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 1.115 by jsr166, Fri Dec 2 14:28:17 2011 UTC vs.
Revision 1.148 by dl, Sat Dec 8 14:10:42 2012 UTC

#	Line 5 | Line 5
5		*/
6
7		package java.util.concurrent;
8	<	import java.util.concurrent.locks.*;
9	<	import java.util.*;
8	>	import java.util.concurrent.atomic.LongAdder;
9	>	import java.util.concurrent.ForkJoinPool;
10	>	import java.util.concurrent.CountedCompleter;
11	>
12	>	import java.util.Comparator;
13	>	import java.util.Arrays;
14	>	import java.util.Map;
15	>	import java.util.Set;
16	>	import java.util.Collection;
17	>	import java.util.AbstractMap;
18	>	import java.util.AbstractSet;
19	>	import java.util.AbstractCollection;
20	>	import java.util.Hashtable;
21	>	import java.util.HashMap;
22	>	import java.util.Iterator;
23	>	import java.util.Enumeration;
24	>	import java.util.ConcurrentModificationException;
25	>	import java.util.NoSuchElementException;
26	>	import java.util.concurrent.ConcurrentMap;
27	>	import java.util.concurrent.ThreadLocalRandom;
28	>	import java.util.concurrent.locks.LockSupport;
29	>	import java.util.concurrent.locks.AbstractQueuedSynchronizer;
30	>	import java.util.concurrent.atomic.AtomicReference;
31	>
32		import java.io.Serializable;
33
34		/**
35		* A hash table supporting full concurrency of retrievals and
36	<	* adjustable expected concurrency for updates. This class obeys the
36	>	* high expected concurrency for updates. This class obeys the
37		* same functional specification as {@link java.util.Hashtable}, and
38		* includes versions of methods corresponding to each method of
39	<	* <tt>Hashtable</tt>. However, even though all operations are
39	>	* {@code Hashtable}. However, even though all operations are
40		* thread-safe, retrieval operations do <em>not</em> entail locking,
41		* and there is <em>not</em> any support for locking the entire table
42		* in a way that prevents all access.  This class is fully
43	<	* interoperable with <tt>Hashtable</tt> in programs that rely on its
43	>	* interoperable with {@code Hashtable} in programs that rely on its
44		* thread safety but not on its synchronization details.
45		*
46	<	* <p> Retrieval operations (including <tt>get</tt>) generally do not
47	<	* block, so may overlap with update operations (including
48	<	* <tt>put</tt> and <tt>remove</tt>). Retrievals reflect the results
49	<	* of the most recently <em>completed</em> update operations holding
50	<	* upon their onset.  For aggregate operations such as <tt>putAll</tt>
51	<	* and <tt>clear</tt>, concurrent retrievals may reflect insertion or
52	<	* removal of only some entries.  Similarly, Iterators and
53	<	* Enumerations return elements reflecting the state of the hash table
54	<	* at some point at or since the creation of the iterator/enumeration.
55	<	* They do <em>not</em> throw {@link ConcurrentModificationException}.
56	<	* However, iterators are designed to be used by only one thread at a time.
57	<	*
58	<	* <p> The allowed concurrency among update operations is guided by
59	<	* the optional <tt>concurrencyLevel</tt> constructor argument
60	<	* (default <tt>16</tt>), which is used as a hint for internal sizing.  The
61	<	* table is internally partitioned to try to permit the indicated
62	<	* number of concurrent updates without contention. Because placement
63	<	* in hash tables is essentially random, the actual concurrency will
64	<	* vary.  Ideally, you should choose a value to accommodate as many
65	<	* threads as will ever concurrently modify the table. Using a
66	<	* significantly higher value than you need can waste space and time,
67	<	* and a significantly lower value can lead to thread contention. But
68	<	* overestimates and underestimates within an order of magnitude do
69	<	* not usually have much noticeable impact. A value of one is
70	<	* appropriate when it is known that only one thread will modify and
71	<	* all others will only read. Also, resizing this or any other kind of
72	<	* hash table is a relatively slow operation, so, when possible, it is
73	<	* a good idea to provide estimates of expected table sizes in
74	<	* constructors.
