)" : v);
3017	>	if ((v = it.advance()) == null)
3018	>	break;
3019	>	sb.append(',').append(' ');
3020		}
3021	+	}
3022	+	return sb.append('}').toString();
3023	+	}
3024
3025	<	while (nextSegmentIndex >= 0) {
3026	<	Segment<K,V> seg = segments[nextSegmentIndex--];
3027	<	if (seg.count != 0) {
3028	<	currentTable = seg.table;
3029	<	for (int j = currentTable.length - 1; j >= 0; --j) {
3030	<	if ( (nextEntry = currentTable[j]) != null) {
3031	<	nextTableIndex = j - 1;
3032	<	return;
3033	<	}
3034	<	}
3035	<	}
3025	>	/**
3026	>	* Compares the specified object with this map for equality.
3027	>	* Returns {@code true} if the given object is a map with the same
3028	>	* mappings as this map.  This operation may return misleading
3029	>	* results if either map is concurrently modified during execution
3030	>	* of this method.
3031	>	*
3032	>	* @param o object to be compared for equality with this map
3033	>	* @return {@code true} if the specified object is equal to this map
3034	>	*/
3035	>	public boolean equals(Object o) {
3036	>	if (o != this) {
3037	>	if (!(o instanceof Map))
3038	>	return false;
3039	>	Map<?,?> m = (Map<?,?>) o;
3040	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
3041	>	Object val;
3042	>	while ((val = it.advance()) != null) {
3043	>	Object v = m.get(it.nextKey);
3044	>	if (v == null || (v != val && !v.equals(val)))
3045	>	return false;
3046	>	}
3047	>	for (Map.Entry<?,?> e : m.entrySet()) {
3048	>	Object mk, mv, v;
3049	>	if ((mk = e.getKey()) == null ||
3050	>	(mv = e.getValue()) == null ||
3051	>	(v = internalGet(mk)) == null ||
3052	>	(mv != v && !mv.equals(v)))
3053	>	return false;
3054		}
3055		}
3056	+	return true;
3057	+	}
3058
3059	<	public boolean hasNext() { return nextEntry != null; }
3059	>	/* ----------------Iterators -------------- */
3060
3061	<	HashEntry<K,V> nextEntry() {
3062	<	if (nextEntry == null)
3061	>	@SuppressWarnings("serial") static final class KeyIterator<K,V>
3062	>	extends Traverser<K,V,Object>
3063	>	implements Spliterator<K>, Enumeration<K> {
3064	>	KeyIterator(ConcurrentHashMap<K, V> map) { super(map); }
3065	>	KeyIterator(ConcurrentHashMap<K, V> map, Traverser<K,V,Object> it) {
3066	>	super(map, it, -1);
3067	>	}
3068	>	public KeyIterator<K,V> split() {
3069	>	if (nextKey != null)
3070	>	throw new IllegalStateException();
3071	>	return new KeyIterator<K,V>(map, this);
3072	>	}
3073	>	@SuppressWarnings("unchecked") public final K next() {
3074	>	if (nextVal == null && advance() == null)
3075		throw new NoSuchElementException();
3076	<	lastReturned = nextEntry;
3077	<	advance();
3078	<	return lastReturned;
3076	>	Object k = nextKey;
3077	>	nextVal = null;
3078	>	return (K) k;
3079		}
3080
3081	<	public void remove() {
3082	<	if (lastReturned == null)
3081	>	public final K nextElement() { return next(); }
3082	>	}
3083	>
3084	>	@SuppressWarnings("serial") static final class ValueIterator<K,V>
3085	>	extends Traverser<K,V,Object>
3086	>	implements Spliterator<V>, Enumeration<V> {
3087	>	ValueIterator(ConcurrentHashMap<K, V> map) { super(map); }
3088	>	ValueIterator(ConcurrentHashMap<K, V> map, Traverser<K,V,Object> it) {
3089	>	super(map, it, -1);
3090	>	}
3091	>	public ValueIterator<K,V> split() {
3092	>	if (nextKey != null)
3093		throw new IllegalStateException();
3094	<	ConcurrentHashMap.this.remove(lastReturned.key);
1105	<	lastReturned = null;
3094	>	return new ValueIterator<K,V>(map, this);
3095		}
1107	–	}
3096
3097	<	final class KeyIterator
3098	<	extends HashIterator
3099	<	implements Iterator<K>, Enumeration<K>
3100	<	{
3101	<	public K next()        { return super.nextEntry().key; }
3102	<	public K nextElement() { return super.nextEntry().key; }
3097	>	@SuppressWarnings("unchecked") public final V next() {
3098	>	Object v;
3099	>	if ((v = nextVal) == null && (v = advance()) == null)
3100	>	throw new NoSuchElementException();
3101	>	nextVal = null;
3102	>	return (V) v;
3103	>	}
3104	>
3105	>	public final V nextElement() { return next(); }
3106		}
3107
3108	<	final class ValueIterator
3109	<	extends HashIterator
3110	<	implements Iterator<V>, Enumeration<V>
3111	<	{
3112	<	public V next()        { return super.nextEntry().value; }
3113	<	public V nextElement() { return super.nextEntry().value; }
3108	>	@SuppressWarnings("serial") static final class EntryIterator<K,V>
3109	>	extends Traverser<K,V,Object>
3110	>	implements Spliterator<Map.Entry<K,V>> {
3111	>	EntryIterator(ConcurrentHashMap<K, V> map) { super(map); }
3112	>	EntryIterator(ConcurrentHashMap<K, V> map, Traverser<K,V,Object> it) {
3113	>	super(map, it, -1);
3114	>	}
3115	>	public EntryIterator<K,V> split() {
3116	>	if (nextKey != null)
3117	>	throw new IllegalStateException();
3118	>	return new EntryIterator<K,V>(map, this);
3119	>	}
3120	>
3121	>	@SuppressWarnings("unchecked") public final Map.Entry<K,V> next() {
3122	>	Object v;
3123	>	if ((v = nextVal) == null && (v = advance()) == null)
3124	>	throw new NoSuchElementException();
3125	>	Object k = nextKey;
3126	>	nextVal = null;
3127	>	return new MapEntry<K,V>((K)k, (V)v, map);
3128	>	}
3129		}
3130
3131		/**
3132	<	* Custom Entry class used by EntryIterator.next(), that relays
1127	<	* setValue changes to the underlying map.
3132	>	* Exported Entry for iterators
3133		*/
3134	<	final class WriteThroughEntry
3135	<	extends AbstractMap.SimpleEntry<K,V>
3136	<	{
3137	<	WriteThroughEntry(K k, V v) {
3138	<	super(k,v);
3134	>	static final class MapEntry<K,V> implements Map.Entry<K, V> {
3135	>	final K key; // non-null
3136	>	V val;       // non-null
3137	>	final ConcurrentHashMap<K, V> map;
3138	>	MapEntry(K key, V val, ConcurrentHashMap<K, V> map) {
3139	>	this.key = key;
3140	>	this.val = val;
3141	>	this.map = map;
3142	>	}
3143	>	public final K getKey()       { return key; }
3144	>	public final V getValue()     { return val; }
3145	>	public final int hashCode()   { return key.hashCode() ^ val.hashCode(); }
3146	>	public final String toString(){ return key + "=" + val; }
3147	>
3148	>	public final boolean equals(Object o) {
3149	>	Object k, v; Map.Entry<?,?> e;
3150	>	return ((o instanceof Map.Entry) &&
3151	>	(k = (e = (Map.Entry<?,?>)o).getKey()) != null &&
3152	>	(v = e.getValue()) != null &&
3153	>	(k == key || k.equals(key)) &&
3154	>	(v == val || v.equals(val)));
3155		}
3156
3157		/**
3158	<	* Set our entry's value and write through to the map. The
3159	<	* value to return is somewhat arbitrary here. Since a
3160	<	* WriteThroughEntry does not necessarily track asynchronous
3161	<	* changes, the most recent "previous" value could be
3162	<	* different from what we return (or could even have been
3163	<	* removed in which case the put will re-establish). We do not
1143	<	* and cannot guarantee more.
3158	>	* Sets our entry's value and writes through to the map. The
3159	>	* value to return is somewhat arbitrary here. Since we do not
3160	>	* necessarily track asynchronous changes, the most recent
3161	>	* "previous" value could be different from what we return (or
3162	>	* could even have been removed in which case the put will
3163	>	* re-establish). We do not and cannot guarantee more.
3164		*/
3165	<	public V setValue(V value) {
3165	>	public final V setValue(V value) {
3166		if (value == null) throw new NullPointerException();
3167	<	V v = super.setValue(value);
3168	<	ConcurrentHashMap.this.put(getKey(), value);
3167	>	V v = val;
3168	>	val = value;
3169	>	map.put(key, value);
3170		return v;
3171		}
3172		}
3173
3174	<	final class EntryIterator
3175	<	extends HashIterator
3176	<	implements Iterator<Entry<K,V>>
3177	<	{
3178	<	public Map.Entry<K,V> next() {
3179	<	HashEntry<K,V> e = super.nextEntry();
3180	<	return new WriteThroughEntry(e.key, e.value);
3174	>	/**
3175	>	* Returns exportable snapshot entry for the given key and value
3176	>	* when write-through can't or shouldn't be used.
3177	>	*/
3178	>	static <K,V> AbstractMap.SimpleEntry<K,V> entryFor(K k, V v) {
3179	>	return new AbstractMap.SimpleEntry<K,V>(k, v);
3180	>	}
3181	>
3182	>	/* ---------------- Serialization Support -------------- */
3183	>
3184	>	/**
3185	>	* Stripped-down version of helper class used in previous version,
3186	>	* declared for the sake of serialization compatibility
3187	>	*/
3188	>	static class Segment<K,V> implements Serializable {
3189	>	private static final long serialVersionUID = 2249069246763182397L;
3190	>	final float loadFactor;
3191	>	Segment(float lf) { this.loadFactor = lf; }
3192	>	}
3193	>
3194	>	/**
3195	>	* Saves the state of the {@code ConcurrentHashMap} instance to a
3196	>	* stream (i.e., serializes it).
3197	>	* @param s the stream
3198	>	* @serialData
3199	>	* the key (Object) and value (Object)
3200	>	* for each key-value mapping, followed by a null pair.
3201	>	* The key-value mappings are emitted in no particular order.
3202	>	*/
3203	>	@SuppressWarnings("unchecked") private void writeObject
3204	>	(java.io.ObjectOutputStream s)
3205	>	throws java.io.IOException {
3206	>	if (segments == null) { // for serialization compatibility
3207	>	segments = (Segment<K,V>[])
3208	>	new Segment<?,?>[DEFAULT_CONCURRENCY_LEVEL];
3209	>	for (int i = 0; i < segments.length; ++i)
3210	>	segments[i] = new Segment<K,V>(LOAD_FACTOR);
3211		}
3212	+	s.defaultWriteObject();
3213	+	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
3214	+	Object v;
3215	+	while ((v = it.advance()) != null) {
3216	+	s.writeObject(it.nextKey);
3217	+	s.writeObject(v);
3218	+	}
3219	+	s.writeObject(null);
3220	+	s.writeObject(null);
3221	+	segments = null; // throw away
3222		}
3223
3224	<	final class KeySet extends AbstractSet<K> {
3225	<	public Iterator<K> iterator() {
3226	<	return new KeyIterator();
3224	>	/**
3225	>	* Reconstitutes the instance from a stream (that is, deserializes it).
3226	>	* @param s the stream
3227	>	*/
3228	>	@SuppressWarnings("unchecked") private void readObject
3229	>	(java.io.ObjectInputStream s)
3230	>	throws java.io.IOException, ClassNotFoundException {
3231	>	s.defaultReadObject();
3232	>	this.segments = null; // unneeded
3233	>
3234	>	// Create all nodes, then place in table once size is known
3235	>	long size = 0L;
3236	>	Node p = null;
3237	>	for (;;) {
3238	>	K k = (K) s.readObject();
3239	>	V v = (V) s.readObject();
3240	>	if (k != null && v != null) {
3241	>	int h = spread(k.hashCode());
3242	>	p = new Node(h, k, v, p);
3243	>	++size;
3244	>	}
3245	>	else
3246	>	break;
3247		}
3248	<	public int size() {
3249	<	return ConcurrentHashMap.this.size();
3248	>	if (p != null) {
3249	>	boolean init = false;
3250	>	int n;
3251	>	if (size >= (long)(MAXIMUM_CAPACITY >>> 1))
3252	>	n = MAXIMUM_CAPACITY;
3253	>	else {
3254	>	int sz = (int)size;
3255	>	n = tableSizeFor(sz + (sz >>> 1) + 1);
3256	>	}
3257	>	int sc = sizeCtl;
3258	>	boolean collide = false;
3259	>	if (n > sc &&
3260	>	U.compareAndSwapInt(this, SIZECTL, sc, -1)) {
3261	>	try {
3262	>	if (table == null) {
3263	>	init = true;
3264	>	Node[] tab = new Node[n];
3265	>	int mask = n - 1;
3266	>	while (p != null) {
3267	>	int j = p.hash & mask;
3268	>	Node next = p.next;
3269	>	Node q = p.next = tabAt(tab, j);
3270	>	setTabAt(tab, j, p);
3271	>	if (!collide && q != null && q.hash == p.hash)
3272	>	collide = true;
3273	>	p = next;
3274	>	}
3275	>	table = tab;
3276	>	addCount(size, -1);
3277	>	sc = n - (n >>> 2);
3278	>	}
3279	>	} finally {
3280	>	sizeCtl = sc;
3281	>	}
3282	>	if (collide) { // rescan and convert to TreeBins
3283	>	Node[] tab = table;
3284	>	for (int i = 0; i < tab.length; ++i) {
3285	>	int c = 0;
3286	>	for (Node e = tabAt(tab, i); e != null; e = e.next) {
3287	>	if (++c > TREE_THRESHOLD &&
3288	>	(e.key instanceof Comparable)) {
3289	>	replaceWithTreeBin(tab, i, e.key);
3290	>	break;
3291	>	}
3292	>	}
3293	>	}
3294	>	}
3295	>	}
3296	>	if (!init) { // Can only happen if unsafely published.
3297	>	while (p != null) {
3298	>	internalPut((K)p.key, (V)p.val, false);
3299	>	p = p.next;
3300	>	}
3301	>	}
3302		}
3303	<	public boolean contains(Object o) {
3304	<	return ConcurrentHashMap.this.containsKey(o);
3303	>	}
3304	>
3305	>	// -------------------------------------------------------
3306	>
3307	>	// Sams
3308	>	/** Interface describing a void action of one argument */
3309	>	public interface Action<A> { void apply(A a); }
3310	>	/** Interface describing a void action of two arguments */
3311	>	public interface BiAction<A,B> { void apply(A a, B b); }
3312	>	/** Interface describing a function of one argument */
3313	>	public interface Fun<A,T> { T apply(A a); }
3314	>	/** Interface describing a function of two arguments */
3315	>	public interface BiFun<A,B,T> { T apply(A a, B b); }
3316	>	/** Interface describing a function of no arguments */
3317	>	public interface Generator<T> { T apply(); }
3318	>	/** Interface describing a function mapping its argument to a double */
3319	>	public interface ObjectToDouble<A> { double apply(A a); }
3320	>	/** Interface describing a function mapping its argument to a long */
3321	>	public interface ObjectToLong<A> { long apply(A a); }
3322	>	/** Interface describing a function mapping its argument to an int */
3323	>	public interface ObjectToInt<A> {int apply(A a); }
3324	>	/** Interface describing a function mapping two arguments to a double */
3325	>	public interface ObjectByObjectToDouble<A,B> { double apply(A a, B b); }
3326	>	/** Interface describing a function mapping two arguments to a long */
3327	>	public interface ObjectByObjectToLong<A,B> { long apply(A a, B b); }
3328	>	/** Interface describing a function mapping two arguments to an int */
3329	>	public interface ObjectByObjectToInt<A,B> {int apply(A a, B b); }
3330	>	/** Interface describing a function mapping a double to a double */
3331	>	public interface DoubleToDouble { double apply(double a); }
3332	>	/** Interface describing a function mapping a long to a long */
3333	>	public interface LongToLong { long apply(long a); }
3334	>	/** Interface describing a function mapping an int to an int */
3335	>	public interface IntToInt { int apply(int a); }
3336	>	/** Interface describing a function mapping two doubles to a double */
3337	>	public interface DoubleByDoubleToDouble { double apply(double a, double b); }
3338	>	/** Interface describing a function mapping two longs to a long */
3339	>	public interface LongByLongToLong { long apply(long a, long b); }
3340	>	/** Interface describing a function mapping two ints to an int */
3341	>	public interface IntByIntToInt { int apply(int a, int b); }
3342	>
3343	>
3344	>	// -------------------------------------------------------
3345	>
3346	>	/**
3347	>	* Performs the given action for each (key, value).
3348	>	*
3349	>	* @param action the action
3350	>	*/
3351	>	public void forEach(BiAction<K,V> action) {
3352	>	ForkJoinTasks.forEach
3353	>	(this, action).invoke();
3354	>	}
3355	>
3356	>	/**
3357	>	* Performs the given action for each non-null transformation
3358	>	* of each (key, value).
3359	>	*
3360	>	* @param transformer a function returning the transformation
3361	>	* for an element, or null of there is no transformation (in
3362	>	* which case the action is not applied).
3363	>	* @param action the action
3364	>	*/
3365	>	public <U> void forEach(BiFun<? super K, ? super V, ? extends U> transformer,
3366	>	Action<U> action) {
3367	>	ForkJoinTasks.forEach
3368	>	(this, transformer, action).invoke();
3369	>	}
3370	>
3371	>	/**
3372	>	* Returns a non-null result from applying the given search
3373	>	* function on each (key, value), or null if none.  Upon
3374	>	* success, further element processing is suppressed and the
3375	>	* results of any other parallel invocations of the search
3376	>	* function are ignored.
3377	>	*
3378	>	* @param searchFunction a function returning a non-null
3379	>	* result on success, else null
3380	>	* @return a non-null result from applying the given search
3381	>	* function on each (key, value), or null if none
3382	>	*/
3383	>	public <U> U search(BiFun<? super K, ? super V, ? extends U> searchFunction) {
3384	>	return ForkJoinTasks.search
3385	>	(this, searchFunction).invoke();
3386	>	}
3387	>
3388	>	/**
3389	>	* Returns the result of accumulating the given transformation
3390	>	* of all (key, value) pairs using the given reducer to
3391	>	* combine values, or null if none.
3392	>	*
3393	>	* @param transformer a function returning the transformation
3394	>	* for an element, or null of there is no transformation (in
3395	>	* which case it is not combined).
3396	>	* @param reducer a commutative associative combining function
3397	>	* @return the result of accumulating the given transformation
3398	>	* of all (key, value) pairs
3399	>	*/
3400	>	public <U> U reduce(BiFun<? super K, ? super V, ? extends U> transformer,
3401	>	BiFun<? super U, ? super U, ? extends U> reducer) {
3402	>	return ForkJoinTasks.reduce
3403	>	(this, transformer, reducer).invoke();
3404	>	}
3405	>
3406	>	/**
3407	>	* Returns the result of accumulating the given transformation
3408	>	* of all (key, value) pairs using the given reducer to
3409	>	* combine values, and the given basis as an identity value.
3410	>	*
3411	>	* @param transformer a function returning the transformation
3412	>	* for an element
3413	>	* @param basis the identity (initial default value) for the reduction
3414	>	* @param reducer a commutative associative combining function
3415	>	* @return the result of accumulating the given transformation
3416	>	* of all (key, value) pairs
3417	>	*/
3418	>	public double reduceToDouble(ObjectByObjectToDouble<? super K, ? super V> transformer,
3419	>	double basis,
3420	>	DoubleByDoubleToDouble reducer) {
3421	>	return ForkJoinTasks.reduceToDouble
3422	>	(this, transformer, basis, reducer).invoke();
3423	>	}
3424	>
3425	>	/**
3426	>	* Returns the result of accumulating the given transformation
3427	>	* of all (key, value) pairs using the given reducer to
3428	>	* combine values, and the given basis as an identity value.
3429	>	*
3430	>	* @param transformer a function returning the transformation
3431	>	* for an element
3432	>	* @param basis the identity (initial default value) for the reduction
3433	>	* @param reducer a commutative associative combining function
3434	>	* @return the result of accumulating the given transformation
3435	>	* of all (key, value) pairs
3436	>	*/
3437	>	public long reduceToLong(ObjectByObjectToLong<? super K, ? super V> transformer,
3438	>	long basis,
3439	>	LongByLongToLong reducer) {
3440	>	return ForkJoinTasks.reduceToLong
3441	>	(this, transformer, basis, reducer).invoke();
3442	>	}
3443	>
3444	>	/**
3445	>	* Returns the result of accumulating the given transformation
3446	>	* of all (key, value) pairs using the given reducer to
3447	>	* combine values, and the given basis as an identity value.
3448	>	*
3449	>	* @param transformer a function returning the transformation
3450	>	* for an element
3451	>	* @param basis the identity (initial default value) for the reduction
3452	>	* @param reducer a commutative associative combining function
3453	>	* @return the result of accumulating the given transformation
3454	>	* of all (key, value) pairs
3455	>	*/
3456	>	public int reduceToInt(ObjectByObjectToInt<? super K, ? super V> transformer,
3457	>	int basis,
3458	>	IntByIntToInt reducer) {
3459	>	return ForkJoinTasks.reduceToInt
3460	>	(this, transformer, basis, reducer).invoke();
3461	>	}
3462	>
3463	>	/**
3464	>	* Performs the given action for each key.
3465	>	*
3466	>	* @param action the action
3467	>	*/
3468	>	public void forEachKey(Action<K> action) {
3469	>	ForkJoinTasks.forEachKey
3470	>	(this, action).invoke();
3471	>	}
3472	>
3473	>	/**
3474	>	* Performs the given action for each non-null transformation
3475	>	* of each key.
3476	>	*
3477	>	* @param transformer a function returning the transformation
3478	>	* for an element, or null of there is no transformation (in
3479	>	* which case the action is not applied).
3480	>	* @param action the action
3481	>	*/
3482	>	public <U> void forEachKey(Fun<? super K, ? extends U> transformer,
3483	>	Action<U> action) {
3484	>	ForkJoinTasks.forEachKey
3485	>	(this, transformer, action).invoke();
3486	>	}
3487	>
3488	>	/**
3489	>	* Returns a non-null result from applying the given search
3490	>	* function on each key, or null if none. Upon success,
3491	>	* further element processing is suppressed and the results of
3492	>	* any other parallel invocations of the search function are
3493	>	* ignored.
3494	>	*
3495	>	* @param searchFunction a function returning a non-null
3496	>	* result on success, else null
3497	>	* @return a non-null result from applying the given search
3498	>	* function on each key, or null if none
3499	>	*/
3500	>	public <U> U searchKeys(Fun<? super K, ? extends U> searchFunction) {
3501	>	return ForkJoinTasks.searchKeys
3502	>	(this, searchFunction).invoke();
3503	>	}
3504	>
3505	>	/**
3506	>	* Returns the result of accumulating all keys using the given
3507	>	* reducer to combine values, or null if none.
3508	>	*
3509	>	* @param reducer a commutative associative combining function
3510	>	* @return the result of accumulating all keys using the given
3511	>	* reducer to combine values, or null if none
3512	>	*/
3513	>	public K reduceKeys(BiFun<? super K, ? super K, ? extends K> reducer) {
3514	>	return ForkJoinTasks.reduceKeys
3515	>	(this, reducer).invoke();
3516	>	}
3517	>
3518	>	/**
3519	>	* Returns the result of accumulating the given transformation
3520	>	* of all keys using the given reducer to combine values, or
3521	>	* null if none.
3522	>	*
3523	>	* @param transformer a function returning the transformation
3524	>	* for an element, or null of there is no transformation (in
3525	>	* which case it is not combined).
3526	>	* @param reducer a commutative associative combining function
3527	>	* @return the result of accumulating the given transformation
3528	>	* of all keys
3529	>	*/
3530	>	public <U> U reduceKeys(Fun<? super K, ? extends U> transformer,
3531	>	BiFun<? super U, ? super U, ? extends U> reducer) {
3532	>	return ForkJoinTasks.reduceKeys
3533	>	(this, transformer, reducer).invoke();
3534	>	}
3535	>
3536	>	/**
3537	>	* Returns the result of accumulating the given transformation
3538	>	* of all keys using the given reducer to combine values, and
3539	>	* the given basis as an identity value.
3540	>	*
3541	>	* @param transformer a function returning the transformation
3542	>	* for an element
3543	>	* @param basis the identity (initial default value) for the reduction
3544	>	* @param reducer a commutative associative combining function
3545	>	* @return  the result of accumulating the given transformation
3546	>	* of all keys
3547	>	*/
3548	>	public double reduceKeysToDouble(ObjectToDouble<? super K> transformer,
3549	>	double basis,
3550	>	DoubleByDoubleToDouble reducer) {
3551	>	return ForkJoinTasks.reduceKeysToDouble
3552	>	(this, transformer, basis, reducer).invoke();
3553	>	}
3554	>
3555	>	/**
3556	>	* Returns the result of accumulating the given transformation
3557	>	* of all keys using the given reducer to combine values, and
3558	>	* the given basis as an identity value.
3559	>	*
3560	>	* @param transformer a function returning the transformation
3561	>	* for an element
3562	>	* @param basis the identity (initial default value) for the reduction
3563	>	* @param reducer a commutative associative combining function
3564	>	* @return the result of accumulating the given transformation
3565	>	* of all keys
3566	>	*/
3567	>	public long reduceKeysToLong(ObjectToLong<? super K> transformer,
3568	>	long basis,
3569	>	LongByLongToLong reducer) {
3570	>	return ForkJoinTasks.reduceKeysToLong
3571	>	(this, transformer, basis, reducer).invoke();
3572	>	}
3573	>
3574	>	/**
3575	>	* Returns the result of accumulating the given transformation
3576	>	* of all keys using the given reducer to combine values, and
3577	>	* the given basis as an identity value.
3578	>	*
3579	>	* @param transformer a function returning the transformation
3580	>	* for an element
3581	>	* @param basis the identity (initial default value) for the reduction
3582	>	* @param reducer a commutative associative combining function
3583	>	* @return the result of accumulating the given transformation
3584	>	* of all keys
3585	>	*/
3586	>	public int reduceKeysToInt(ObjectToInt<? super K> transformer,
3587	>	int basis,
3588	>	IntByIntToInt reducer) {
3589	>	return ForkJoinTasks.reduceKeysToInt
3590	>	(this, transformer, basis, reducer).invoke();
3591	>	}
3592	>
3593	>	/**
3594	>	* Performs the given action for each value.
3595	>	*
3596	>	* @param action the action
3597	>	*/
3598	>	public void forEachValue(Action<V> action) {
3599	>	ForkJoinTasks.forEachValue
3600	>	(this, action).invoke();
3601	>	}
3602	>
3603	>	/**
3604	>	* Performs the given action for each non-null transformation
3605	>	* of each value.
3606	>	*
3607	>	* @param transformer a function returning the transformation
3608	>	* for an element, or null of there is no transformation (in
3609	>	* which case the action is not applied).
3610	>	*/
3611	>	public <U> void forEachValue(Fun<? super V, ? extends U> transformer,
3612	>	Action<U> action) {
3613	>	ForkJoinTasks.forEachValue
3614	>	(this, transformer, action).invoke();
3615	>	}
3616	>
3617	>	/**
3618	>	* Returns a non-null result from applying the given search
3619	>	* function on each value, or null if none.  Upon success,
3620	>	* further element processing is suppressed and the results of
3621	>	* any other parallel invocations of the search function are
3622	>	* ignored.
3623	>	*
3624	>	* @param searchFunction a function returning a non-null
3625	>	* result on success, else null
3626	>	* @return a non-null result from applying the given search
3627	>	* function on each value, or null if none
3628	>	*
3629	>	*/
3630	>	public <U> U searchValues(Fun<? super V, ? extends U> searchFunction) {
3631	>	return ForkJoinTasks.searchValues
3632	>	(this, searchFunction).invoke();
3633	>	}
3634	>
3635	>	/**
3636	>	* Returns the result of accumulating all values using the
3637	>	* given reducer to combine values, or null if none.
3638	>	*
3639	>	* @param reducer a commutative associative combining function
3640	>	* @return  the result of accumulating all values
3641	>	*/
3642	>	public V reduceValues(BiFun<? super V, ? super V, ? extends V> reducer) {
3643	>	return ForkJoinTasks.reduceValues
3644	>	(this, reducer).invoke();
3645	>	}
3646	>
3647	>	/**
3648	>	* Returns the result of accumulating the given transformation
3649	>	* of all values using the given reducer to combine values, or
3650	>	* null if none.
3651	>	*
3652	>	* @param transformer a function returning the transformation
3653	>	* for an element, or null of there is no transformation (in
3654	>	* which case it is not combined).
3655	>	* @param reducer a commutative associative combining function
3656	>	* @return the result of accumulating the given transformation
3657	>	* of all values
3658	>	*/
3659	>	public <U> U reduceValues(Fun<? super V, ? extends U> transformer,
3660	>	BiFun<? super U, ? super U, ? extends U> reducer) {
3661	>	return ForkJoinTasks.reduceValues
3662	>	(this, transformer, reducer).invoke();
3663	>	}
3664	>
3665	>	/**
3666	>	* Returns the result of accumulating the given transformation
3667	>	* of all values using the given reducer to combine values,
3668	>	* and the given basis as an identity value.
3669	>	*
3670	>	* @param transformer a function returning the transformation
3671	>	* for an element
3672	>	* @param basis the identity (initial default value) for the reduction
3673	>	* @param reducer a commutative associative combining function
3674	>	* @return the result of accumulating the given transformation
3675	>	* of all values
3676	>	*/
3677	>	public double reduceValuesToDouble(ObjectToDouble<? super V> transformer,
3678	>	double basis,
3679	>	DoubleByDoubleToDouble reducer) {
3680	>	return ForkJoinTasks.reduceValuesToDouble
3681	>	(this, transformer, basis, reducer).invoke();
3682	>	}
3683	>
3684	>	/**
3685	>	* Returns the result of accumulating the given transformation
3686	>	* of all values using the given reducer to combine values,
3687	>	* and the given basis as an identity value.
3688	>	*
3689	>	* @param transformer a function returning the transformation
3690	>	* for an element
3691	>	* @param basis the identity (initial default value) for the reduction
3692	>	* @param reducer a commutative associative combining function
3693	>	* @return the result of accumulating the given transformation
3694	>	* of all values
3695	>	*/
3696	>	public long reduceValuesToLong(ObjectToLong<? super V> transformer,
3697	>	long basis,
3698	>	LongByLongToLong reducer) {
3699	>	return ForkJoinTasks.reduceValuesToLong
3700	>	(this, transformer, basis, reducer).invoke();
3701	>	}
3702	>
3703	>	/**
3704	>	* Returns the result of accumulating the given transformation
3705	>	* of all values using the given reducer to combine values,
3706	>	* and the given basis as an identity value.
3707	>	*
3708	>	* @param transformer a function returning the transformation
3709	>	* for an element
3710	>	* @param basis the identity (initial default value) for the reduction
3711	>	* @param reducer a commutative associative combining function
3712	>	* @return the result of accumulating the given transformation
3713	>	* of all values
3714	>	*/
3715	>	public int reduceValuesToInt(ObjectToInt<? super V> transformer,
3716	>	int basis,
3717	>	IntByIntToInt reducer) {
3718	>	return ForkJoinTasks.reduceValuesToInt
3719	>	(this, transformer, basis, reducer).invoke();
3720	>	}
3721	>
3722	>	/**
3723	>	* Performs the given action for each entry.
3724	>	*
3725	>	* @param action the action
3726	>	*/
3727	>	public void forEachEntry(Action<Map.Entry<K,V>> action) {
3728	>	ForkJoinTasks.forEachEntry
3729	>	(this, action).invoke();
3730	>	}
3731	>
3732	>	/**
3733	>	* Performs the given action for each non-null transformation
3734	>	* of each entry.
3735	>	*
3736	>	* @param transformer a function returning the transformation
3737	>	* for an element, or null of there is no transformation (in
3738	>	* which case the action is not applied).
3739	>	* @param action the action
3740	>	*/
3741	>	public <U> void forEachEntry(Fun<Map.Entry<K,V>, ? extends U> transformer,
3742	>	Action<U> action) {
3743	>	ForkJoinTasks.forEachEntry
3744	>	(this, transformer, action).invoke();
3745	>	}
3746	>
3747	>	/**
3748	>	* Returns a non-null result from applying the given search
3749	>	* function on each entry, or null if none.  Upon success,
3750	>	* further element processing is suppressed and the results of
3751	>	* any other parallel invocations of the search function are
3752	>	* ignored.
3753	>	*
3754	>	* @param searchFunction a function returning a non-null
3755	>	* result on success, else null
3756	>	* @return a non-null result from applying the given search
3757	>	* function on each entry, or null if none
3758	>	*/
3759	>	public <U> U searchEntries(Fun<Map.Entry<K,V>, ? extends U> searchFunction) {
3760	>	return ForkJoinTasks.searchEntries
3761	>	(this, searchFunction).invoke();
3762	>	}
3763	>
3764	>	/**
3765	>	* Returns the result of accumulating all entries using the
3766	>	* given reducer to combine values, or null if none.
3767	>	*
3768	>	* @param reducer a commutative associative combining function
3769	>	* @return the result of accumulating all entries
3770	>	*/
3771	>	public Map.Entry<K,V> reduceEntries(BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer) {
3772	>	return ForkJoinTasks.reduceEntries
3773	>	(this, reducer).invoke();
3774	>	}
3775	>
3776	>	/**
3777	>	* Returns the result of accumulating the given transformation
3778	>	* of all entries using the given reducer to combine values,
3779	>	* or null if none.
3780	>	*
3781	>	* @param transformer a function returning the transformation
3782	>	* for an element, or null of there is no transformation (in
3783	>	* which case it is not combined).
3784	>	* @param reducer a commutative associative combining function
3785	>	* @return the result of accumulating the given transformation
3786	>	* of all entries
3787	>	*/
3788	>	public <U> U reduceEntries(Fun<Map.Entry<K,V>, ? extends U> transformer,
3789	>	BiFun<? super U, ? super U, ? extends U> reducer) {
3790	>	return ForkJoinTasks.reduceEntries
3791	>	(this, transformer, reducer).invoke();
3792	>	}
3793	>
3794	>	/**
3795	>	* Returns the result of accumulating the given transformation
3796	>	* of all entries using the given reducer to combine values,
3797	>	* and the given basis as an identity value.
3798	>	*
3799	>	* @param transformer a function returning the transformation
3800	>	* for an element
3801	>	* @param basis the identity (initial default value) for the reduction
3802	>	* @param reducer a commutative associative combining function
3803	>	* @return the result of accumulating the given transformation
3804	>	* of all entries
3805	>	*/
3806	>	public double reduceEntriesToDouble(ObjectToDouble<Map.Entry<K,V>> transformer,
3807	>	double basis,
3808	>	DoubleByDoubleToDouble reducer) {
3809	>	return ForkJoinTasks.reduceEntriesToDouble
3810	>	(this, transformer, basis, reducer).invoke();
3811	>	}
3812	>
3813	>	/**
3814	>	* Returns the result of accumulating the given transformation
3815	>	* of all entries using the given reducer to combine values,
3816	>	* and the given basis as an identity value.
3817	>	*
3818	>	* @param transformer a function returning the transformation
3819	>	* for an element
3820	>	* @param basis the identity (initial default value) for the reduction
3821	>	* @param reducer a commutative associative combining function
3822	>	* @return  the result of accumulating the given transformation
3823	>	* of all entries
3824	>	*/
3825	>	public long reduceEntriesToLong(ObjectToLong<Map.Entry<K,V>> transformer,
3826	>	long basis,
3827	>	LongByLongToLong reducer) {
3828	>	return ForkJoinTasks.reduceEntriesToLong
3829	>	(this, transformer, basis, reducer).invoke();
3830	>	}
3831	>
3832	>	/**
3833	>	* Returns the result of accumulating the given transformation
3834	>	* of all entries using the given reducer to combine values,
3835	>	* and the given basis as an identity value.
3836	>	*
3837	>	* @param transformer a function returning the transformation
3838	>	* for an element
3839	>	* @param basis the identity (initial default value) for the reduction
3840	>	* @param reducer a commutative associative combining function
3841	>	* @return the result of accumulating the given transformation
3842	>	* of all entries
3843	>	*/
3844	>	public int reduceEntriesToInt(ObjectToInt<Map.Entry<K,V>> transformer,
3845	>	int basis,
3846	>	IntByIntToInt reducer) {
3847	>	return ForkJoinTasks.reduceEntriesToInt
3848	>	(this, transformer, basis, reducer).invoke();
3849	>	}
3850	>
3851	>	/* ----------------Views -------------- */
3852	>
3853	>	/**
3854	>	* Base class for views.
3855	>	*/
3856	>	static abstract class CHMView<K, V> {
3857	>	final ConcurrentHashMap<K, V> map;
3858	>	CHMView(ConcurrentHashMap<K, V> map)  { this.map = map; }
3859	>
3860	>	/**
3861	>	* Returns the map backing this view.