46	>	* <p>Retrieval operations (including {@code get}) generally do not
47	>	* block, so may overlap with update operations (including {@code put}
48	>	* and {@code remove}). Retrievals reflect the results of the most
49	>	* recently <em>completed</em> update operations holding upon their
50	>	* onset. (More formally, an update operation for a given key bears a
51	>	* <em>happens-before</em> relation with any (non-null) retrieval for
52	>	* that key reporting the updated value.)  For aggregate operations
53	>	* such as {@code putAll} and {@code clear}, concurrent retrievals may
54	>	* reflect insertion or removal of only some entries.  Similarly,
55	>	* Iterators and Enumerations return elements reflecting the state of
56	>	* the hash table at some point at or since the creation of the
57	>	* iterator/enumeration.  They do <em>not</em> throw {@link
58	>	* ConcurrentModificationException}.  However, iterators are designed
59	>	* to be used by only one thread at a time.  Bear in mind that the
60	>	* results of aggregate status methods including {@code size}, {@code
61	>	* isEmpty}, and {@code containsValue} are typically useful only when
62	>	* a map is not undergoing concurrent updates in other threads.
63	>	* Otherwise the results of these methods reflect transient states
64	>	* that may be adequate for monitoring or estimation purposes, but not
65	>	* for program control.
66	>	*
67	>	* <p>The table is dynamically expanded when there are too many
68	>	* collisions (i.e., keys that have distinct hash codes but fall into
69	>	* the same slot modulo the table size), with the expected average
70	>	* effect of maintaining roughly two bins per mapping (corresponding
71	>	* to a 0.75 load factor threshold for resizing). There may be much
72	>	* variance around this average as mappings are added and removed, but
73	>	* overall, this maintains a commonly accepted time/space tradeoff for
74	>	* hash tables.  However, resizing this or any other kind of hash
75	>	* table may be a relatively slow operation. When possible, it is a
76	>	* good idea to provide a size estimate as an optional {@code
77	>	* initialCapacity} constructor argument. An additional optional
78	>	* {@code loadFactor} constructor argument provides a further means of
79	>	* customizing initial table capacity by specifying the table density
80	>	* to be used in calculating the amount of space to allocate for the
81	>	* given number of elements.  Also, for compatibility with previous
82	>	* versions of this class, constructors may optionally specify an
83	>	* expected {@code concurrencyLevel} as an additional hint for
84	>	* internal sizing.  Note that using many keys with exactly the same
85	>	* {@code hashCode()} is a sure way to slow down performance of any
86	>	* hash table.
87	>	*
88	>	* <p>A {@link Set} projection of a ConcurrentHashMap may be created
89	>	* (using {@link #newKeySet()} or {@link #newKeySet(int)}), or viewed
90	>	* (using {@link #keySet(Object)} when only keys are of interest, and the
91	>	* mapped values are (perhaps transiently) not used or all take the
92	>	* same mapping value.
93	>	*
94	>	* <p>A ConcurrentHashMap can be used as scalable frequency map (a
95	>	* form of histogram or multiset) by using {@link LongAdder} values
96	>	* and initializing via {@link #computeIfAbsent}. For example, to add
97	>	* a count to a {@code ConcurrentHashMap<String,LongAdder> freqs}, you
98	>	* can use {@code freqs.computeIfAbsent(k -> new
99	>	* LongAdder()).increment();}
100		*
101		* <p>This class and its views and iterators implement all of the
102		* <em>optional</em> methods of the {@link Map} and {@link Iterator}
103		* interfaces.
104		*
105	<	* <p> Like {@link Hashtable} but unlike {@link HashMap}, this class
106	<	* does <em>not</em> allow <tt>null</tt> to be used as a key or value.
105	>	* <p>Like {@link Hashtable} but unlike {@link HashMap}, this class
106	>	* does <em>not</em> allow {@code null} to be used as a key or value.
107	>	*
108	>	* <p>ConcurrentHashMaps support parallel operations using the {@link
109	>	* ForkJoinPool#commonPool}. (Tasks that may be used in other contexts
110	>	* are available in class {@link ForkJoinTasks}). These operations are
111	>	* designed to be safely, and often sensibly, applied even with maps
112	>	* that are being concurrently updated by other threads; for example,
113	>	* when computing a snapshot summary of the values in a shared
114	>	* registry.  There are three kinds of operation, each with four
115	>	* forms, accepting functions with Keys, Values, Entries, and (Key,
116	>	* Value) arguments and/or return values. (The first three forms are
117	>	* also available via the {@link #keySet()}, {@link #values()} and
118	>	* {@link #entrySet()} views). Because the elements of a
119	>	* ConcurrentHashMap are not ordered in any particular way, and may be
120	>	* processed in different orders in different parallel executions, the
121	>	* correctness of supplied functions should not depend on any
122	>	* ordering, or on any other objects or values that may transiently
123	>	* change while computation is in progress; and except for forEach
124	>	* actions, should ideally be side-effect-free.
125	>	*
126	>	* <ul>
127	>	* <li> forEach: Perform a given action on each element.
128	>	* A variant form applies a given transformation on each element
129	>	* before performing the action.</li>
130	>	*
131	>	* <li> search: Return the first available non-null result of
132	>	* applying a given function on each element; skipping further
133	>	* search when a result is found.</li>
134	>	*
135	>	* <li> reduce: Accumulate each element.  The supplied reduction
136	>	* function cannot rely on ordering (more formally, it should be
137	>	* both associative and commutative).  There are five variants:
138	>	*
139	>	* <ul>
140	>	*
141	>	* <li> Plain reductions. (There is not a form of this method for
142	>	* (key, value) function arguments since there is no corresponding
143	>	* return type.)</li>
144	>	*
145	>	* <li> Mapped reductions that accumulate the results of a given
146	>	* function applied to each element.</li>
147	>	*
148	>	* <li> Reductions to scalar doubles, longs, and ints, using a
149	>	* given basis value.</li>
150	>	*
151	>	* </li>
152	>	* </ul>
153	>	* </ul>
154	>	*
155	>	* <p>The concurrency properties of bulk operations follow
156	>	* from those of ConcurrentHashMap: Any non-null result returned
157	>	* from {@code get(key)} and related access methods bears a
158	>	* happens-before relation with the associated insertion or
159	>	* update.  The result of any bulk operation reflects the
160	>	* composition of these per-element relations (but is not
161	>	* necessarily atomic with respect to the map as a whole unless it
162	>	* is somehow known to be quiescent).  Conversely, because keys
163	>	* and values in the map are never null, null serves as a reliable
164	>	* atomic indicator of the current lack of any result.  To
165	>	* maintain this property, null serves as an implicit basis for
166	>	* all non-scalar reduction operations. For the double, long, and
167	>	* int versions, the basis should be one that, when combined with
168	>	* any other value, returns that other value (more formally, it
169	>	* should be the identity element for the reduction). Most common
170	>	* reductions have these properties; for example, computing a sum
171	>	* with basis 0 or a minimum with basis MAX_VALUE.
172	>	*
173	>	* <p>Search and transformation functions provided as arguments
174	>	* should similarly return null to indicate the lack of any result
175	>	* (in which case it is not used). In the case of mapped
176	>	* reductions, this also enables transformations to serve as
177	>	* filters, returning null (or, in the case of primitive
178	>	* specializations, the identity basis) if the element should not
179	>	* be combined. You can create compound transformations and
180	>	* filterings by composing them yourself under this "null means
181	>	* there is nothing there now" rule before using them in search or
182	>	* reduce operations.
183	>	*
184	>	* <p>Methods accepting and/or returning Entry arguments maintain
185	>	* key-value associations. They may be useful for example when
186	>	* finding the key for the greatest value. Note that "plain" Entry
187	>	* arguments can be supplied using {@code new
188	>	* AbstractMap.SimpleEntry(k,v)}.