3862	>	*
3863	>	* @return the map backing this view
3864	>	*/
3865	>	public ConcurrentHashMap<K,V> getMap() { return map; }
3866	>
3867	>	public final int size()                 { return map.size(); }
3868	>	public final boolean isEmpty()          { return map.isEmpty(); }
3869	>	public final void clear()               { map.clear(); }
3870	>
3871	>	// implementations below rely on concrete classes supplying these
3872	>	abstract public Iterator<?> iterator();
3873	>	abstract public boolean contains(Object o);
3874	>	abstract public boolean remove(Object o);
3875	>
3876	>	private static final String oomeMsg = "Required array size too large";
3877	>
3878	>	public final Object[] toArray() {
3879	>	long sz = map.mappingCount();
3880	>	if (sz > (long)(MAX_ARRAY_SIZE))
3881	>	throw new OutOfMemoryError(oomeMsg);
3882	>	int n = (int)sz;
3883	>	Object[] r = new Object[n];
3884	>	int i = 0;
3885	>	Iterator<?> it = iterator();
3886	>	while (it.hasNext()) {
3887	>	if (i == n) {
3888	>	if (n >= MAX_ARRAY_SIZE)
3889	>	throw new OutOfMemoryError(oomeMsg);
3890	>	if (n >= MAX_ARRAY_SIZE - (MAX_ARRAY_SIZE >>> 1) - 1)
3891	>	n = MAX_ARRAY_SIZE;
3892	>	else
3893	>	n += (n >>> 1) + 1;
3894	>	r = Arrays.copyOf(r, n);
3895	>	}
3896	>	r[i++] = it.next();
3897	>	}
3898	>	return (i == n) ? r : Arrays.copyOf(r, i);
3899	>	}
3900	>
3901	>	@SuppressWarnings("unchecked") public final <T> T[] toArray(T[] a) {
3902	>	long sz = map.mappingCount();
3903	>	if (sz > (long)(MAX_ARRAY_SIZE))
3904	>	throw new OutOfMemoryError(oomeMsg);
3905	>	int m = (int)sz;
3906	>	T[] r = (a.length >= m) ? a :
3907	>	(T[])java.lang.reflect.Array
3908	>	.newInstance(a.getClass().getComponentType(), m);
3909	>	int n = r.length;
3910	>	int i = 0;
3911	>	Iterator<?> it = iterator();
3912	>	while (it.hasNext()) {
3913	>	if (i == n) {
3914	>	if (n >= MAX_ARRAY_SIZE)
3915	>	throw new OutOfMemoryError(oomeMsg);
3916	>	if (n >= MAX_ARRAY_SIZE - (MAX_ARRAY_SIZE >>> 1) - 1)
3917	>	n = MAX_ARRAY_SIZE;
3918	>	else
3919	>	n += (n >>> 1) + 1;
3920	>	r = Arrays.copyOf(r, n);
3921	>	}
3922	>	r[i++] = (T)it.next();
3923	>	}
3924	>	if (a == r && i < n) {
3925	>	r[i] = null; // null-terminate
3926	>	return r;
3927	>	}
3928	>	return (i == n) ? r : Arrays.copyOf(r, i);
3929	>	}
3930	>
3931	>	public final int hashCode() {
3932	>	int h = 0;
3933	>	for (Iterator<?> it = iterator(); it.hasNext();)
3934	>	h += it.next().hashCode();
3935	>	return h;
3936	>	}
3937	>
3938	>	public final String toString() {
3939	>	StringBuilder sb = new StringBuilder();
3940	>	sb.append('[');
3941	>	Iterator<?> it = iterator();
3942	>	if (it.hasNext()) {
3943	>	for (;;) {
3944	>	Object e = it.next();
3945	>	sb.append(e == this ? "(this Collection)" : e);
3946	>	if (!it.hasNext())
3947	>	break;
3948	>	sb.append(',').append(' ');
3949	>	}
3950	>	}
3951	>	return sb.append(']').toString();
3952	>	}
3953	>
3954	>	public final boolean containsAll(Collection<?> c) {
3955	>	if (c != this) {
3956	>	for (Iterator<?> it = c.iterator(); it.hasNext();) {
3957	>	Object e = it.next();
3958	>	if (e == null || !contains(e))
3959	>	return false;
3960	>	}
3961	>	}
3962	>	return true;
3963	>	}
3964	>
3965	>	public final boolean removeAll(Collection<?> c) {
3966	>	boolean modified = false;
3967	>	for (Iterator<?> it = iterator(); it.hasNext();) {
3968	>	if (c.contains(it.next())) {
3969	>	it.remove();
3970	>	modified = true;
3971	>	}
3972	>	}
3973	>	return modified;
3974	>	}
3975	>
3976	>	public final boolean retainAll(Collection<?> c) {
3977	>	boolean modified = false;
3978	>	for (Iterator<?> it = iterator(); it.hasNext();) {
3979	>	if (!c.contains(it.next())) {
3980	>	it.remove();
3981	>	modified = true;
3982	>	}
3983	>	}
3984	>	return modified;
3985	>	}
3986	>
3987	>	}
3988	>
3989	>	/**
3990	>	* A view of a ConcurrentHashMap as a {@link Set} of keys, in
3991	>	* which additions may optionally be enabled by mapping to a
3992	>	* common value.  This class cannot be directly instantiated. See
3993	>	* {@link #keySet}, {@link #keySet(Object)}, {@link #newKeySet()},
3994	>	* {@link #newKeySet(int)}.
3995	>	*/
3996	>	public static class KeySetView<K,V> extends CHMView<K,V>
3997	>	implements Set<K>, java.io.Serializable {
3998	>	private static final long serialVersionUID = 7249069246763182397L;
3999	>	private final V value;
4000	>	KeySetView(ConcurrentHashMap<K, V> map, V value) {  // non-public
4001	>	super(map);
4002	>	this.value = value;
4003	>	}
4004	>
4005	>	/**
4006	>	* Returns the default mapped value for additions,
4007	>	* or {@code null} if additions are not supported.
4008	>	*
4009	>	* @return the default mapped value for additions, or {@code null}
4010	>	* if not supported.
4011	>	*/
4012	>	public V getMappedValue() { return value; }
4013	>
4014	>	// implement Set API
4015	>
4016	>	public boolean contains(Object o) { return map.containsKey(o); }
4017	>	public boolean remove(Object o)   { return map.remove(o) != null; }
4018	>
4019	>	/**
4020	>	* Returns a "weakly consistent" iterator that will never
4021	>	* throw {@link ConcurrentModificationException}, and
4022	>	* guarantees to traverse elements as they existed upon
4023	>	* construction of the iterator, and may (but is not
4024	>	* guaranteed to) reflect any modifications subsequent to
4025	>	* construction.
4026	>	*
4027	>	* @return an iterator over the keys of this map
4028	>	*/
4029	>	public Iterator<K> iterator()     { return new KeyIterator<K,V>(map); }
4030	>	public boolean add(K e) {
4031	>	V v;
4032	>	if ((v = value) == null)
4033	>	throw new UnsupportedOperationException();
4034	>	if (e == null)
4035	>	throw new NullPointerException();
4036	>	return map.internalPut(e, v, true) == null;
4037	>	}
4038	>	public boolean addAll(Collection<? extends K> c) {
4039	>	boolean added = false;
4040	>	V v;
4041	>	if ((v = value) == null)
4042	>	throw new UnsupportedOperationException();
4043	>	for (K e : c) {
4044	>	if (e == null)
4045	>	throw new NullPointerException();
4046	>	if (map.internalPut(e, v, true) == null)
4047	>	added = true;
4048	>	}
4049	>	return added;
4050		}
4051	<	public boolean remove(Object o) {
4052	<	return ConcurrentHashMap.this.remove(o) != null;
4051	>	public boolean equals(Object o) {
4052	>	Set<?> c;
4053	>	return ((o instanceof Set) &&
4054	>	((c = (Set<?>)o) == this ||
4055	>	(containsAll(c) && c.containsAll(this))));
4056		}
4057	<	public void clear() {
4058	<	ConcurrentHashMap.this.clear();
4057	>
4058	>	/**
4059	>	* Performs the given action for each key.
4060	>	*
4061	>	* @param action the action
4062	>	*/
4063	>	public void forEach(Action<K> action) {
4064	>	ForkJoinTasks.forEachKey
4065	>	(map, action).invoke();
4066		}
4067	<	public Object[] toArray() {
4068	<	Collection<K> c = new ArrayList<K>(size());
4069	<	for (K k : this)
4070	<	c.add(k);
4071	<	return c.toArray();
4067	>
4068	>	/**
4069	>	* Performs the given action for each non-null transformation
4070	>	* of each key.
4071	>	*
4072	>	* @param transformer a function returning the transformation
4073	>	* for an element, or null of there is no transformation (in
4074	>	* which case the action is not applied).
4075	>	* @param action the action
4076	>	*/
4077	>	public <U> void forEach(Fun<? super K, ? extends U> transformer,
4078	>	Action<U> action) {
4079	>	ForkJoinTasks.forEachKey
4080	>	(map, transformer, action).invoke();
4081		}
4082	<	public <T> T[] toArray(T[] a) {
4083	<	Collection<K> c = new ArrayList<K>();
4084	<	for (K k : this)
4085	<	c.add(k);
4086	<	return c.toArray(a);
4082	>
4083	>	/**
4084	>	* Returns a non-null result from applying the given search
4085	>	* function on each key, or null if none. Upon success,
4086	>	* further element processing is suppressed and the results of
4087	>	* any other parallel invocations of the search function are
4088	>	* ignored.
4089	>	*
4090	>	* @param searchFunction a function returning a non-null
4091	>	* result on success, else null
4092	>	* @return a non-null result from applying the given search
4093	>	* function on each key, or null if none
4094	>	*/
4095	>	public <U> U search(Fun<? super K, ? extends U> searchFunction) {
4096	>	return ForkJoinTasks.searchKeys
4097	>	(map, searchFunction).invoke();
4098	>	}
4099	>
4100	>	/**
4101	>	* Returns the result of accumulating all keys using the given
4102	>	* reducer to combine values, or null if none.
4103	>	*
4104	>	* @param reducer a commutative associative combining function
4105	>	* @return the result of accumulating all keys using the given
4106	>	* reducer to combine values, or null if none
4107	>	*/
4108	>	public K reduce(BiFun<? super K, ? super K, ? extends K> reducer) {
4109	>	return ForkJoinTasks.reduceKeys
4110	>	(map, reducer).invoke();
4111	>	}
4112	>
4113	>	/**
4114	>	* Returns the result of accumulating the given transformation
4115	>	* of all keys using the given reducer to combine values, and
4116	>	* the given basis as an identity value.
4117	>	*
4118	>	* @param transformer a function returning the transformation
4119	>	* for an element
4120	>	* @param basis the identity (initial default value) for the reduction
4121	>	* @param reducer a commutative associative combining function
4122	>	* @return  the result of accumulating the given transformation
4123	>	* of all keys
4124	>	*/
4125	>	public double reduceToDouble(ObjectToDouble<? super K> transformer,
4126	>	double basis,
4127	>	DoubleByDoubleToDouble reducer) {
4128	>	return ForkJoinTasks.reduceKeysToDouble
4129	>	(map, transformer, basis, reducer).invoke();
4130	>	}
4131	>
4132	>	/**
4133	>	* Returns the result of accumulating the given transformation
4134	>	* of all keys using the given reducer to combine values, and
4135	>	* the given basis as an identity value.
4136	>	*
4137	>	* @param transformer a function returning the transformation
4138	>	* for an element
4139	>	* @param basis the identity (initial default value) for the reduction
4140	>	* @param reducer a commutative associative combining function
4141	>	* @return the result of accumulating the given transformation
4142	>	* of all keys
4143	>	*/
4144	>	public long reduceToLong(ObjectToLong<? super K> transformer,
4145	>	long basis,
4146	>	LongByLongToLong reducer) {
4147	>	return ForkJoinTasks.reduceKeysToLong
4148	>	(map, transformer, basis, reducer).invoke();
4149	>	}
4150	>
4151	>	/**
4152	>	* Returns the result of accumulating the given transformation
4153	>	* of all keys using the given reducer to combine values, and
4154	>	* the given basis as an identity value.
4155	>	*
4156	>	* @param transformer a function returning the transformation
4157	>	* for an element
4158	>	* @param basis the identity (initial default value) for the reduction
4159	>	* @param reducer a commutative associative combining function
4160	>	* @return the result of accumulating the given transformation
4161	>	* of all keys
4162	>	*/
4163	>	public int reduceToInt(ObjectToInt<? super K> transformer,
4164	>	int basis,
4165	>	IntByIntToInt reducer) {
4166	>	return ForkJoinTasks.reduceKeysToInt
4167	>	(map, transformer, basis, reducer).invoke();
4168		}
4169	+
4170		}
4171
4172	<	final class Values extends AbstractCollection<V> {
4173	<	public Iterator<V> iterator() {
4174	<	return new ValueIterator();
4172	>	/**
4173	>	* A view of a ConcurrentHashMap as a {@link Collection} of
4174	>	* values, in which additions are disabled. This class cannot be
4175	>	* directly instantiated. See {@link #values},
4176	>	*
4177	>	* <p>The view's {@code iterator} is a "weakly consistent" iterator
4178	>	* that will never throw {@link ConcurrentModificationException},
4179	>	* and guarantees to traverse elements as they existed upon
4180	>	* construction of the iterator, and may (but is not guaranteed to)
4181	>	* reflect any modifications subsequent to construction.
4182	>	*/
4183	>	public static final class ValuesView<K,V> extends CHMView<K,V>
4184	>	implements Collection<V> {
4185	>	ValuesView(ConcurrentHashMap<K, V> map)   { super(map); }
4186	>	public final boolean contains(Object o) { return map.containsValue(o); }
4187	>	public final boolean remove(Object o) {
4188	>	if (o != null) {
4189	>	Iterator<V> it = new ValueIterator<K,V>(map);
4190	>	while (it.hasNext()) {
4191	>	if (o.equals(it.next())) {
4192	>	it.remove();
4193	>	return true;
4194	>	}
4195	>	}
4196	>	}
4197	>	return false;
4198	>	}
4199	>
4200	>	/**
4201	>	* Returns a "weakly consistent" iterator that will never
4202	>	* throw {@link ConcurrentModificationException}, and
4203	>	* guarantees to traverse elements as they existed upon
4204	>	* construction of the iterator, and may (but is not
4205	>	* guaranteed to) reflect any modifications subsequent to
4206	>	* construction.
4207	>	*
4208	>	* @return an iterator over the values of this map
4209	>	*/
4210	>	public final Iterator<V> iterator() {
4211	>	return new ValueIterator<K,V>(map);
4212		}
4213	<	public int size() {
4214	<	return ConcurrentHashMap.this.size();
4213	>	public final boolean add(V e) {
4214	>	throw new UnsupportedOperationException();
4215		}
4216	<	public boolean contains(Object o) {
4217	<	return ConcurrentHashMap.this.containsValue(o);
4216	>	public final boolean addAll(Collection<? extends V> c) {
4217	>	throw new UnsupportedOperationException();
4218	>	}
4219	>
4220	>	/**
4221	>	* Performs the given action for each value.
4222	>	*
4223	>	* @param action the action
4224	>	*/
4225	>	public void forEach(Action<V> action) {
4226	>	ForkJoinTasks.forEachValue
4227	>	(map, action).invoke();
4228	>	}
4229	>
4230	>	/**
4231	>	* Performs the given action for each non-null transformation
4232	>	* of each value.
4233	>	*
4234	>	* @param transformer a function returning the transformation
4235	>	* for an element, or null of there is no transformation (in
4236	>	* which case the action is not applied).
4237	>	*/
4238	>	public <U> void forEach(Fun<? super V, ? extends U> transformer,
4239	>	Action<U> action) {
4240	>	ForkJoinTasks.forEachValue
4241	>	(map, transformer, action).invoke();
4242	>	}
4243	>
4244	>	/**
4245	>	* Returns a non-null result from applying the given search
4246	>	* function on each value, or null if none.  Upon success,
4247	>	* further element processing is suppressed and the results of
4248	>	* any other parallel invocations of the search function are
4249	>	* ignored.
4250	>	*
4251	>	* @param searchFunction a function returning a non-null
4252	>	* result on success, else null
4253	>	* @return a non-null result from applying the given search
4254	>	* function on each value, or null if none
4255	>	*
4256	>	*/
4257	>	public <U> U search(Fun<? super V, ? extends U> searchFunction) {
4258	>	return ForkJoinTasks.searchValues
4259	>	(map, searchFunction).invoke();
4260	>	}
4261	>
4262	>	/**
4263	>	* Returns the result of accumulating all values using the
4264	>	* given reducer to combine values, or null if none.
4265	>	*
4266	>	* @param reducer a commutative associative combining function
4267	>	* @return  the result of accumulating all values
4268	>	*/
4269	>	public V reduce(BiFun<? super V, ? super V, ? extends V> reducer) {
4270	>	return ForkJoinTasks.reduceValues
4271	>	(map, reducer).invoke();
4272	>	}
4273	>
4274	>	/**
4275	>	* Returns the result of accumulating the given transformation
4276	>	* of all values using the given reducer to combine values, or
4277	>	* null if none.
4278	>	*
4279	>	* @param transformer a function returning the transformation
4280	>	* for an element, or null of there is no transformation (in
4281	>	* which case it is not combined).
4282	>	* @param reducer a commutative associative combining function
4283	>	* @return the result of accumulating the given transformation
4284	>	* of all values
4285	>	*/
4286	>	public <U> U reduce(Fun<? super V, ? extends U> transformer,
4287	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4288	>	return ForkJoinTasks.reduceValues
4289	>	(map, transformer, reducer).invoke();
4290		}
4291	<	public void clear() {
4292	<	ConcurrentHashMap.this.clear();
4291	>
4292	>	/**
4293	>	* Returns the result of accumulating the given transformation
4294	>	* of all values using the given reducer to combine values,
4295	>	* and the given basis as an identity value.
4296	>	*
4297	>	* @param transformer a function returning the transformation
4298	>	* for an element
4299	>	* @param basis the identity (initial default value) for the reduction
4300	>	* @param reducer a commutative associative combining function
4301	>	* @return the result of accumulating the given transformation
4302	>	* of all values
4303	>	*/
4304	>	public double reduceToDouble(ObjectToDouble<? super V> transformer,
4305	>	double basis,
4306	>	DoubleByDoubleToDouble reducer) {
4307	>	return ForkJoinTasks.reduceValuesToDouble
4308	>	(map, transformer, basis, reducer).invoke();
4309		}
4310	<	public Object[] toArray() {
4311	<	Collection<V> c = new ArrayList<V>(size());
4312	<	for (V v : this)
4313	<	c.add(v);
4314	<	return c.toArray();
4310	>
4311	>	/**
4312	>	* Returns the result of accumulating the given transformation
4313	>	* of all values using the given reducer to combine values,
4314	>	* and the given basis as an identity value.
4315	>	*
4316	>	* @param transformer a function returning the transformation
4317	>	* for an element
4318	>	* @param basis the identity (initial default value) for the reduction
4319	>	* @param reducer a commutative associative combining function
4320	>	* @return the result of accumulating the given transformation
4321	>	* of all values
4322	>	*/
4323	>	public long reduceToLong(ObjectToLong<? super V> transformer,
4324	>	long basis,
4325	>	LongByLongToLong reducer) {
4326	>	return ForkJoinTasks.reduceValuesToLong
4327	>	(map, transformer, basis, reducer).invoke();
4328		}
4329	<	public <T> T[] toArray(T[] a) {
4330	<	Collection<V> c = new ArrayList<V>(size());
4331	<	for (V v : this)
4332	<	c.add(v);
4333	<	return c.toArray(a);
4329	>
4330	>	/**
4331	>	* Returns the result of accumulating the given transformation
4332	>	* of all values using the given reducer to combine values,
4333	>	* and the given basis as an identity value.