189	>	*
190	>	* <p>Bulk operations may complete abruptly, throwing an
191	>	* exception encountered in the application of a supplied
192	>	* function. Bear in mind when handling such exceptions that other
193	>	* concurrently executing functions could also have thrown
194	>	* exceptions, or would have done so if the first exception had
195	>	* not occurred.
196	>	*
197	>	* <p>Parallel speedups for bulk operations compared to sequential
198	>	* processing are common but not guaranteed.  Operations involving
199	>	* brief functions on small maps may execute more slowly than
200	>	* sequential loops if the underlying work to parallelize the
201	>	* computation is more expensive than the computation itself.
202	>	* Similarly, parallelization may not lead to much actual parallelism
203	>	* if all processors are busy performing unrelated tasks.
204	>	*
205	>	* <p>All arguments to all task methods must be non-null.
206	>	*
207	>	* <p><em>jsr166e note: During transition, this class
208	>	* uses nested functional interfaces with different names but the
209	>	* same forms as those expected for JDK8.</em>
210		*
211		* <p>This class is a member of the
212		* <a href="{@docRoot}/../technotes/guides/collections/index.html">
#	Line 67 | Line 217 | import java.io.Serializable;
217		* @param <K> the type of keys maintained by this map
218		* @param <V> the type of mapped values
219		*/
220	<	public class ConcurrentHashMap<K, V> extends AbstractMap<K, V>
221	<	implements ConcurrentMap<K, V>, Serializable {
220	>	public class ConcurrentHashMap<K, V>
221	>	implements ConcurrentMap<K, V>, Serializable {
222		private static final long serialVersionUID = 7249069246763182397L;
223
224	+	/**
225	+	* A partitionable iterator. A Spliterator can be traversed
226	+	* directly, but can also be partitioned (before traversal) by
227	+	* creating another Spliterator that covers a non-overlapping
228	+	* portion of the elements, and so may be amenable to parallel
229	+	* execution.
230	+	*
231	+	* <p>This interface exports a subset of expected JDK8
232	+	* functionality.
233	+	*
234	+	* <p>Sample usage: Here is one (of the several) ways to compute
235	+	* the sum of the values held in a map using the ForkJoin
236	+	* framework. As illustrated here, Spliterators are well suited to
237	+	* designs in which a task repeatedly splits off half its work
238	+	* into forked subtasks until small enough to process directly,
239	+	* and then joins these subtasks. Variants of this style can also
240	+	* be used in completion-based designs.
241	+	*
242	+	* <pre>
243	+	* {@code ConcurrentHashMap<String, Long> m = ...
244	+	* // split as if have 8 * parallelism, for load balance
245	+	* int n = m.size();
246	+	* int p = aForkJoinPool.getParallelism() * 8;
247	+	* int split = (n < p)? n : p;
248	+	* long sum = aForkJoinPool.invoke(new SumValues(m.valueSpliterator(), split, null));
249	+	* // ...
250	+	* static class SumValues extends RecursiveTask<Long> {
251	+	*   final Spliterator<Long> s;
252	+	*   final int split;             // split while > 1
253	+	*   final SumValues nextJoin;    // records forked subtasks to join
254	+	*   SumValues(Spliterator<Long> s, int depth, SumValues nextJoin) {
255	+	*     this.s = s; this.depth = depth; this.nextJoin = nextJoin;
256	+	*   }
257	+	*   public Long compute() {
258	+	*     long sum = 0;
259	+	*     SumValues subtasks = null; // fork subtasks
260	+	*     for (int s = split >>> 1; s > 0; s >>>= 1)
261	+	*       (subtasks = new SumValues(s.split(), s, subtasks)).fork();
262	+	*     while (s.hasNext())        // directly process remaining elements
263	+	*       sum += s.next();
264	+	*     for (SumValues t = subtasks; t != null; t = t.nextJoin)
265	+	*       sum += t.join();         // collect subtask results
266	+	*     return sum;
267	+	*   }
268	+	* }
269	+	* }</pre>
270	+	*/
271	+	public static interface Spliterator<T> extends Iterator<T> {
272	+	/**
273	+	* Returns a Spliterator covering approximately half of the
274	+	* elements, guaranteed not to overlap with those subsequently
275	+	* returned by this Spliterator.  After invoking this method,
276	+	* the current Spliterator will <em>not</em> produce any of
277	+	* the elements of the returned Spliterator, but the two
278	+	* Spliterators together will produce all of the elements that
279	+	* would have been produced by this Spliterator had this
280	+	* method not been called. The exact number of elements
281	+	* produced by the returned Spliterator is not guaranteed, and
282	+	* may be zero (i.e., with {@code hasNext()} reporting {@code
283	+	* false}) if this Spliterator cannot be further split.