4334	>	*
4335	>	* @param transformer a function returning the transformation
4336	>	* for an element
4337	>	* @param basis the identity (initial default value) for the reduction
4338	>	* @param reducer a commutative associative combining function
4339	>	* @return the result of accumulating the given transformation
4340	>	* of all values
4341	>	*/
4342	>	public int reduceToInt(ObjectToInt<? super V> transformer,
4343	>	int basis,
4344	>	IntByIntToInt reducer) {
4345	>	return ForkJoinTasks.reduceValuesToInt
4346	>	(map, transformer, basis, reducer).invoke();
4347		}
4348	+
4349		}
4350
4351	<	final class EntrySet extends AbstractSet<Map.Entry<K,V>> {
4352	<	public Iterator<Map.Entry<K,V>> iterator() {
4353	<	return new EntryIterator();
4351	>	/**
4352	>	* A view of a ConcurrentHashMap as a {@link Set} of (key, value)
4353	>	* entries.  This class cannot be directly instantiated. See
4354	>	* {@link #entrySet}.
4355	>	*/
4356	>	public static final class EntrySetView<K,V> extends CHMView<K,V>
4357	>	implements Set<Map.Entry<K,V>> {
4358	>	EntrySetView(ConcurrentHashMap<K, V> map) { super(map); }
4359	>	public final boolean contains(Object o) {
4360	>	Object k, v, r; Map.Entry<?,?> e;
4361	>	return ((o instanceof Map.Entry) &&
4362	>	(k = (e = (Map.Entry<?,?>)o).getKey()) != null &&
4363	>	(r = map.get(k)) != null &&
4364	>	(v = e.getValue()) != null &&
4365	>	(v == r || v.equals(r)));
4366	>	}
4367	>	public final boolean remove(Object o) {
4368	>	Object k, v; Map.Entry<?,?> e;
4369	>	return ((o instanceof Map.Entry) &&
4370	>	(k = (e = (Map.Entry<?,?>)o).getKey()) != null &&
4371	>	(v = e.getValue()) != null &&
4372	>	map.remove(k, v));
4373		}
4374	<	public boolean contains(Object o) {
4375	<	if (!(o instanceof Map.Entry))
4376	<	return false;
4377	<	Map.Entry<?,?> e = (Map.Entry<?,?>)o;
4378	<	V v = ConcurrentHashMap.this.get(e.getKey());
4379	<	return v != null && v.equals(e.getValue());
4374	>
4375	>	/**
4376	>	* Returns a "weakly consistent" iterator that will never
4377	>	* throw {@link ConcurrentModificationException}, and
4378	>	* guarantees to traverse elements as they existed upon
4379	>	* construction of the iterator, and may (but is not
4380	>	* guaranteed to) reflect any modifications subsequent to
4381	>	* construction.
4382	>	*
4383	>	* @return an iterator over the entries of this map
4384	>	*/
4385	>	public final Iterator<Map.Entry<K,V>> iterator() {
4386	>	return new EntryIterator<K,V>(map);
4387		}
4388	<	public boolean remove(Object o) {
4389	<	if (!(o instanceof Map.Entry))
4390	<	return false;
4391	<	Map.Entry<?,?> e = (Map.Entry<?,?>)o;
4392	<	return ConcurrentHashMap.this.remove(e.getKey(), e.getValue());
4388	>
4389	>	public final boolean add(Entry<K,V> e) {
4390	>	K key = e.getKey();
4391	>	V value = e.getValue();
4392	>	if (key == null || value == null)
4393	>	throw new NullPointerException();
4394	>	return map.internalPut(key, value, false) == null;
4395	>	}
4396	>	public final boolean addAll(Collection<? extends Entry<K,V>> c) {
4397	>	boolean added = false;
4398	>	for (Entry<K,V> e : c) {
4399	>	if (add(e))
4400	>	added = true;
4401	>	}
4402	>	return added;
4403	>	}
4404	>	public boolean equals(Object o) {
4405	>	Set<?> c;
4406	>	return ((o instanceof Set) &&
4407	>	((c = (Set<?>)o) == this ||
4408	>	(containsAll(c) && c.containsAll(this))));
4409	>	}
4410	>
4411	>	/**
4412	>	* Performs the given action for each entry.
4413	>	*
4414	>	* @param action the action
4415	>	*/
4416	>	public void forEach(Action<Map.Entry<K,V>> action) {
4417	>	ForkJoinTasks.forEachEntry
4418	>	(map, action).invoke();
4419	>	}
4420	>
4421	>	/**
4422	>	* Performs the given action for each non-null transformation
4423	>	* of each entry.
4424	>	*
4425	>	* @param transformer a function returning the transformation
4426	>	* for an element, or null of there is no transformation (in
4427	>	* which case the action is not applied).
4428	>	* @param action the action
4429	>	*/
4430	>	public <U> void forEach(Fun<Map.Entry<K,V>, ? extends U> transformer,
4431	>	Action<U> action) {
4432	>	ForkJoinTasks.forEachEntry
4433	>	(map, transformer, action).invoke();
4434	>	}
4435	>
4436	>	/**
4437	>	* Returns a non-null result from applying the given search
4438	>	* function on each entry, or null if none.  Upon success,
4439	>	* further element processing is suppressed and the results of
4440	>	* any other parallel invocations of the search function are
4441	>	* ignored.
4442	>	*
4443	>	* @param searchFunction a function returning a non-null
4444	>	* result on success, else null
4445	>	* @return a non-null result from applying the given search
4446	>	* function on each entry, or null if none
4447	>	*/
4448	>	public <U> U search(Fun<Map.Entry<K,V>, ? extends U> searchFunction) {
4449	>	return ForkJoinTasks.searchEntries
4450	>	(map, searchFunction).invoke();
4451	>	}
4452	>
4453	>	/**
4454	>	* Returns the result of accumulating all entries using the
4455	>	* given reducer to combine values, or null if none.
4456	>	*
4457	>	* @param reducer a commutative associative combining function
4458	>	* @return the result of accumulating all entries
4459	>	*/
4460	>	public Map.Entry<K,V> reduce(BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer) {
4461	>	return ForkJoinTasks.reduceEntries
4462	>	(map, reducer).invoke();
4463		}
4464	<	public int size() {
4465	<	return ConcurrentHashMap.this.size();
4464	>
4465	>	/**
4466	>	* Returns the result of accumulating the given transformation
4467	>	* of all entries using the given reducer to combine values,
4468	>	* or null if none.
4469	>	*
4470	>	* @param transformer a function returning the transformation
4471	>	* for an element, or null of there is no transformation (in
4472	>	* which case it is not combined).
4473	>	* @param reducer a commutative associative combining function
4474	>	* @return the result of accumulating the given transformation
4475	>	* of all entries
4476	>	*/
4477	>	public <U> U reduce(Fun<Map.Entry<K,V>, ? extends U> transformer,
4478	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4479	>	return ForkJoinTasks.reduceEntries
4480	>	(map, transformer, reducer).invoke();
4481	>	}
4482	>
4483	>	/**
4484	>	* Returns the result of accumulating the given transformation
4485	>	* of all entries using the given reducer to combine values,
4486	>	* and the given basis as an identity value.
4487	>	*
4488	>	* @param transformer a function returning the transformation
4489	>	* for an element
4490	>	* @param basis the identity (initial default value) for the reduction
4491	>	* @param reducer a commutative associative combining function
4492	>	* @return the result of accumulating the given transformation
4493	>	* of all entries
4494	>	*/
4495	>	public double reduceToDouble(ObjectToDouble<Map.Entry<K,V>> transformer,
4496	>	double basis,
4497	>	DoubleByDoubleToDouble reducer) {
4498	>	return ForkJoinTasks.reduceEntriesToDouble
4499	>	(map, transformer, basis, reducer).invoke();
4500		}
4501	<	public void clear() {
4502	<	ConcurrentHashMap.this.clear();
4501	>
4502	>	/**
4503	>	* Returns the result of accumulating the given transformation
4504	>	* of all entries using the given reducer to combine values,
4505	>	* and the given basis as an identity value.
4506	>	*
4507	>	* @param transformer a function returning the transformation
4508	>	* for an element
4509	>	* @param basis the identity (initial default value) for the reduction
4510	>	* @param reducer a commutative associative combining function
4511	>	* @return  the result of accumulating the given transformation
4512	>	* of all entries
4513	>	*/
4514	>	public long reduceToLong(ObjectToLong<Map.Entry<K,V>> transformer,
4515	>	long basis,
4516	>	LongByLongToLong reducer) {
4517	>	return ForkJoinTasks.reduceEntriesToLong
4518	>	(map, transformer, basis, reducer).invoke();
4519		}
4520	<	public Object[] toArray() {
4521	<	Collection<Map.Entry<K,V>> c = new ArrayList<Map.Entry<K,V>>(size());
4522	<	for (Map.Entry<K,V> e : this)
4523	<	c.add(e);
4524	<	return c.toArray();
4525	<	}
4526	<	public <T> T[] toArray(T[] a) {
4527	<	Collection<Map.Entry<K,V>> c = new ArrayList<Map.Entry<K,V>>(size());
4528	<	for (Map.Entry<K,V> e : this)
4529	<	c.add(e);
4530	<	return c.toArray(a);
4520	>
4521	>	/**
4522	>	* Returns the result of accumulating the given transformation
4523	>	* of all entries using the given reducer to combine values,
4524	>	* and the given basis as an identity value.
4525	>	*
4526	>	* @param transformer a function returning the transformation
4527	>	* for an element
4528	>	* @param basis the identity (initial default value) for the reduction
4529	>	* @param reducer a commutative associative combining function
4530	>	* @return the result of accumulating the given transformation
4531	>	* of all entries
4532	>	*/
4533	>	public int reduceToInt(ObjectToInt<Map.Entry<K,V>> transformer,
4534	>	int basis,
4535	>	IntByIntToInt reducer) {
4536	>	return ForkJoinTasks.reduceEntriesToInt
4537	>	(map, transformer, basis, reducer).invoke();
4538		}
4539
4540		}
4541
4542	<	/* ---------------- Serialization Support -------------- */
4542	>	// ---------------------------------------------------------------------
4543
4544		/**
4545	<	* Save the state of the <tt>ConcurrentHashMap</tt> instance to a
4546	<	* stream (i.e., serialize it).
4547	<	* @param s the stream
4548	<	* @serialData
4549	<	* the key (Object) and value (Object)
4550	<	* for each key-value mapping, followed by a null pair.
4551	<	* The key-value mappings are emitted in no particular order.
4545	>	* Predefined tasks for performing bulk parallel operations on
4546	>	* ConcurrentHashMaps. These tasks follow the forms and rules used
4547	>	* for bulk operations. Each method has the same name, but returns
4548	>	* a task rather than invoking it. These methods may be useful in
4549	>	* custom applications such as submitting a task without waiting
4550	>	* for completion, using a custom pool, or combining with other
4551	>	* tasks.
4552		*/
4553	<	private void writeObject(java.io.ObjectOutputStream s) throws IOException  {
4554	<	s.defaultWriteObject();
4553	>	public static class ForkJoinTasks {
4554	>	private ForkJoinTasks() {}
4555
4556	<	for (int k = 0; k < segments.length; ++k) {
4557	<	Segment<K,V> seg = segments[k];
4558	<	seg.lock();
4559	<	try {
4560	<	HashEntry<K,V>[] tab = seg.table;
4561	<	for (int i = 0; i < tab.length; ++i) {
4562	<	for (HashEntry<K,V> e = tab[i]; e != null; e = e.next) {
4563	<	s.writeObject(e.key);
4564	<	s.writeObject(e.value);
4556	>	/**
4557	>	* Returns a task that when invoked, performs the given
4558	>	* action for each (key, value)
4559	>	*
4560	>	* @param map the map
4561	>	* @param action the action
4562	>	* @return the task
4563	>	*/
4564	>	public static <K,V> ForkJoinTask<Void> forEach
4565	>	(ConcurrentHashMap<K,V> map,
4566	>	BiAction<K,V> action) {
4567	>	if (action == null) throw new NullPointerException();
4568	>	return new ForEachMappingTask<K,V>(map, null, -1, action);
4569	>	}
4570	>
4571	>	/**
4572	>	* Returns a task that when invoked, performs the given
4573	>	* action for each non-null transformation of each (key, value)
4574	>	*
4575	>	* @param map the map
4576	>	* @param transformer a function returning the transformation
4577	>	* for an element, or null if there is no transformation (in
4578	>	* which case the action is not applied)
4579	>	* @param action the action
4580	>	* @return the task
4581	>	*/
4582	>	public static <K,V,U> ForkJoinTask<Void> forEach
4583	>	(ConcurrentHashMap<K,V> map,
4584	>	BiFun<? super K, ? super V, ? extends U> transformer,
4585	>	Action<U> action) {
4586	>	if (transformer == null || action == null)
4587	>	throw new NullPointerException();
4588	>	return new ForEachTransformedMappingTask<K,V,U>
4589	>	(map, null, -1, 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 (key,
4595	>	* value), or null if none. Upon success, further element
4596	>	* processing is suppressed and the results of any other
4597	>	* parallel 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> search
4605	>	(ConcurrentHashMap<K,V> map,
4606	>	BiFun<? super K, ? super V, ? extends U> searchFunction) {
4607	>	if (searchFunction == null) throw new NullPointerException();
4608	>	return new SearchMappingsTask<K,V,U>
4609	>	(map, null, -1, searchFunction,
4610	>	new AtomicReference<U>());
4611	>	}
4612	>
4613	>	/**
4614	>	* Returns a task that when invoked, returns the result of
4615	>	* accumulating the given transformation of all (key, value) pairs
4616	>	* using the given reducer to combine values, or null if none.
4617	>	*
4618	>	* @param map the map
4619	>	* @param transformer a function returning the transformation
4620	>	* for an element, or null if there is no transformation (in
4621	>	* which case it is not combined).
4622	>	* @param reducer a commutative associative combining function
4623	>	* @return the task
4624	>	*/
4625	>	public static <K,V,U> ForkJoinTask<U> reduce
4626	>	(ConcurrentHashMap<K,V> map,
4627	>	BiFun<? super K, ? super V, ? extends U> transformer,
4628	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4629	>	if (transformer == null || reducer == null)
4630	>	throw new NullPointerException();
4631	>	return new MapReduceMappingsTask<K,V,U>
4632	>	(map, null, -1, null, transformer, reducer);
4633	>	}
4634	>
4635	>	/**
4636	>	* Returns a task that when invoked, returns the result of
4637	>	* accumulating the given transformation of all (key, value) pairs
4638	>	* using the given reducer to combine values, and the given
4639	>	* basis as an identity value.
4640	>	*
4641	>	* @param map the map
4642	>	* @param transformer a function returning the transformation
4643	>	* for an element
4644	>	* @param basis the identity (initial default value) for the reduction
4645	>	* @param reducer a commutative associative combining function
4646	>	* @return the task
4647	>	*/
4648	>	public static <K,V> ForkJoinTask<Double> reduceToDouble
4649	>	(ConcurrentHashMap<K,V> map,
4650	>	ObjectByObjectToDouble<? super K, ? super V> transformer,
4651	>	double basis,
4652	>	DoubleByDoubleToDouble reducer) {
4653	>	if (transformer == null || reducer == null)
4654	>	throw new NullPointerException();
4655	>	return new MapReduceMappingsToDoubleTask<K,V>
4656	>	(map, null, -1, null, transformer, basis, reducer);
4657	>	}
4658	>
4659	>	/**
4660	>	* Returns a task that when invoked, returns the result of
4661	>	* accumulating the given transformation of all (key, value) pairs
4662	>	* using the given reducer to combine values, and the given
4663	>	* basis as an identity value.
4664	>	*
4665	>	* @param map the map
4666	>	* @param transformer a function returning the transformation
4667	>	* for an element
4668	>	* @param basis the identity (initial default value) for the reduction
4669	>	* @param reducer a commutative associative combining function
4670	>	* @return the task
4671	>	*/
4672	>	public static <K,V> ForkJoinTask<Long> reduceToLong
4673	>	(ConcurrentHashMap<K,V> map,
4674	>	ObjectByObjectToLong<? super K, ? super V> transformer,
4675	>	long basis,
4676	>	LongByLongToLong reducer) {
4677	>	if (transformer == null || reducer == null)
4678	>	throw new NullPointerException();
4679	>	return new MapReduceMappingsToLongTask<K,V>
4680	>	(map, null, -1, null, transformer, basis, reducer);
4681	>	}
4682	>
4683	>	/**
4684	>	* Returns a task that when invoked, returns the result of
4685	>	* accumulating the given transformation of all (key, value) pairs
4686	>	* using the given reducer to combine values, and the given
4687	>	* basis as an identity value.
4688	>	*
4689	>	* @param transformer a function returning the transformation
4690	>	* for an element
4691	>	* @param basis the identity (initial default value) for the reduction
4692	>	* @param reducer a commutative associative combining function
4693	>	* @return the task
4694	>	*/
4695	>	public static <K,V> ForkJoinTask<Integer> reduceToInt
4696	>	(ConcurrentHashMap<K,V> map,
4697	>	ObjectByObjectToInt<? super K, ? super V> transformer,
4698	>	int basis,
4699	>	IntByIntToInt reducer) {
4700	>	if (transformer == null || reducer == null)
4701	>	throw new NullPointerException();
4702	>	return new MapReduceMappingsToIntTask<K,V>
4703	>	(map, null, -1, null, transformer, basis, reducer);
4704	>	}
4705	>
4706	>	/**
4707	>	* Returns a task that when invoked, performs the given action
4708	>	* for each key.
4709	>	*
4710	>	* @param map the map
4711	>	* @param action the action
4712	>	* @return the task
4713	>	*/
4714	>	public static <K,V> ForkJoinTask<Void> forEachKey
4715	>	(ConcurrentHashMap<K,V> map,
4716	>	Action<K> action) {
4717	>	if (action == null) throw new NullPointerException();
4718	>	return new ForEachKeyTask<K,V>(map, null, -1, action);
4719	>	}
4720	>
4721	>	/**
4722	>	* Returns a task that when invoked, performs the given action
4723	>	* for each non-null transformation of each key.
4724	>	*
4725	>	* @param map the map
4726	>	* @param transformer a function returning the transformation
4727	>	* for an element, or null if there is no transformation (in
4728	>	* which case the action is not applied)
4729	>	* @param action the action
4730	>	* @return the task
4731	>	*/
4732	>	public static <K,V,U> ForkJoinTask<Void> forEachKey
4733	>	(ConcurrentHashMap<K,V> map,
4734	>	Fun<? super K, ? extends U> transformer,
4735	>	Action<U> action) {
4736	>	if (transformer == null || action == null)
4737	>	throw new NullPointerException();
4738	>	return new ForEachTransformedKeyTask<K,V,U>
4739	>	(map, null, -1, transformer, action);
4740	>	}
4741	>
4742	>	/**
4743	>	* Returns a task that when invoked, returns a non-null result
4744	>	* from applying the given search function on each key, or
4745	>	* null if none.  Upon success, further element processing is
4746	>	* suppressed and the results of any other parallel
4747	>	* invocations of the search function are ignored.