284	+	*
285	+	* @return a Spliterator covering approximately half of the
286	+	* elements
287	+	* @throws IllegalStateException if this Spliterator has
288	+	* already commenced traversing elements
289	+	*/
290	+	Spliterator<T> split();
291	+	}
292	+
293	+
294		/*
295	<	* The basic strategy is to subdivide the table among Segments,
296	<	* each of which itself is a concurrently readable hash table.  To
297	<	* reduce footprint, all but one segments are constructed only
298	<	* when first needed (see ensureSegment). To maintain visibility
299	<	* in the presence of lazy construction, accesses to segments as
300	<	* well as elements of segment's table must use volatile access,
301	<	* which is done via Unsafe within methods segmentAt etc
302	<	* below. These provide the functionality of AtomicReferenceArrays
303	<	* but reduce the levels of indirection. Additionally,
304	<	* volatile-writes of table elements and entry "next" fields
305	<	* within locked operations use the cheaper "lazySet" forms of
306	<	* writes (via putOrderedObject) because these writes are always
307	<	* followed by lock releases that maintain sequential consistency
308	<	* of table updates.
309	<	*
310	<	* Historical note: The previous version of this class relied
311	<	* heavily on "final" fields, which avoided some volatile reads at
312	<	* the expense of a large initial footprint.  Some remnants of
313	<	* that design (including forced construction of segment 0) exist
314	<	* to ensure serialization compatibility.
295	>	* Overview:
296	>	*
297	>	* The primary design goal of this hash table is to maintain
298	>	* concurrent readability (typically method get(), but also
299	>	* iterators and related methods) while minimizing update
300	>	* contention. Secondary goals are to keep space consumption about
301	>	* the same or better than java.util.HashMap, and to support high
302	>	* initial insertion rates on an empty table by many threads.
303	>	*
304	>	* Each key-value mapping is held in a Node.  Because Node fields
305	>	* can contain special values, they are defined using plain Object
306	>	* types. Similarly in turn, all internal methods that use them
307	>	* work off Object types. And similarly, so do the internal
308	>	* methods of auxiliary iterator and view classes.  All public
309	>	* generic typed methods relay in/out of these internal methods,
310	>	* supplying null-checks and casts as needed. This also allows
311	>	* many of the public methods to be factored into a smaller number
312	>	* of internal methods (although sadly not so for the five
313	>	* variants of put-related operations). The validation-based
314	>	* approach explained below leads to a lot of code sprawl because
315	>	* retry-control precludes factoring into smaller methods.
316	>	*
317	>	* The table is lazily initialized to a power-of-two size upon the
318	>	* first insertion.  Each bin in the table normally contains a
319	>	* list of Nodes (most often, the list has only zero or one Node).
320	>	* Table accesses require volatile/atomic reads, writes, and
321	>	* CASes.  Because there is no other way to arrange this without
322	>	* adding further indirections, we use intrinsics
323	>	* (sun.misc.Unsafe) operations.  The lists of nodes within bins
324	>	* are always accurately traversable under volatile reads, so long
325	>	* as lookups check hash code and non-nullness of value before
326	>	* checking key equality.
327	>	*
328	>	* We use the top two bits of Node hash fields for control
329	>	* purposes -- they are available anyway because of addressing
330	>	* constraints.  As explained further below, these top bits are
331	>	* used as follows:
332	>	*  00 - Normal
333	>	*  01 - Locked
334	>	*  11 - Locked and may have a thread waiting for lock
335	>	*  10 - Node is a forwarding node
336	>	*
337	>	* The lower 30 bits of each Node's hash field contain a
338	>	* transformation of the key's hash code, except for forwarding
339	>	* nodes, for which the lower bits are zero (and so always have
340	>	* hash field == MOVED).