4748	>	*
4749	>	* @param map the map
4750	>	* @param searchFunction a function returning a non-null
4751	>	* result on success, else null
4752	>	* @return the task
4753	>	*/
4754	>	public static <K,V,U> ForkJoinTask<U> searchKeys
4755	>	(ConcurrentHashMap<K,V> map,
4756	>	Fun<? super K, ? extends U> searchFunction) {
4757	>	if (searchFunction == null) throw new NullPointerException();
4758	>	return new SearchKeysTask<K,V,U>
4759	>	(map, null, -1, searchFunction,
4760	>	new AtomicReference<U>());
4761	>	}
4762	>
4763	>	/**
4764	>	* Returns a task that when invoked, returns the result of
4765	>	* accumulating all keys using the given reducer to combine
4766	>	* values, or null if none.
4767	>	*
4768	>	* @param map the map
4769	>	* @param reducer a commutative associative combining function
4770	>	* @return the task
4771	>	*/
4772	>	public static <K,V> ForkJoinTask<K> reduceKeys
4773	>	(ConcurrentHashMap<K,V> map,
4774	>	BiFun<? super K, ? super K, ? extends K> reducer) {
4775	>	if (reducer == null) throw new NullPointerException();
4776	>	return new ReduceKeysTask<K,V>
4777	>	(map, null, -1, null, reducer);
4778	>	}
4779	>
4780	>	/**
4781	>	* Returns a task that when invoked, returns the result of
4782	>	* accumulating the given transformation of all keys using the given
4783	>	* reducer to combine values, or null if none.
4784	>	*
4785	>	* @param map the map
4786	>	* @param transformer a function returning the transformation
4787	>	* for an element, or null if there is no transformation (in
4788	>	* which case it is not combined).
4789	>	* @param reducer a commutative associative combining function
4790	>	* @return the task
4791	>	*/
4792	>	public static <K,V,U> ForkJoinTask<U> reduceKeys
4793	>	(ConcurrentHashMap<K,V> map,
4794	>	Fun<? super K, ? extends U> transformer,
4795	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4796	>	if (transformer == null || reducer == null)
4797	>	throw new NullPointerException();
4798	>	return new MapReduceKeysTask<K,V,U>
4799	>	(map, null, -1, null, transformer, reducer);
4800	>	}
4801	>
4802	>	/**
4803	>	* Returns a task that when invoked, returns the result of
4804	>	* accumulating the given transformation of all keys using the given
4805	>	* reducer to combine values, and the given basis as an
4806	>	* identity value.
4807	>	*
4808	>	* @param map the map
4809	>	* @param transformer a function returning the transformation
4810	>	* for an element
4811	>	* @param basis the identity (initial default value) for the reduction
4812	>	* @param reducer a commutative associative combining function
4813	>	* @return the task
4814	>	*/
4815	>	public static <K,V> ForkJoinTask<Double> reduceKeysToDouble
4816	>	(ConcurrentHashMap<K,V> map,
4817	>	ObjectToDouble<? super K> transformer,
4818	>	double basis,
4819	>	DoubleByDoubleToDouble reducer) {
4820	>	if (transformer == null || reducer == null)
4821	>	throw new NullPointerException();
4822	>	return new MapReduceKeysToDoubleTask<K,V>
4823	>	(map, null, -1, null, transformer, basis, reducer);
4824	>	}
4825	>
4826	>	/**
4827	>	* Returns a task that when invoked, returns the result of
4828	>	* accumulating the given transformation of all keys using the given
4829	>	* reducer to combine values, and the given basis as an
4830	>	* identity value.
4831	>	*
4832	>	* @param map the map
4833	>	* @param transformer a function returning the transformation
4834	>	* for an element
4835	>	* @param basis the identity (initial default value) for the reduction
4836	>	* @param reducer a commutative associative combining function
4837	>	* @return the task
4838	>	*/
4839	>	public static <K,V> ForkJoinTask<Long> reduceKeysToLong
4840	>	(ConcurrentHashMap<K,V> map,
4841	>	ObjectToLong<? super K> transformer,
4842	>	long basis,
4843	>	LongByLongToLong reducer) {
4844	>	if (transformer == null || reducer == null)
4845	>	throw new NullPointerException();
4846	>	return new MapReduceKeysToLongTask<K,V>
4847	>	(map, null, -1, null, transformer, basis, reducer);
4848	>	}
4849	>
4850	>	/**
4851	>	* Returns a task that when invoked, returns the result of
4852	>	* accumulating the given transformation of all keys using the given
4853	>	* reducer to combine values, and the given basis as an
4854	>	* identity value.
4855	>	*
4856	>	* @param map the map
4857	>	* @param transformer a function returning the transformation
4858	>	* for an element
4859	>	* @param basis the identity (initial default value) for the reduction
4860	>	* @param reducer a commutative associative combining function
4861	>	* @return the task
4862	>	*/
4863	>	public static <K,V> ForkJoinTask<Integer> reduceKeysToInt
4864	>	(ConcurrentHashMap<K,V> map,
4865	>	ObjectToInt<? super K> transformer,
4866	>	int basis,
4867	>	IntByIntToInt reducer) {
4868	>	if (transformer == null || reducer == null)
4869	>	throw new NullPointerException();
4870	>	return new MapReduceKeysToIntTask<K,V>
4871	>	(map, null, -1, null, transformer, basis, reducer);
4872	>	}
4873	>
4874	>	/**
4875	>	* Returns a task that when invoked, performs the given action
4876	>	* for each value.
4877	>	*
4878	>	* @param map the map
4879	>	* @param action the action
4880	>	*/
4881	>	public static <K,V> ForkJoinTask<Void> forEachValue
4882	>	(ConcurrentHashMap<K,V> map,
4883	>	Action<V> action) {
4884	>	if (action == null) throw new NullPointerException();
4885	>	return new ForEachValueTask<K,V>(map, null, -1, action);
4886	>	}
4887	>
4888	>	/**
4889	>	* Returns a task that when invoked, performs the given action
4890	>	* for each non-null transformation of each value.
4891	>	*
4892	>	* @param map the map
4893	>	* @param transformer a function returning the transformation
4894	>	* for an element, or null if there is no transformation (in
4895	>	* which case the action is not applied)
4896	>	* @param action the action
4897	>	*/
4898	>	public static <K,V,U> ForkJoinTask<Void> forEachValue
4899	>	(ConcurrentHashMap<K,V> map,
4900	>	Fun<? super V, ? extends U> transformer,
4901	>	Action<U> action) {
4902	>	if (transformer == null || action == null)
4903	>	throw new NullPointerException();
4904	>	return new ForEachTransformedValueTask<K,V,U>
4905	>	(map, null, -1, transformer, action);
4906	>	}
4907	>
4908	>	/**
4909	>	* Returns a task that when invoked, returns a non-null result
4910	>	* from applying the given search function on each value, or
4911	>	* null if none.  Upon success, further element processing is
4912	>	* suppressed and the results of any other parallel
4913	>	* invocations of the search function are ignored.
4914	>	*
4915	>	* @param map the map
4916	>	* @param searchFunction a function returning a non-null
4917	>	* result on success, else null
4918	>	* @return the task
4919	>	*/
4920	>	public static <K,V,U> ForkJoinTask<U> searchValues
4921	>	(ConcurrentHashMap<K,V> map,
4922	>	Fun<? super V, ? extends U> searchFunction) {
4923	>	if (searchFunction == null) throw new NullPointerException();
4924	>	return new SearchValuesTask<K,V,U>
4925	>	(map, null, -1, searchFunction,
4926	>	new AtomicReference<U>());
4927	>	}
4928	>
4929	>	/**
4930	>	* Returns a task that when invoked, returns the result of
4931	>	* accumulating all values using the given reducer to combine
4932	>	* values, or null if none.
4933	>	*
4934	>	* @param map the map
4935	>	* @param reducer a commutative associative combining function
4936	>	* @return the task
4937	>	*/
4938	>	public static <K,V> ForkJoinTask<V> reduceValues
4939	>	(ConcurrentHashMap<K,V> map,
4940	>	BiFun<? super V, ? super V, ? extends V> reducer) {
4941	>	if (reducer == null) throw new NullPointerException();
4942	>	return new ReduceValuesTask<K,V>
4943	>	(map, null, -1, null, reducer);
4944	>	}
4945	>
4946	>	/**
4947	>	* Returns a task that when invoked, returns the result of
4948	>	* accumulating the given transformation of all values using the
4949	>	* given reducer to combine values, or null if none.
4950	>	*
4951	>	* @param map the map
4952	>	* @param transformer a function returning the transformation
4953	>	* for an element, or null if there is no transformation (in
4954	>	* which case it is not combined).
4955	>	* @param reducer a commutative associative combining function
4956	>	* @return the task
4957	>	*/
4958	>	public static <K,V,U> ForkJoinTask<U> reduceValues
4959	>	(ConcurrentHashMap<K,V> map,
4960	>	Fun<? super V, ? extends U> transformer,
4961	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4962	>	if (transformer == null || reducer == null)
4963	>	throw new NullPointerException();
4964	>	return new MapReduceValuesTask<K,V,U>
4965	>	(map, null, -1, null, transformer, reducer);
4966	>	}
4967	>
4968	>	/**
4969	>	* Returns a task that when invoked, returns the result of
4970	>	* accumulating the given transformation of all values using the
4971	>	* given reducer to combine values, and the given basis as an
4972	>	* identity value.
4973	>	*
4974	>	* @param map the map
4975	>	* @param transformer a function returning the transformation
4976	>	* for an element
4977	>	* @param basis the identity (initial default value) for the reduction
4978	>	* @param reducer a commutative associative combining function
4979	>	* @return the task
4980	>	*/
4981	>	public static <K,V> ForkJoinTask<Double> reduceValuesToDouble
4982	>	(ConcurrentHashMap<K,V> map,
4983	>	ObjectToDouble<? super V> transformer,
4984	>	double basis,
4985	>	DoubleByDoubleToDouble reducer) {
4986	>	if (transformer == null || reducer == null)
4987	>	throw new NullPointerException();
4988	>	return new MapReduceValuesToDoubleTask<K,V>
4989	>	(map, null, -1, null, transformer, basis, reducer);
4990	>	}
4991	>
4992	>	/**
4993	>	* Returns a task that when invoked, returns the result of
4994	>	* accumulating the given transformation of all values using the
4995	>	* given reducer to combine values, and the given basis as an
4996	>	* identity value.
4997	>	*
4998	>	* @param map the map
4999	>	* @param transformer a function returning the transformation
5000	>	* for an element
5001	>	* @param basis the identity (initial default value) for the reduction
5002	>	* @param reducer a commutative associative combining function
5003	>	* @return the task
5004	>	*/
5005	>	public static <K,V> ForkJoinTask<Long> reduceValuesToLong
5006	>	(ConcurrentHashMap<K,V> map,
5007	>	ObjectToLong<? super V> transformer,
5008	>	long basis,
5009	>	LongByLongToLong reducer) {
5010	>	if (transformer == null || reducer == null)
5011	>	throw new NullPointerException();
5012	>	return new MapReduceValuesToLongTask<K,V>
5013	>	(map, null, -1, null, transformer, basis, reducer);
5014	>	}
5015	>
5016	>	/**
5017	>	* Returns a task that when invoked, returns the result of
5018	>	* accumulating the given transformation of all values using the
5019	>	* given reducer to combine values, and the given basis as an
5020	>	* identity value.
5021	>	*
5022	>	* @param map the map
5023	>	* @param transformer a function returning the transformation
5024	>	* for an element
5025	>	* @param basis the identity (initial default value) for the reduction
5026	>	* @param reducer a commutative associative combining function
5027	>	* @return the task
5028	>	*/
5029	>	public static <K,V> ForkJoinTask<Integer> reduceValuesToInt
5030	>	(ConcurrentHashMap<K,V> map,
5031	>	ObjectToInt<? super V> transformer,
5032	>	int basis,
5033	>	IntByIntToInt reducer) {
5034	>	if (transformer == null || reducer == null)
5035	>	throw new NullPointerException();
5036	>	return new MapReduceValuesToIntTask<K,V>
5037	>	(map, null, -1, null, transformer, basis, reducer);
5038	>	}
5039	>
5040	>	/**
5041	>	* Returns a task that when invoked, perform the given action
5042	>	* for each entry.
5043	>	*
5044	>	* @param map the map
5045	>	* @param action the action
5046	>	*/
5047	>	public static <K,V> ForkJoinTask<Void> forEachEntry
5048	>	(ConcurrentHashMap<K,V> map,
5049	>	Action<Map.Entry<K,V>> action) {
5050	>	if (action == null) throw new NullPointerException();
5051	>	return new ForEachEntryTask<K,V>(map, null, -1, action);
5052	>	}
5053	>
5054	>	/**
5055	>	* Returns a task that when invoked, perform the given action
5056	>	* for each non-null transformation of each entry.
5057	>	*
5058	>	* @param map the map
5059	>	* @param transformer a function returning the transformation
5060	>	* for an element, or null if there is no transformation (in
5061	>	* which case the action is not applied)
5062	>	* @param action the action
5063	>	*/
5064	>	public static <K,V,U> ForkJoinTask<Void> forEachEntry
5065	>	(ConcurrentHashMap<K,V> map,
5066	>	Fun<Map.Entry<K,V>, ? extends U> transformer,
5067	>	Action<U> action) {
5068	>	if (transformer == null || action == null)
5069	>	throw new NullPointerException();
5070	>	return new ForEachTransformedEntryTask<K,V,U>
5071	>	(map, null, -1, transformer, action);
5072	>	}
5073	>
5074	>	/**
5075	>	* Returns a task that when invoked, returns a non-null result
5076	>	* from applying the given search function on each entry, or
5077	>	* null if none.  Upon success, further element processing is
5078	>	* suppressed and the results of any other parallel
5079	>	* invocations of the search function are ignored.
5080	>	*
5081	>	* @param map the map
5082	>	* @param searchFunction a function returning a non-null
5083	>	* result on success, else null
5084	>	* @return the task
5085	>	*/
5086	>	public static <K,V,U> ForkJoinTask<U> searchEntries
5087	>	(ConcurrentHashMap<K,V> map,
5088	>	Fun<Map.Entry<K,V>, ? extends U> searchFunction) {
5089	>	if (searchFunction == null) throw new NullPointerException();
5090	>	return new SearchEntriesTask<K,V,U>
5091	>	(map, null, -1, searchFunction,
5092	>	new AtomicReference<U>());
5093	>	}
5094	>
5095	>	/**
5096	>	* Returns a task that when invoked, returns the result of
5097	>	* accumulating all entries using the given reducer to combine
5098	>	* values, or null if none.
5099	>	*
5100	>	* @param map the map
5101	>	* @param reducer a commutative associative combining function
5102	>	* @return the task
5103	>	*/
5104	>	public static <K,V> ForkJoinTask<Map.Entry<K,V>> reduceEntries
5105	>	(ConcurrentHashMap<K,V> map,
5106	>	BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer) {
5107	>	if (reducer == null) throw new NullPointerException();
5108	>	return new ReduceEntriesTask<K,V>
5109	>	(map, null, -1, null, reducer);
5110	>	}
5111	>
5112	>	/**
5113	>	* Returns a task that when invoked, returns the result of
5114	>	* accumulating the given transformation of all entries using the
5115	>	* given reducer to combine values, or null if none.
5116	>	*
5117	>	* @param map the map
5118	>	* @param transformer a function returning the transformation
5119	>	* for an element, or null if there is no transformation (in
5120	>	* which case it is not combined).
5121	>	* @param reducer a commutative associative combining function
5122	>	* @return the task
5123	>	*/
5124	>	public static <K,V,U> ForkJoinTask<U> reduceEntries
5125	>	(ConcurrentHashMap<K,V> map,
5126	>	Fun<Map.Entry<K,V>, ? extends U> transformer,
5127	>	BiFun<? super U, ? super U, ? extends U> reducer) {
5128	>	if (transformer == null || reducer == null)
5129	>	throw new NullPointerException();
5130	>	return new MapReduceEntriesTask<K,V,U>
5131	>	(map, null, -1, null, transformer, reducer);
5132	>	}
5133	>
5134	>	/**
5135	>	* Returns a task that when invoked, returns the result of
5136	>	* accumulating the given transformation of all entries using the
5137	>	* given reducer to combine values, and the given basis as an
5138	>	* identity value.
5139	>	*
5140	>	* @param map the map
5141	>	* @param transformer a function returning the transformation
5142	>	* for an element
5143	>	* @param basis the identity (initial default value) for the reduction
5144	>	* @param reducer a commutative associative combining function
5145	>	* @return the task
5146	>	*/
5147	>	public static <K,V> ForkJoinTask<Double> reduceEntriesToDouble
5148	>	(ConcurrentHashMap<K,V> map,
5149	>	ObjectToDouble<Map.Entry<K,V>> transformer,
5150	>	double basis,
5151	>	DoubleByDoubleToDouble reducer) {
5152	>	if (transformer == null || reducer == null)
5153	>	throw new NullPointerException();
5154	>	return new MapReduceEntriesToDoubleTask<K,V>
5155	>	(map, null, -1, null, transformer, basis, reducer);
5156	>	}
5157	>
5158	>	/**
5159	>	* Returns a task that when invoked, returns the result of
5160	>	* accumulating the given transformation of all entries using the
5161	>	* given reducer to combine values, and the given basis as an
5162	>	* identity value.
5163	>	*
5164	>	* @param map the map
5165	>	* @param transformer a function returning the transformation
5166	>	* for an element
5167	>	* @param basis the identity (initial default value) for the reduction
5168	>	* @param reducer a commutative associative combining function
5169	>	* @return the task
5170	>	*/
5171	>	public static <K,V> ForkJoinTask<Long> reduceEntriesToLong
5172	>	(ConcurrentHashMap<K,V> map,
5173	>	ObjectToLong<Map.Entry<K,V>> transformer,
5174	>	long basis,
5175	>	LongByLongToLong reducer) {
5176	>	if (transformer == null || reducer == null)
5177	>	throw new NullPointerException();
5178	>	return new MapReduceEntriesToLongTask<K,V>
5179	>	(map, null, -1, null, transformer, basis, reducer);
5180	>	}
5181	>
5182	>	/**
5183	>	* Returns a task that when invoked, returns the result of
5184	>	* accumulating the given transformation of all entries using the
5185	>	* given reducer to combine values, and the given basis as an
5186	>	* identity value.
5187	>	*
5188	>	* @param map the map
5189	>	* @param transformer a function returning the transformation
5190	>	* for an element
5191	>	* @param basis the identity (initial default value) for the reduction
5192	>	* @param reducer a commutative associative combining function
5193	>	* @return the task
5194	>	*/
5195	>	public static <K,V> ForkJoinTask<Integer> reduceEntriesToInt
5196	>	(ConcurrentHashMap<K,V> map,
5197	>	ObjectToInt<Map.Entry<K,V>> transformer,
5198	>	int basis,
5199	>	IntByIntToInt reducer) {
5200	>	if (transformer == null || reducer == null)
5201	>	throw new NullPointerException();
5202	>	return new MapReduceEntriesToIntTask<K,V>
5203	>	(map, null, -1, null, transformer, basis, reducer);
5204	>	}
5205	>	}
5206	>
5207	>	// -------------------------------------------------------
5208	>
5209	>	/*
5210	>	* Task classes. Coded in a regular but ugly format/style to
5211	>	* simplify checks that each variant differs in the right way from
5212	>	* others. The null screenings exist because compilers cannot tell
5213	>	* that we've already null-checked task arguments, so we force
5214	>	* simplest hoisted bypass to help avoid convoluted traps.
5215	>	*/
5216	>
5217	>	@SuppressWarnings("serial") static final class ForEachKeyTask<K,V>
5218	>	extends Traverser<K,V,Void> {
5219	>	final Action<K> action;
5220	>	ForEachKeyTask
5221	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5222	>	Action<K> action) {
5223	>	super(m, p, b);
5224	>	this.action = action;
5225	>	}
5226	>	@SuppressWarnings("unchecked") public final void compute() {
5227	>	final Action<K> action;
5228	>	if ((action = this.action) != null) {
5229	>	for (int b; (b = preSplit()) > 0;)
5230	>	new ForEachKeyTask<K,V>(map, this, b, action).fork();
5231	>	while (advance() != null)
5232	>	action.apply((K)nextKey);
5233	>	propagateCompletion();
5234	>	}
5235	>	}
5236	>	}
5237	>
5238	>	@SuppressWarnings("serial") static final class ForEachValueTask<K,V>
5239	>	extends Traverser<K,V,Void> {
5240	>	final Action<V> action;
5241	>	ForEachValueTask
5242	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5243	>	Action<V> action) {
5244	>	super(m, p, b);
5245	>	this.action = action;
5246	>	}
5247	>	@SuppressWarnings("unchecked") public final void compute() {
5248	>	final Action<V> action;
5249	>	if ((action = this.action) != null) {
5250	>	for (int b; (b = preSplit()) > 0;)
5251	>	new ForEachValueTask<K,V>(map, this, b, action).fork();
5252	>	Object v;
5253	>	while ((v = advance()) != null)
5254	>	action.apply((V)v);
5255	>	propagateCompletion();
5256	>	}
5257	>	}
5258	>	}
5259	>
5260	>	@SuppressWarnings("serial") static final class ForEachEntryTask<K,V>
5261	>	extends Traverser<K,V,Void> {
5262	>	final Action<Entry<K,V>> action;
5263	>	ForEachEntryTask
5264	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5265	>	Action<Entry<K,V>> action) {
5266	>	super(m, p, b);
5267	>	this.action = action;
5268	>	}
5269	>	@SuppressWarnings("unchecked") public final void compute() {
5270	>	final Action<Entry<K,V>> action;
5271	>	if ((action = this.action) != null) {
5272	>	for (int b; (b = preSplit()) > 0;)
5273	>	new ForEachEntryTask<K,V>(map, this, b, action).fork();
5274	>	Object v;
5275	>	while ((v = advance()) != null)
5276	>	action.apply(entryFor((K)nextKey, (V)v));
5277	>	propagateCompletion();
5278	>	}
5279	>	}
5280	>	}
5281	>
5282	>	@SuppressWarnings("serial") static final class ForEachMappingTask<K,V>
5283	>	extends Traverser<K,V,Void> {
5284	>	final BiAction<K,V> action;
5285	>	ForEachMappingTask
5286	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5287	>	BiAction<K,V> action) {
5288	>	super(m, p, b);
5289	>	this.action = action;
5290	>	}
5291	>	@SuppressWarnings("unchecked") public final void compute() {
5292	>	final BiAction<K,V> action;
5293	>	if ((action = this.action) != null) {
5294	>	for (int b; (b = preSplit()) > 0;)
5295	>	new ForEachMappingTask<K,V>(map, this, b, action).fork();
5296	>	Object v;
5297	>	while ((v = advance()) != null)
5298	>	action.apply((K)nextKey, (V)v);
5299	>	propagateCompletion();
5300	>	}
5301	>	}
5302	>	}
5303	>
5304	>	@SuppressWarnings("serial") static final class ForEachTransformedKeyTask<K,V,U>
5305	>	extends Traverser<K,V,Void> {
5306	>	final Fun<? super K, ? extends U> transformer;
5307	>	final Action<U> action;
5308	>	ForEachTransformedKeyTask
5309	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5310	>	Fun<? super K, ? extends U> transformer, Action<U> action) {
5311	>	super(m, p, b);
5312	>	this.transformer = transformer; this.action = action;
5313	>	}
5314	>	@SuppressWarnings("unchecked") public final void compute() {
5315	>	final Fun<? super K, ? extends U> transformer;
5316	>	final Action<U> action;
5317	>	if ((transformer = this.transformer) != null &&
5318	>	(action = this.action) != null) {
5319	>	for (int b; (b = preSplit()) > 0;)
5320	>	new ForEachTransformedKeyTask<K,V,U>
5321	>	(map, this, b, transformer, action).fork();
5322	>	U u;
5323	>	while (advance() != null) {
5324	>	if ((u = transformer.apply((K)nextKey)) != null)
5325	>	action.apply(u);
5326	>	}
5327	>	propagateCompletion();
5328	>	}
5329	>	}
5330	>	}
5331	>
5332	>	@SuppressWarnings("serial") static final class ForEachTransformedValueTask<K,V,U>
5333	>	extends Traverser<K,V,Void> {
5334	>	final Fun<? super V, ? extends U> transformer;
5335	>	final Action<U> action;
5336	>	ForEachTransformedValueTask
5337	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5338	>	Fun<? super V, ? extends U> transformer, Action<U> action) {
5339	>	super(m, p, b);
5340	>	this.transformer = transformer; this.action = action;
5341	>	}
5342	>	@SuppressWarnings("unchecked") public final void compute() {
5343	>	final Fun<? super V, ? extends U> transformer;
5344	>	final Action<U> action;
5345	>	if ((transformer = this.transformer) != null &&
5346	>	(action = this.action) != null) {
5347	>	for (int b; (b = preSplit()) > 0;)
5348	>	new ForEachTransformedValueTask<K,V,U>
5349	>	(map, this, b, transformer, action).fork();
5350	>	Object v; U u;
5351	>	while ((v = advance()) != null) {
5352	>	if ((u = transformer.apply((V)v)) != null)
5353	>	action.apply(u);
5354	>	}
5355	>	propagateCompletion();
5356	>	}
5357	>	}
5358	>	}
5359	>
5360	>	@SuppressWarnings("serial") static final class ForEachTransformedEntryTask<K,V,U>
5361	>	extends Traverser<K,V,Void> {
5362	>	final Fun<Map.Entry<K,V>, ? extends U> transformer;
5363	>	final Action<U> action;
5364	>	ForEachTransformedEntryTask
5365	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5366	>	Fun<Map.Entry<K,V>, ? extends U> transformer, Action<U> action) {
5367	>	super(m, p, b);
5368	>	this.transformer = transformer; this.action = action;
5369	>	}
5370	>	@SuppressWarnings("unchecked") public final void compute() {
5371	>	final Fun<Map.Entry<K,V>, ? extends U> transformer;
5372	>	final Action<U> action;
5373	>	if ((transformer = this.transformer) != null &&
5374	>	(action = this.action) != null) {
5375	>	for (int b; (b = preSplit()) > 0;)
5376	>	new ForEachTransformedEntryTask<K,V,U>
5377	>	(map, this, b, transformer, action).fork();
5378	>	Object v; U u;
5379	>	while ((v = advance()) != null) {
5380	>	if ((u = transformer.apply(entryFor((K)nextKey,
5381	>	(V)v))) != null)
5382	>	action.apply(u);
5383	>	}
5384	>	propagateCompletion();
5385	>	}
5386	>	}
5387	>	}
5388	>
5389	>	@SuppressWarnings("serial") static final class ForEachTransformedMappingTask<K,V,U>
5390	>	extends Traverser<K,V,Void> {
5391	>	final BiFun<? super K, ? super V, ? extends U> transformer;
5392	>	final Action<U> action;
5393	>	ForEachTransformedMappingTask
5394	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5395	>	BiFun<? super K, ? super V, ? extends U> transformer,
5396	>	Action<U> action) {
5397	>	super(m, p, b);
5398	>	this.transformer = transformer; this.action = action;
5399	>	}
5400	>	@SuppressWarnings("unchecked") public final void compute() {
5401	>	final BiFun<? super K, ? super V, ? extends U> transformer;
5402	>	final Action<U> action;
5403	>	if ((transformer = this.transformer) != null &&
5404	>	(action = this.action) != null) {
5405	>	for (int b; (b = preSplit()) > 0;)
5406	>	new ForEachTransformedMappingTask<K,V,U>
5407	>	(map, this, b, transformer, action).fork();
5408	>	Object v; U u;
5409	>	while ((v = advance()) != null) {
5410	>	if ((u = transformer.apply((K)nextKey, (V)v)) != null)
5411	>	action.apply(u);
5412	>	}
5413	>	propagateCompletion();
5414	>	}
5415	>	}
5416	>	}
5417	>
5418	>	@SuppressWarnings("serial") static final class SearchKeysTask<K,V,U>
5419	>	extends Traverser<K,V,U> {
5420	>	final Fun<? super K, ? extends U> searchFunction;
5421	>	final AtomicReference<U> result;
5422	>	SearchKeysTask
5423	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5424	>	Fun<? super K, ? extends U> searchFunction,
5425	>	AtomicReference<U> result) {
5426	>	super(m, p, b);
5427	>	this.searchFunction = searchFunction; this.result = result;
5428	>	}
5429	>	public final U getRawResult() { return result.get(); }
5430	>	@SuppressWarnings("unchecked") public final void compute() {
5431	>	final Fun<? super K, ? extends U> searchFunction;
5432	>	final AtomicReference<U> result;
5433	>	if ((searchFunction = this.searchFunction) != null &&
5434	>	(result = this.result) != null) {
5435	>	for (int b;;) {
5436	>	if (result.get() != null)
5437	>	return;
5438	>	if ((b = preSplit()) <= 0)
5439	>	break;
5440	>	new SearchKeysTask<K,V,U>
5441	>	(map, this, b, searchFunction, result).fork();
5442	>	}
5443	>	while (result.get() == null) {
5444	>	U u;
5445	>	if (advance() == null) {
5446	>	propagateCompletion();
5447	>	break;
5448	>	}
5449	>	if ((u = searchFunction.apply((K)nextKey)) != null) {
5450	>	if (result.compareAndSet(null, u))
5451	>	quietlyCompleteRoot();
5452	>	break;
5453		}
5454		}
1283	–	} finally {
1284	–	seg.unlock();
5455		}
5456		}
1287	–	s.writeObject(null);
1288	–	s.writeObject(null);
5457		}
5458
5459	<	/**
5460	<	* Reconstitute the <tt>ConcurrentHashMap</tt> instance from a
5461	<	* stream (i.e., deserialize it).
5462	<	* @param s the stream
5463	<	*/
5464	<	private void readObject(java.io.ObjectInputStream s)
5465	<	throws IOException, ClassNotFoundException  {
5466	<	s.defaultReadObject();
5459	>	@SuppressWarnings("serial") static final class SearchValuesTask<K,V,U>
5460	>	extends Traverser<K,V,U> {
5461	>	final Fun<? super V, ? extends U> searchFunction;
5462	>	final AtomicReference<U> result;
5463	>	SearchValuesTask
5464	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5465	>	Fun<? super V, ? extends U> searchFunction,
5466	>	AtomicReference<U> result) {
5467	>	super(m, p, b);
5468	>	this.searchFunction = searchFunction; this.result = result;
5469	>	}
5470	>	public final U getRawResult() { return result.get(); }
5471	>	@SuppressWarnings("unchecked") public final void compute() {
5472	>	final Fun<? super V, ? extends U> searchFunction;
5473	>	final AtomicReference<U> result;
5474	>	if ((searchFunction = this.searchFunction) != null &&
5475	>	(result = this.result) != null) {
5476	>	for (int b;;) {
5477	>	if (result.get() != null)
5478	>	return;
5479	>	if ((b = preSplit()) <= 0)
5480	>	break;
5481	>	new SearchValuesTask<K,V,U>
5482	>	(map, this, b, searchFunction, result).fork();
5483	>	}
5484	>	while (result.get() == null) {
5485	>	Object v; U u;
5486	>	if ((v = advance()) == null) {
5487	>	propagateCompletion();
5488	>	break;
5489	>	}
5490	>	if ((u = searchFunction.apply((V)v)) != null) {
5491	>	if (result.compareAndSet(null, u))
5492	>	quietlyCompleteRoot();
5493	>	break;
5494	>	}
5495	>	}
5496	>	}
5497	>	}
5498	>	}
5499
5500	<	// Initialize each segment to be minimally sized, and let grow.
5501	<	for (int i = 0; i < segments.length; ++i) {
5502	<	segments[i].setTable(new HashEntry[1]);
5500	>	@SuppressWarnings("serial") static final class SearchEntriesTask<K,V,U>
5501	>	extends Traverser<K,V,U> {
5502	>	final Fun<Entry<K,V>, ? extends U> searchFunction;
5503	>	final AtomicReference<U> result;
5504	>	SearchEntriesTask
5505	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5506	>	Fun<Entry<K,V>, ? extends U> searchFunction,
5507	>	AtomicReference<U> result) {
5508	>	super(m, p, b);
5509	>	this.searchFunction = searchFunction; this.result = result;
5510	>	}
5511	>	public final U getRawResult() { return result.get(); }
5512	>	@SuppressWarnings("unchecked") public final void compute() {
5513	>	final Fun<Entry<K,V>, ? extends U> searchFunction;
5514	>	final AtomicReference<U> result;
5515	>	if ((searchFunction = this.searchFunction) != null &&
5516	>	(result = this.result) != null) {
5517	>	for (int b;;) {
5518	>	if (result.get() != null)
5519	>	return;
5520	>	if ((b = preSplit()) <= 0)
5521	>	break;
5522	>	new SearchEntriesTask<K,V,U>
5523	>	(map, this, b, searchFunction, result).fork();
5524	>	}
5525	>	while (result.get() == null) {
5526	>	Object v; U u;
5527	>	if ((v = advance()) == null) {
5528	>	propagateCompletion();
5529	>	break;
5530	>	}
5531	>	if ((u = searchFunction.apply(entryFor((K)nextKey,
5532	>	(V)v))) != null) {
5533	>	if (result.compareAndSet(null, u))
5534	>	quietlyCompleteRoot();
5535	>	return;
5536	>	}
5537	>	}
5538	>	}
5539		}
5540	+	}
5541
5542	<	// Read the keys and values, and put the mappings in the table
5543	<	for (;;) {
5544	<	K key = (K) s.readObject();
5545	<	V value = (V) s.readObject();
5546	<	if (key == null)
5547	<	break;
5548	<	put(key, value);
5542	>	@SuppressWarnings("serial") static final class SearchMappingsTask<K,V,U>
5543	>	extends Traverser<K,V,U> {
5544	>	final BiFun<? super K, ? super V, ? extends U> searchFunction;
5545	>	final AtomicReference<U> result;
5546	>	SearchMappingsTask
5547	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5548	>	BiFun<? super K, ? super V, ? extends U> searchFunction,
5549	>	AtomicReference<U> result) {
5550	>	super(m, p, b);
5551	>	this.searchFunction = searchFunction; this.result = result;
5552	>	}
5553	>	public final U getRawResult() { return result.get(); }
5554	>	@SuppressWarnings("unchecked") public final void compute() {
5555	>	final BiFun<? super K, ? super V, ? extends U> searchFunction;
5556	>	final AtomicReference<U> result;
5557	>	if ((searchFunction = this.searchFunction) != null &&
5558	>	(result = this.result) != null) {
5559	>	for (int b;;) {
5560	>	if (result.get() != null)
5561	>	return;
5562	>	if ((b = preSplit()) <= 0)
5563	>	break;
5564	>	new SearchMappingsTask<K,V,U>
5565	>	(map, this, b, searchFunction, result).fork();
5566	>	}
5567	>	while (result.get() == null) {
5568	>	Object v; U u;
5569	>	if ((v = advance()) == null) {
5570	>	propagateCompletion();
5571	>	break;
5572	>	}
5573	>	if ((u = searchFunction.apply((K)nextKey, (V)v)) != null) {
5574	>	if (result.compareAndSet(null, u))
5575	>	quietlyCompleteRoot();
5576	>	break;
5577	>	}
5578	>	}
5579	>	}
5580	>	}
5581	>	}
5582	>
5583	>	@SuppressWarnings("serial") static final class ReduceKeysTask<K,V>
5584	>	extends Traverser<K,V,K> {
5585	>	final BiFun<? super K, ? super K, ? extends K> reducer;
5586	>	K result;
5587	>	ReduceKeysTask<K,V> rights, nextRight;
5588	>	ReduceKeysTask
5589	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5590	>	ReduceKeysTask<K,V> nextRight,
5591	>	BiFun<? super K, ? super K, ? extends K> reducer) {
5592	>	super(m, p, b); this.nextRight = nextRight;
5593	>	this.reducer = reducer;
5594	>	}
5595	>	public final K getRawResult() { return result; }
5596	>	@SuppressWarnings("unchecked") public final void compute() {
5597	>	final BiFun<? super K, ? super K, ? extends K> reducer;
5598	>	if ((reducer = this.reducer) != null) {
5599	>	for (int b; (b = preSplit()) > 0;)
5600	>	(rights = new ReduceKeysTask<K,V>
5601	>	(map, this, b, rights, reducer)).fork();
5602	>	K r = null;
5603	>	while (advance() != null) {
5604	>	K u = (K)nextKey;
5605	>	r = (r == null) ? u : reducer.apply(r, u);
5606	>	}
5607	>	result = r;
5608	>	CountedCompleter<?> c;
5609	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5610	>	ReduceKeysTask<K,V>
5611	>	t = (ReduceKeysTask<K,V>)c,
5612	>	s = t.rights;
5613	>	while (s != null) {
5614	>	K tr, sr;
5615	>	if ((sr = s.result) != null)
5616	>	t.result = (((tr = t.result) == null) ? sr :
5617	>	reducer.apply(tr, sr));
5618	>	s = t.rights = s.nextRight;
5619	>	}
5620	>	}
5621	>	}
5622	>	}
5623	>	}
5624	>