341	>	*
342	>	* Insertion (via put or its variants) of the first node in an
343	>	* empty bin is performed by just CASing it to the bin.  This is
344	>	* by far the most common case for put operations under most
345	>	* key/hash distributions.  Other update operations (insert,
346	>	* delete, and replace) require locks.  We do not want to waste
347	>	* the space required to associate a distinct lock object with
348	>	* each bin, so instead use the first node of a bin list itself as
349	>	* a lock. Blocking support for these locks relies on the builtin
350	>	* "synchronized" monitors.  However, we also need a tryLock
351	>	* construction, so we overlay these by using bits of the Node
352	>	* hash field for lock control (see above), and so normally use
353	>	* builtin monitors only for blocking and signalling using
354	>	* wait/notifyAll constructions. See Node.tryAwaitLock.
355	>	*
356	>	* Using the first node of a list as a lock does not by itself
357	>	* suffice though: When a node is locked, any update must first
358	>	* validate that it is still the first node after locking it, and
359	>	* retry if not. Because new nodes are always appended to lists,
360	>	* once a node is first in a bin, it remains first until deleted
361	>	* or the bin becomes invalidated (upon resizing).  However,
362	>	* operations that only conditionally update may inspect nodes
363	>	* until the point of update. This is a converse of sorts to the
364	>	* lazy locking technique described by Herlihy & Shavit.
365	>	*
366	>	* The main disadvantage of per-bin locks is that other update
367	>	* operations on other nodes in a bin list protected by the same
368	>	* lock can stall, for example when user equals() or mapping
369	>	* functions take a long time.  However, statistically, under
370	>	* random hash codes, this is not a common problem.  Ideally, the
371	>	* frequency of nodes in bins follows a Poisson distribution
372	>	* (http://en.wikipedia.org/wiki/Poisson_distribution) with a
373	>	* parameter of about 0.5 on average, given the resizing threshold
374	>	* of 0.75, although with a large variance because of resizing
375	>	* granularity. Ignoring variance, the expected occurrences of
376	>	* list size k are (exp(-0.5) * pow(0.5, k) / factorial(k)). The
377	>	* first values are:
378	>	*
379	>	* 0:    0.60653066
380	>	* 1:    0.30326533
381	>	* 2:    0.07581633
382	>	* 3:    0.01263606
383	>	* 4:    0.00157952
384	>	* 5:    0.00015795
385	>	* 6:    0.00001316
386	>	* 7:    0.00000094
387	>	* 8:    0.00000006
388	>	* more: less than 1 in ten million
389	>	*
390	>	* Lock contention probability for two threads accessing distinct
391	>	* elements is roughly 1 / (8 * #elements) under random hashes.
392	>	*
393	>	* Actual hash code distributions encountered in practice
394	>	* sometimes deviate significantly from uniform randomness.  This
395	>	* includes the case when N > (1<<30), so some keys MUST collide.
396	>	* Similarly for dumb or hostile usages in which multiple keys are
397	>	* designed to have identical hash codes. Also, although we guard
398	>	* against the worst effects of this (see method spread), sets of
399	>	* hashes may differ only in bits that do not impact their bin
400	>	* index for a given power-of-two mask.  So we use a secondary
401	>	* strategy that applies when the number of nodes in a bin exceeds
402	>	* a threshold, and at least one of the keys implements
403	>	* Comparable.  These TreeBins use a balanced tree to hold nodes
404	>	* (a specialized form of red-black trees), bounding search time
405	>	* to O(log N).  Each search step in a TreeBin is around twice as
406	>	* slow as in a regular list, but given that N cannot exceed
407	>	* (1<<64) (before running out of addresses) this bounds search
408	>	* steps, lock hold times, etc, to reasonable constants (roughly
409	>	* 100 nodes inspected per operation worst case) so long as keys
410	>	* are Comparable (which is very common -- String, Long, etc).