5625	>	@SuppressWarnings("serial") static final class ReduceValuesTask<K,V>
5626	>	extends Traverser<K,V,V> {
5627	>	final BiFun<? super V, ? super V, ? extends V> reducer;
5628	>	V result;
5629	>	ReduceValuesTask<K,V> rights, nextRight;
5630	>	ReduceValuesTask
5631	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5632	>	ReduceValuesTask<K,V> nextRight,
5633	>	BiFun<? super V, ? super V, ? extends V> reducer) {
5634	>	super(m, p, b); this.nextRight = nextRight;
5635	>	this.reducer = reducer;
5636	>	}
5637	>	public final V getRawResult() { return result; }
5638	>	@SuppressWarnings("unchecked") public final void compute() {
5639	>	final BiFun<? super V, ? super V, ? extends V> reducer;
5640	>	if ((reducer = this.reducer) != null) {
5641	>	for (int b; (b = preSplit()) > 0;)
5642	>	(rights = new ReduceValuesTask<K,V>
5643	>	(map, this, b, rights, reducer)).fork();
5644	>	V r = null;
5645	>	Object v;
5646	>	while ((v = advance()) != null) {
5647	>	V u = (V)v;
5648	>	r = (r == null) ? u : reducer.apply(r, u);
5649	>	}
5650	>	result = r;
5651	>	CountedCompleter<?> c;
5652	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5653	>	ReduceValuesTask<K,V>
5654	>	t = (ReduceValuesTask<K,V>)c,
5655	>	s = t.rights;
5656	>	while (s != null) {
5657	>	V tr, sr;
5658	>	if ((sr = s.result) != null)
5659	>	t.result = (((tr = t.result) == null) ? sr :
5660	>	reducer.apply(tr, sr));
5661	>	s = t.rights = s.nextRight;
5662	>	}
5663	>	}
5664	>	}
5665	>	}
5666	>	}
5667	>
5668	>	@SuppressWarnings("serial") static final class ReduceEntriesTask<K,V>
5669	>	extends Traverser<K,V,Map.Entry<K,V>> {
5670	>	final BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer;
5671	>	Map.Entry<K,V> result;
5672	>	ReduceEntriesTask<K,V> rights, nextRight;
5673	>	ReduceEntriesTask
5674	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5675	>	ReduceEntriesTask<K,V> nextRight,
5676	>	BiFun<Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer) {
5677	>	super(m, p, b); this.nextRight = nextRight;
5678	>	this.reducer = reducer;
5679	>	}
5680	>	public final Map.Entry<K,V> getRawResult() { return result; }
5681	>	@SuppressWarnings("unchecked") public final void compute() {
5682	>	final BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer;
5683	>	if ((reducer = this.reducer) != null) {
5684	>	for (int b; (b = preSplit()) > 0;)
5685	>	(rights = new ReduceEntriesTask<K,V>
5686	>	(map, this, b, rights, reducer)).fork();
5687	>	Map.Entry<K,V> r = null;
5688	>	Object v;
5689	>	while ((v = advance()) != null) {
5690	>	Map.Entry<K,V> u = entryFor((K)nextKey, (V)v);
5691	>	r = (r == null) ? u : reducer.apply(r, u);
5692	>	}
5693	>	result = r;
5694	>	CountedCompleter<?> c;
5695	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5696	>	ReduceEntriesTask<K,V>
5697	>	t = (ReduceEntriesTask<K,V>)c,
5698	>	s = t.rights;
5699	>	while (s != null) {
5700	>	Map.Entry<K,V> tr, sr;
5701	>	if ((sr = s.result) != null)
5702	>	t.result = (((tr = t.result) == null) ? sr :
5703	>	reducer.apply(tr, sr));
5704	>	s = t.rights = s.nextRight;
5705	>	}
5706	>	}
5707	>	}
5708	>	}
5709	>	}
5710	>
5711	>	@SuppressWarnings("serial") static final class MapReduceKeysTask<K,V,U>
5712	>	extends Traverser<K,V,U> {
5713	>	final Fun<? super K, ? extends U> transformer;
5714	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5715	>	U result;
5716	>	MapReduceKeysTask<K,V,U> rights, nextRight;
5717	>	MapReduceKeysTask
5718	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5719	>	MapReduceKeysTask<K,V,U> nextRight,
5720	>	Fun<? super K, ? extends U> transformer,
5721	>	BiFun<? super U, ? super U, ? extends U> reducer) {
5722	>	super(m, p, b); this.nextRight = nextRight;
5723	>	this.transformer = transformer;
5724	>	this.reducer = reducer;
5725	>	}
5726	>	public final U getRawResult() { return result; }
5727	>	@SuppressWarnings("unchecked") public final void compute() {
5728	>	final Fun<? super K, ? extends U> transformer;
5729	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5730	>	if ((transformer = this.transformer) != null &&
5731	>	(reducer = this.reducer) != null) {
5732	>	for (int b; (b = preSplit()) > 0;)
5733	>	(rights = new MapReduceKeysTask<K,V,U>
5734	>	(map, this, b, rights, transformer, reducer)).fork();
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	>	CountedCompleter<?> c;
5742	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5743	>	MapReduceKeysTask<K,V,U>
5744	>	t = (MapReduceKeysTask<K,V,U>)c,
5745	>	s = t.rights;
5746	>	while (s != null) {
5747	>	U tr, sr;
5748	>	if ((sr = s.result) != null)
5749	>	t.result = (((tr = t.result) == null) ? sr :
5750	>	reducer.apply(tr, sr));
5751	>	s = t.rights = s.nextRight;
5752	>	}
5753	>	}
5754	>	}
5755	>	}
5756	>	}
5757	>
5758	>	@SuppressWarnings("serial") static final class MapReduceValuesTask<K,V,U>
5759	>	extends Traverser<K,V,U> {
5760	>	final Fun<? super V, ? extends U> transformer;
5761	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5762	>	U result;
5763	>	MapReduceValuesTask<K,V,U> rights, nextRight;
5764	>	MapReduceValuesTask
5765	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5766	>	MapReduceValuesTask<K,V,U> nextRight,
5767	>	Fun<? super V, ? extends U> transformer,
5768	>	BiFun<? super U, ? super U, ? extends U> reducer) {
5769	>	super(m, p, b); this.nextRight = nextRight;
5770	>	this.transformer = transformer;
5771	>	this.reducer = reducer;
5772	>	}
5773	>	public final U getRawResult() { return result; }
5774	>	@SuppressWarnings("unchecked") public final void compute() {
5775	>	final Fun<? super V, ? extends U> transformer;
5776	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5777	>	if ((transformer = this.transformer) != null &&
5778	>	(reducer = this.reducer) != null) {
5779	>	for (int b; (b = preSplit()) > 0;)
5780	>	(rights = new MapReduceValuesTask<K,V,U>
5781	>	(map, this, b, rights, transformer, reducer)).fork();
5782	>	U r = null, u;
5783	>	Object v;
5784	>	while ((v = advance()) != null) {
5785	>	if ((u = transformer.apply((V)v)) != null)
5786	>	r = (r == null) ? u : reducer.apply(r, u);
5787	>	}
5788	>	result = r;
5789	>	CountedCompleter<?> c;
5790	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5791	>	MapReduceValuesTask<K,V,U>
5792	>	t = (MapReduceValuesTask<K,V,U>)c,
5793	>	s = t.rights;
5794	>	while (s != null) {
5795	>	U tr, sr;
5796	>	if ((sr = s.result) != null)
5797	>	t.result = (((tr = t.result) == null) ? sr :
5798	>	reducer.apply(tr, sr));
5799	>	s = t.rights = s.nextRight;
5800	>	}
5801	>	}
5802	>	}
5803	>	}
5804	>	}
5805	>
5806	>	@SuppressWarnings("serial") static final class MapReduceEntriesTask<K,V,U>
5807	>	extends Traverser<K,V,U> {
5808	>	final Fun<Map.Entry<K,V>, ? extends U> transformer;
5809	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5810	>	U result;
5811	>	MapReduceEntriesTask<K,V,U> rights, nextRight;
5812	>	MapReduceEntriesTask
5813	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5814	>	MapReduceEntriesTask<K,V,U> nextRight,
5815	>	Fun<Map.Entry<K,V>, ? extends U> transformer,
5816	>	BiFun<? super U, ? super U, ? extends U> reducer) {
5817	>	super(m, p, b); this.nextRight = nextRight;
5818	>	this.transformer = transformer;
5819	>	this.reducer = reducer;
5820	>	}
5821	>	public final U getRawResult() { return result; }
5822	>	@SuppressWarnings("unchecked") public final void compute() {
5823	>	final Fun<Map.Entry<K,V>, ? extends U> transformer;
5824	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5825	>	if ((transformer = this.transformer) != null &&
5826	>	(reducer = this.reducer) != null) {
5827	>	for (int b; (b = preSplit()) > 0;)
5828	>	(rights = new MapReduceEntriesTask<K,V,U>
5829	>	(map, this, b, rights, transformer, reducer)).fork();
5830	>	U r = null, u;
5831	>	Object v;
5832	>	while ((v = advance()) != null) {
5833	>	if ((u = transformer.apply(entryFor((K)nextKey,
5834	>	(V)v))) != null)
5835	>	r = (r == null) ? u : reducer.apply(r, u);
5836	>	}
5837	>	result = r;
5838	>	CountedCompleter<?> c;
5839	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5840	>	MapReduceEntriesTask<K,V,U>
5841	>	t = (MapReduceEntriesTask<K,V,U>)c,
5842	>	s = t.rights;
5843	>	while (s != null) {
5844	>	U tr, sr;
5845	>	if ((sr = s.result) != null)
5846	>	t.result = (((tr = t.result) == null) ? sr :
5847	>	reducer.apply(tr, sr));
5848	>	s = t.rights = s.nextRight;
5849	>	}
5850	>	}
5851	>	}
5852	>	}
5853	>	}
5854	>
5855	>	@SuppressWarnings("serial") static final class MapReduceMappingsTask<K,V,U>
5856	>	extends Traverser<K,V,U> {
5857	>	final BiFun<? super K, ? super V, ? extends U> transformer;
5858	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5859	>	U result;
5860	>	MapReduceMappingsTask<K,V,U> rights, nextRight;
5861	>	MapReduceMappingsTask
5862	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5863	>	MapReduceMappingsTask<K,V,U> nextRight,
5864	>	BiFun<? super K, ? super V, ? extends U> transformer,
5865	>	BiFun<? super U, ? super U, ? extends U> reducer) {
5866	>	super(m, p, b); this.nextRight = nextRight;
5867	>	this.transformer = transformer;
5868	>	this.reducer = reducer;
5869	>	}
5870	>	public final U getRawResult() { return result; }
5871	>	@SuppressWarnings("unchecked") public final void compute() {
5872	>	final BiFun<? super K, ? super V, ? extends U> transformer;
5873	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5874	>	if ((transformer = this.transformer) != null &&
5875	>	(reducer = this.reducer) != null) {
5876	>	for (int b; (b = preSplit()) > 0;)
5877	>	(rights = new MapReduceMappingsTask<K,V,U>
5878	>	(map, this, b, rights, transformer, reducer)).fork();
5879	>	U r = null, u;
5880	>	Object v;
5881	>	while ((v = advance()) != null) {
5882	>	if ((u = transformer.apply((K)nextKey, (V)v)) != null)
5883	>	r = (r == null) ? u : reducer.apply(r, u);
5884	>	}
5885	>	result = r;
5886	>	CountedCompleter<?> c;
5887	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5888	>	MapReduceMappingsTask<K,V,U>
5889	>	t = (MapReduceMappingsTask<K,V,U>)c,
5890	>	s = t.rights;
5891	>	while (s != null) {
5892	>	U tr, sr;
5893	>	if ((sr = s.result) != null)
5894	>	t.result = (((tr = t.result) == null) ? sr :
5895	>	reducer.apply(tr, sr));
5896	>	s = t.rights = s.nextRight;
5897	>	}
5898	>	}
5899	>	}
5900	>	}
5901	>	}
5902	>
5903	>	@SuppressWarnings("serial") static final class MapReduceKeysToDoubleTask<K,V>
5904	>	extends Traverser<K,V,Double> {
5905	>	final ObjectToDouble<? super K> transformer;
5906	>	final DoubleByDoubleToDouble reducer;
5907	>	final double basis;
5908	>	double result;
5909	>	MapReduceKeysToDoubleTask<K,V> rights, nextRight;
5910	>	MapReduceKeysToDoubleTask
5911	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5912	>	MapReduceKeysToDoubleTask<K,V> nextRight,
5913	>	ObjectToDouble<? super K> transformer,
5914	>	double basis,
5915	>	DoubleByDoubleToDouble reducer) {
5916	>	super(m, p, b); this.nextRight = nextRight;
5917	>	this.transformer = transformer;
5918	>	this.basis = basis; this.reducer = reducer;
5919	>	}
5920	>	public final Double getRawResult() { return result; }
5921	>	@SuppressWarnings("unchecked") public final void compute() {
5922	>	final ObjectToDouble<? super K> transformer;
5923	>	final DoubleByDoubleToDouble reducer;
5924	>	if ((transformer = this.transformer) != null &&
5925	>	(reducer = this.reducer) != null) {
5926	>	double r = this.basis;
5927	>	for (int b; (b = preSplit()) > 0;)
5928	>	(rights = new MapReduceKeysToDoubleTask<K,V>
5929	>	(map, this, b, rights, transformer, r, reducer)).fork();
5930	>	while (advance() != null)
5931	>	r = reducer.apply(r, transformer.apply((K)nextKey));
5932	>	result = r;
5933	>	CountedCompleter<?> c;
5934	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5935	>	MapReduceKeysToDoubleTask<K,V>
5936	>	t = (MapReduceKeysToDoubleTask<K,V>)c,
5937	>	s = t.rights;
5938	>	while (s != null) {
5939	>	t.result = reducer.apply(t.result, s.result);
5940	>	s = t.rights = s.nextRight;
5941	>	}
5942	>	}
5943	>	}
5944	>	}
5945	>	}
5946	>
5947	>	@SuppressWarnings("serial") static final class MapReduceValuesToDoubleTask<K,V>
5948	>	extends Traverser<K,V,Double> {
5949	>	final ObjectToDouble<? super V> transformer;
5950	>	final DoubleByDoubleToDouble reducer;
5951	>	final double basis;
5952	>	double result;
5953	>	MapReduceValuesToDoubleTask<K,V> rights, nextRight;
5954	>	MapReduceValuesToDoubleTask
5955	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5956	>	MapReduceValuesToDoubleTask<K,V> nextRight,
5957	>	ObjectToDouble<? super V> transformer,
5958	>	double basis,
5959	>	DoubleByDoubleToDouble reducer) {
5960	>	super(m, p, b); this.nextRight = nextRight;
5961	>	this.transformer = transformer;
5962	>	this.basis = basis; this.reducer = reducer;
5963	>	}
5964	>	public final Double getRawResult() { return result; }
5965	>	@SuppressWarnings("unchecked") public final void compute() {
5966	>	final ObjectToDouble<? super V> transformer;
5967	>	final DoubleByDoubleToDouble reducer;
5968	>	if ((transformer = this.transformer) != null &&
5969	>	(reducer = this.reducer) != null) {
5970	>	double r = this.basis;
5971	>	for (int b; (b = preSplit()) > 0;)
5972	>	(rights = new MapReduceValuesToDoubleTask<K,V>
5973	>	(map, this, b, rights, transformer, r, reducer)).fork();
5974	>	Object v;
5975	>	while ((v = advance()) != null)
5976	>	r = reducer.apply(r, transformer.apply((V)v));
5977	>	result = r;
5978	>	CountedCompleter<?> c;
5979	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5980	>	MapReduceValuesToDoubleTask<K,V>
5981	>	t = (MapReduceValuesToDoubleTask<K,V>)c,
5982	>	s = t.rights;
5983	>	while (s != null) {
5984	>	t.result = reducer.apply(t.result, s.result);
5985	>	s = t.rights = s.nextRight;
5986	>	}
5987	>	}
5988	>	}
5989	>	}
5990	>	}
5991	>
5992	>	@SuppressWarnings("serial") static final class MapReduceEntriesToDoubleTask<K,V>
5993	>	extends Traverser<K,V,Double> {
5994	>	final ObjectToDouble<Map.Entry<K,V>> transformer;
5995	>	final DoubleByDoubleToDouble reducer;
5996	>	final double basis;
5997	>	double result;
5998	>	MapReduceEntriesToDoubleTask<K,V> rights, nextRight;
5999	>	MapReduceEntriesToDoubleTask
6000	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6001	>	MapReduceEntriesToDoubleTask<K,V> nextRight,
6002	>	ObjectToDouble<Map.Entry<K,V>> transformer,
6003	>	double basis,
6004	>	DoubleByDoubleToDouble reducer) {
6005	>	super(m, p, b); this.nextRight = nextRight;
6006	>	this.transformer = transformer;
6007	>	this.basis = basis; this.reducer = reducer;
6008	>	}
6009	>	public final Double getRawResult() { return result; }
6010	>	@SuppressWarnings("unchecked") public final void compute() {
6011	>	final ObjectToDouble<Map.Entry<K,V>> transformer;
6012	>	final DoubleByDoubleToDouble reducer;
6013	>	if ((transformer = this.transformer) != null &&
6014	>	(reducer = this.reducer) != null) {
6015	>	double r = this.basis;
6016	>	for (int b; (b = preSplit()) > 0;)
6017	>	(rights = new MapReduceEntriesToDoubleTask<K,V>
6018	>	(map, this, b, rights, transformer, r, reducer)).fork();
6019	>	Object v;
6020	>	while ((v = advance()) != null)
6021	>	r = reducer.apply(r, transformer.apply(entryFor((K)nextKey,
6022	>	(V)v)));
6023	>	result = r;
6024	>	CountedCompleter<?> c;
6025	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6026	>	MapReduceEntriesToDoubleTask<K,V>
6027	>	t = (MapReduceEntriesToDoubleTask<K,V>)c,
6028	>	s = t.rights;
6029	>	while (s != null) {
6030	>	t.result = reducer.apply(t.result, s.result);
6031	>	s = t.rights = s.nextRight;
6032	>	}
6033	>	}
6034	>	}
6035	>	}
6036	>	}
6037	>
6038	>	@SuppressWarnings("serial") static final class MapReduceMappingsToDoubleTask<K,V>
6039	>	extends Traverser<K,V,Double> {
6040	>	final ObjectByObjectToDouble<? super K, ? super V> transformer;
6041	>	final DoubleByDoubleToDouble reducer;
6042	>	final double basis;
6043	>	double result;
6044	>	MapReduceMappingsToDoubleTask<K,V> rights, nextRight;
6045	>	MapReduceMappingsToDoubleTask
6046	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6047	>	MapReduceMappingsToDoubleTask<K,V> nextRight,
6048	>	ObjectByObjectToDouble<? super K, ? super V> transformer,
6049	>	double basis,
6050	>	DoubleByDoubleToDouble reducer) {
6051	>	super(m, p, b); this.nextRight = nextRight;
6052	>	this.transformer = transformer;
6053	>	this.basis = basis; this.reducer = reducer;
6054	>	}
6055	>	public final Double getRawResult() { return result; }
6056	>	@SuppressWarnings("unchecked") public final void compute() {
6057	>	final ObjectByObjectToDouble<? super K, ? super V> transformer;
6058	>	final DoubleByDoubleToDouble reducer;
6059	>	if ((transformer = this.transformer) != null &&
6060	>	(reducer = this.reducer) != null) {
6061	>	double r = this.basis;
6062	>	for (int b; (b = preSplit()) > 0;)
6063	>	(rights = new MapReduceMappingsToDoubleTask<K,V>
6064	>	(map, this, b, rights, transformer, r, reducer)).fork();
6065	>	Object v;
6066	>	while ((v = advance()) != null)
6067	>	r = reducer.apply(r, transformer.apply((K)nextKey, (V)v));
6068	>	result = r;
6069	>	CountedCompleter<?> c;
6070	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6071	>	MapReduceMappingsToDoubleTask<K,V>
6072	>	t = (MapReduceMappingsToDoubleTask<K,V>)c,
6073	>	s = t.rights;
6074	>	while (s != null) {
6075	>	t.result = reducer.apply(t.result, s.result);
6076	>	s = t.rights = s.nextRight;
6077	>	}
6078	>	}
6079	>	}
6080	>	}
6081	>	}
6082	>
6083	>	@SuppressWarnings("serial") static final class MapReduceKeysToLongTask<K,V>
6084	>	extends Traverser<K,V,Long> {
6085	>	final ObjectToLong<? super K> transformer;
6086	>	final LongByLongToLong reducer;
6087	>	final long basis;
6088	>	long result;
6089	>	MapReduceKeysToLongTask<K,V> rights, nextRight;
6090	>	MapReduceKeysToLongTask
6091	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6092	>	MapReduceKeysToLongTask<K,V> nextRight,
6093	>	ObjectToLong<? super K> transformer,
6094	>	long basis,
6095	>	LongByLongToLong reducer) {
6096	>	super(m, p, b); this.nextRight = nextRight;
6097	>	this.transformer = transformer;
6098	>	this.basis = basis; this.reducer = reducer;
6099	>	}
6100	>	public final Long getRawResult() { return result; }
6101	>	@SuppressWarnings("unchecked") public final void compute() {
6102	>	final ObjectToLong<? super K> transformer;
6103	>	final LongByLongToLong reducer;
6104	>	if ((transformer = this.transformer) != null &&
6105	>	(reducer = this.reducer) != null) {
6106	>	long r = this.basis;
6107	>	for (int b; (b = preSplit()) > 0;)
6108	>	(rights = new MapReduceKeysToLongTask<K,V>
6109	>	(map, this, b, rights, transformer, r, reducer)).fork();
6110	>	while (advance() != null)
6111	>	r = reducer.apply(r, transformer.apply((K)nextKey));
6112	>	result = r;
6113	>	CountedCompleter<?> c;
6114	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6115	>	MapReduceKeysToLongTask<K,V>
6116	>	t = (MapReduceKeysToLongTask<K,V>)c,
6117	>	s = t.rights;
6118	>	while (s != null) {
6119	>	t.result = reducer.apply(t.result, s.result);
6120	>	s = t.rights = s.nextRight;
6121	>	}
6122	>	}
6123	>	}
6124	>	}
6125	>	}
6126	>
6127	>	@SuppressWarnings("serial") static final class MapReduceValuesToLongTask<K,V>
6128	>	extends Traverser<K,V,Long> {
6129	>	final ObjectToLong<? super V> transformer;
6130	>	final LongByLongToLong reducer;
6131	>	final long basis;
6132	>	long result;
6133	>	MapReduceValuesToLongTask<K,V> rights, nextRight;
6134	>	MapReduceValuesToLongTask
6135	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6136	>	MapReduceValuesToLongTask<K,V> nextRight,
6137	>	ObjectToLong<? super V> transformer,
6138	>	long basis,
6139	>	LongByLongToLong reducer) {
6140	>	super(m, p, b); this.nextRight = nextRight;
6141	>	this.transformer = transformer;
6142	>	this.basis = basis; this.reducer = reducer;
6143	>	}
6144	>	public final Long getRawResult() { return result; }
6145	>	@SuppressWarnings("unchecked") public final void compute() {
6146	>	final ObjectToLong<? super V> transformer;
6147	>	final LongByLongToLong reducer;
6148	>	if ((transformer = this.transformer) != null &&
6149	>	(reducer = this.reducer) != null) {
6150	>	long r = this.basis;
6151	>	for (int b; (b = preSplit()) > 0;)
6152	>	(rights = new MapReduceValuesToLongTask<K,V>
6153	>	(map, this, b, rights, transformer, r, reducer)).fork();
6154	>	Object v;
6155	>	while ((v = advance()) != null)
6156	>	r = reducer.apply(r, transformer.apply((V)v));
6157	>	result = r;
6158	>	CountedCompleter<?> c;
6159	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6160	>	MapReduceValuesToLongTask<K,V>
6161	>	t = (MapReduceValuesToLongTask<K,V>)c,
6162	>	s = t.rights;
6163	>	while (s != null) {
6164	>	t.result = reducer.apply(t.result, s.result);
6165	>	s = t.rights = s.nextRight;
6166	>	}
6167	>	}
6168	>	}
6169	>	}
6170	>	}
6171	>
6172	>	@SuppressWarnings("serial") static final class MapReduceEntriesToLongTask<K,V>
6173	>	extends Traverser<K,V,Long> {
6174	>	final ObjectToLong<Map.Entry<K,V>> transformer;
6175	>	final LongByLongToLong reducer;
6176	>	final long basis;
6177	>	long result;
6178	>	MapReduceEntriesToLongTask<K,V> rights, nextRight;
6179	>	MapReduceEntriesToLongTask
6180	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6181	>	MapReduceEntriesToLongTask<K,V> nextRight,
6182	>	ObjectToLong<Map.Entry<K,V>> transformer,
6183	>	long basis,
6184	>	LongByLongToLong reducer) {
6185	>	super(m, p, b); this.nextRight = nextRight;
6186	>	this.transformer = transformer;
6187	>	this.basis = basis; this.reducer = reducer;
6188	>	}
6189	>	public final Long getRawResult() { return result; }
6190	>	@SuppressWarnings("unchecked") public final void compute() {
6191	>	final ObjectToLong<Map.Entry<K,V>> transformer;
6192	>	final LongByLongToLong reducer;
6193	>	if ((transformer = this.transformer) != null &&
6194	>	(reducer = this.reducer) != null) {
6195	>	long r = this.basis;
6196	>	for (int b; (b = preSplit()) > 0;)
6197	>	(rights = new MapReduceEntriesToLongTask<K,V>
6198	>	(map, this, b, rights, transformer, r, reducer)).fork();
6199	>	Object v;
6200	>	while ((v = advance()) != null)
6201	>	r = reducer.apply(r, transformer.apply(entryFor((K)nextKey,
6202	>	(V)v)));
6203	>	result = r;
6204	>	CountedCompleter<?> c;
6205	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6206	>	MapReduceEntriesToLongTask<K,V>
6207	>	t = (MapReduceEntriesToLongTask<K,V>)c,
6208	>	s = t.rights;
6209	>	while (s != null) {
6210	>	t.result = reducer.apply(t.result, s.result);
6211	>	s = t.rights = s.nextRight;
6212	>	}
6213	>	}
6214	>	}
6215	>	}
6216	>	}
6217	>
6218	>	@SuppressWarnings("serial") static final class MapReduceMappingsToLongTask<K,V>
6219	>	extends Traverser<K,V,Long> {
6220	>	final ObjectByObjectToLong<? super K, ? super V> transformer;
6221	>	final LongByLongToLong reducer;
6222	>	final long basis;
6223	>	long result;
6224	>	MapReduceMappingsToLongTask<K,V> rights, nextRight;
6225	>	MapReduceMappingsToLongTask
6226	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6227	>	MapReduceMappingsToLongTask<K,V> nextRight,
6228	>	ObjectByObjectToLong<? super K, ? super V> transformer,
6229	>	long basis,
6230	>	LongByLongToLong reducer) {
6231	>	super(m, p, b); this.nextRight = nextRight;
6232	>	this.transformer = transformer;
6233	>	this.basis = basis; this.reducer = reducer;
6234	>	}
6235	>	public final Long getRawResult() { return result; }
6236	>	@SuppressWarnings("unchecked") public final void compute() {
6237	>	final ObjectByObjectToLong<? super K, ? super V> transformer;
6238	>	final LongByLongToLong reducer;
6239	>	if ((transformer = this.transformer) != null &&
6240	>	(reducer = this.reducer) != null) {
6241	>	long r = this.basis;
6242	>	for (int b; (b = preSplit()) > 0;)
6243	>	(rights = new MapReduceMappingsToLongTask<K,V>
6244	>	(map, this, b, rights, transformer, r, reducer)).fork();
6245	>	Object v;
6246	>	while ((v = advance()) != null)
6247	>	r = reducer.apply(r, transformer.apply((K)nextKey, (V)v));
6248	>	result = r;
6249	>	CountedCompleter<?> c;
6250	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6251	>	MapReduceMappingsToLongTask<K,V>
6252	>	t = (MapReduceMappingsToLongTask<K,V>)c,
6253	>	s = t.rights;
6254	>	while (s != null) {
6255	>	t.result = reducer.apply(t.result, s.result);
6256	>	s = t.rights = s.nextRight;
6257	>	}
6258	>	}
6259	>	}
6260	>	}
6261	>	}
6262	>
6263	>	@SuppressWarnings("serial") static final class MapReduceKeysToIntTask<K,V>
6264	>	extends Traverser<K,V,Integer> {
6265	>	final ObjectToInt<? super K> transformer;
6266	>	final IntByIntToInt reducer;
6267	>	final int basis;
6268	>	int result;
6269	>	MapReduceKeysToIntTask<K,V> rights, nextRight;
6270	>	MapReduceKeysToIntTask
6271	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6272	>	MapReduceKeysToIntTask<K,V> nextRight,
6273	>	ObjectToInt<? super K> transformer,
6274	>	int basis,
6275	>	IntByIntToInt reducer) {
6276	>	super(m, p, b); this.nextRight = nextRight;
6277	>	this.transformer = transformer;
6278	>	this.basis = basis; this.reducer = reducer;
6279	>	}
6280	>	public final Integer getRawResult() { return result; }
6281	>	@SuppressWarnings("unchecked") public final void compute() {
6282	>	final ObjectToInt<? super K> transformer;
6283	>	final IntByIntToInt reducer;
6284	>	if ((transformer = this.transformer) != null &&
6285	>	(reducer = this.reducer) != null) {
6286	>	int r = this.basis;
6287	>	for (int b; (b = preSplit()) > 0;)
6288	>	(rights = new MapReduceKeysToIntTask<K,V>
6289	>	(map, this, b, rights, transformer, r, reducer)).fork();
6290	>	while (advance() != null)
6291	>	r = reducer.apply(r, transformer.apply((K)nextKey));
6292	>	result = r;
6293	>	CountedCompleter<?> c;
6294	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6295	>	MapReduceKeysToIntTask<K,V>
6296	>	t = (MapReduceKeysToIntTask<K,V>)c,
6297	>	s = t.rights;
6298	>	while (s != null) {
6299	>	t.result = reducer.apply(t.result, s.result);
6300	>	s = t.rights = s.nextRight;
6301	>	}
6302	>	}
6303	>	}
6304	>	}
6305	>	}
6306	>
6307	>	@SuppressWarnings("serial") static final class MapReduceValuesToIntTask<K,V>
6308	>	extends Traverser<K,V,Integer> {
6309	>	final ObjectToInt<? super V> transformer;
6310	>	final IntByIntToInt reducer;
6311	>	final int basis;
6312	>	int result;
6313	>	MapReduceValuesToIntTask<K,V> rights, nextRight;
6314	>	MapReduceValuesToIntTask
6315	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6316	>	MapReduceValuesToIntTask<K,V> nextRight,
6317	>	ObjectToInt<? super V> transformer,
6318	>	int basis,
6319	>	IntByIntToInt reducer) {
6320	>	super(m, p, b); this.nextRight = nextRight;
6321	>	this.transformer = transformer;
6322	>	this.basis = basis; this.reducer = reducer;
6323	>	}
6324	>	public final Integer getRawResult() { return result; }
6325	>	@SuppressWarnings("unchecked") public final void compute() {
6326	>	final ObjectToInt<? super V> transformer;
6327	>	final IntByIntToInt reducer;
6328	>	if ((transformer = this.transformer) != null &&
6329	>	(reducer = this.reducer) != null) {
6330	>	int r = this.basis;
6331	>	for (int b; (b = preSplit()) > 0;)
6332	>	(rights = new MapReduceValuesToIntTask<K,V>
6333	>	(map, this, b, rights, transformer, r, reducer)).fork();
6334	>	Object v;
6335	>	while ((v = advance()) != null)
6336	>	r = reducer.apply(r, transformer.apply((V)v));
6337	>	result = r;
6338	>	CountedCompleter<?> c;
6339	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6340	>	MapReduceValuesToIntTask<K,V>
6341	>	t = (MapReduceValuesToIntTask<K,V>)c,
6342	>	s = t.rights;
6343	>	while (s != null) {
6344	>	t.result = reducer.apply(t.result, s.result);
6345	>	s = t.rights = s.nextRight;
6346	>	}
6347	>	}
6348	>	}
6349	>	}
6350	>	}
6351	>
6352	>	@SuppressWarnings("serial") static final class MapReduceEntriesToIntTask<K,V>
6353	>	extends Traverser<K,V,Integer> {
6354	>	final ObjectToInt<Map.Entry<K,V>> transformer;
6355	>	final IntByIntToInt reducer;
6356	>	final int basis;
6357	>	int result;
6358	>	MapReduceEntriesToIntTask<K,V> rights, nextRight;
6359	>	MapReduceEntriesToIntTask
6360	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6361	>	MapReduceEntriesToIntTask<K,V> nextRight,
6362	>	ObjectToInt<Map.Entry<K,V>> transformer,
6363	>	int basis,
6364	>	IntByIntToInt reducer) {
6365	>	super(m, p, b); this.nextRight = nextRight;
6366	>	this.transformer = transformer;
6367	>	this.basis = basis; this.reducer = reducer;
6368	>	}
6369	>	public final Integer getRawResult() { return result; }
6370	>	@SuppressWarnings("unchecked") public final void compute() {
6371	>	final ObjectToInt<Map.Entry<K,V>> transformer;
6372	>	final IntByIntToInt reducer;
6373	>	if ((transformer = this.transformer) != null &&
6374	>	(reducer = this.reducer) != null) {
6375	>	int r = this.basis;
6376	>	for (int b; (b = preSplit()) > 0;)
6377	>	(rights = new MapReduceEntriesToIntTask<K,V>
6378	>	(map, this, b, rights, transformer, r, reducer)).fork();
6379	>	Object v;
6380	>	while ((v = advance()) != null)
6381	>	r = reducer.apply(r, transformer.apply(entryFor((K)nextKey,
6382	>	(V)v)));
6383	>	result = r;
6384	>	CountedCompleter<?> c;
6385	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6386	>	MapReduceEntriesToIntTask<K,V>
6387	>	t = (MapReduceEntriesToIntTask<K,V>)c,
6388	>	s = t.rights;
6389	>	while (s != null) {
6390	>	t.result = reducer.apply(t.result, s.result);
6391	>	s = t.rights = s.nextRight;
6392	>	}
6393	>	}
6394	>	}
6395	>	}
6396	>	}
6397	>
6398	>	@SuppressWarnings("serial") static final class MapReduceMappingsToIntTask<K,V>
6399	>	extends Traverser<K,V,Integer> {
6400	>	final ObjectByObjectToInt<? super K, ? super V> transformer;
6401	>	final IntByIntToInt reducer;
6402	>	final int basis;
6403	>	int result;
6404	>	MapReduceMappingsToIntTask<K,V> rights, nextRight;
6405	>	MapReduceMappingsToIntTask
6406	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6407	>	MapReduceMappingsToIntTask<K,V> nextRight,
6408	>	ObjectByObjectToInt<? super K, ? super V> transformer,
6409	>	int basis,
6410	>	IntByIntToInt reducer) {
6411	>	super(m, p, b); this.nextRight = nextRight;
6412	>	this.transformer = transformer;
6413	>	this.basis = basis; this.reducer = reducer;
6414	>	}
6415	>	public final Integer getRawResult() { return result; }
6416	>	@SuppressWarnings("unchecked") public final void compute() {
6417	>	final ObjectByObjectToInt<? super K, ? super V> transformer;
6418	>	final IntByIntToInt reducer;
6419	>	if ((transformer = this.transformer) != null &&
6420	>	(reducer = this.reducer) != null) {
6421	>	int r = this.basis;
6422	>	for (int b; (b = preSplit()) > 0;)
6423	>	(rights = new MapReduceMappingsToIntTask<K,V>
6424	>	(map, this, b, rights, transformer, r, reducer)).fork();
6425	>	Object v;
6426	>	while ((v = advance()) != null)
6427	>	r = reducer.apply(r, transformer.apply((K)nextKey, (V)v));
6428	>	result = r;
6429	>	CountedCompleter<?> c;
6430	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6431	>	MapReduceMappingsToIntTask<K,V>
6432	>	t = (MapReduceMappingsToIntTask<K,V>)c,
6433	>	s = t.rights;
6434	>	while (s != null) {
6435	>	t.result = reducer.apply(t.result, s.result);
6436	>	s = t.rights = s.nextRight;
6437	>	}
6438	>	}
6439	>	}
6440	>	}
6441	>	}
6442	>
6443	>	// Unsafe mechanics
6444	>	private static final sun.misc.Unsafe U;
6445	>	private static final long SIZECTL;
6446	>	private static final long TRANSFERINDEX;
6447	>	private static final long TRANSFERORIGIN;
6448	>	private static final long BASECOUNT;
6449	>	private static final long COUNTERBUSY;
6450	>	private static final long CELLVALUE;
6451	>	private static final long ABASE;
6452	>	private static final int ASHIFT;
6453	>
6454	>	static {
6455	>	int ss;
6456	>	try {
6457	>	U = sun.misc.Unsafe.getUnsafe();
6458	>	Class<?> k = ConcurrentHashMap.class;
6459	>	SIZECTL = U.objectFieldOffset
6460	>	(k.getDeclaredField("sizeCtl"));
6461	>	TRANSFERINDEX = U.objectFieldOffset
6462	>	(k.getDeclaredField("transferIndex"));
6463	>	TRANSFERORIGIN = U.objectFieldOffset
6464	>	(k.getDeclaredField("transferOrigin"));
6465	>	BASECOUNT = U.objectFieldOffset
6466	>	(k.getDeclaredField("baseCount"));
6467	>	COUNTERBUSY = U.objectFieldOffset
6468	>	(k.getDeclaredField("counterBusy"));
6469	>	Class<?> ck = CounterCell.class;
6470	>	CELLVALUE = U.objectFieldOffset
6471	>	(ck.getDeclaredField("value"));
6472	>	Class<?> sc = Node[].class;
6473	>	ABASE = U.arrayBaseOffset(sc);
6474	>	ss = U.arrayIndexScale(sc);
6475	>	ASHIFT = 31 - Integer.numberOfLeadingZeros(ss);
6476	>	} catch (Exception e) {
6477	>	throw new Error(e);
6478		}
6479	+	if ((ss & (ss-1)) != 0)
6480	+	throw new Error("data type scale not a power of two");
6481		}
6482		}

Diff Legend

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