411	>	* TreeBin nodes (TreeNodes) also maintain the same "next"
412	>	* traversal pointers as regular nodes, so can be traversed in
413	>	* iterators in the same way.
414	>	*
415	>	* The table is resized when occupancy exceeds a percentage
416	>	* threshold (nominally, 0.75, but see below).  Only a single
417	>	* thread performs the resize (using field "sizeCtl", to arrange
418	>	* exclusion), but the table otherwise remains usable for reads
419	>	* and updates. Resizing proceeds by transferring bins, one by
420	>	* one, from the table to the next table.  Because we are using
421	>	* power-of-two expansion, the elements from each bin must either
422	>	* stay at same index, or move with a power of two offset. We
423	>	* eliminate unnecessary node creation by catching cases where old
424	>	* nodes can be reused because their next fields won't change.  On
425	>	* average, only about one-sixth of them need cloning when a table
426	>	* doubles. The nodes they replace will be garbage collectable as
427	>	* soon as they are no longer referenced by any reader thread that
428	>	* may be in the midst of concurrently traversing table.  Upon
429	>	* transfer, the old table bin contains only a special forwarding
430	>	* node (with hash field "MOVED") that contains the next table as
431	>	* its key. On encountering a forwarding node, access and update
432	>	* operations restart, using the new table.
433	>	*
434	>	* Each bin transfer requires its bin lock. However, unlike other
435	>	* cases, a transfer can skip a bin if it fails to acquire its
436	>	* lock, and revisit it later (unless it is a TreeBin). Method
437	>	* rebuild maintains a buffer of TRANSFER_BUFFER_SIZE bins that
438	>	* have been skipped because of failure to acquire a lock, and
439	>	* blocks only if none are available (i.e., only very rarely).
440	>	* The transfer operation must also ensure that all accessible
441	>	* bins in both the old and new table are usable by any traversal.
442	>	* When there are no lock acquisition failures, this is arranged
443	>	* simply by proceeding from the last bin (table.length - 1) up
444	>	* towards the first.  Upon seeing a forwarding node, traversals
445	>	* (see class Iter) arrange to move to the new table
446	>	* without revisiting nodes.  However, when any node is skipped
447	>	* during a transfer, all earlier table bins may have become
448	>	* visible, so are initialized with a reverse-forwarding node back
449	>	* to the old table until the new ones are established. (This
450	>	* sometimes requires transiently locking a forwarding node, which
451	>	* is possible under the above encoding.) These more expensive
452	>	* mechanics trigger only when necessary.
453	>	*
454	>	* The traversal scheme also applies to partial traversals of
455	>	* ranges of bins (via an alternate Traverser constructor)
456	>	* to support partitioned aggregate operations.  Also, read-only
457	>	* operations give up if ever forwarded to a null table, which
458	>	* provides support for shutdown-style clearing, which is also not
459	>	* currently implemented.
460	>	*
461	>	* Lazy table initialization minimizes footprint until first use,
462	>	* and also avoids resizings when the first operation is from a
463	>	* putAll, constructor with map argument, or deserialization.
464	>	* These cases attempt to override the initial capacity settings,
465	>	* but harmlessly fail to take effect in cases of races.
466	>	*
467	>	* The element count is maintained using a LongAdder, which avoids
468	>	* contention on updates but can encounter cache thrashing if read
469	>	* too frequently during concurrent access. To avoid reading so
470	>	* often, resizing is attempted either when a bin lock is
471	>	* contended, or upon adding to a bin already holding two or more
472	>	* nodes (checked before adding in the xIfAbsent methods, after
473	>	* adding in others). Under uniform hash distributions, the
474	>	* probability of this occurring at threshold is around 13%,
475	>	* meaning that only about 1 in 8 puts check threshold (and after
476	>	* resizing, many fewer do so). But this approximation has high
477	>	* variance for small table sizes, so we check on any collision
478	>	* for sizes <= 64. The bulk putAll operation further reduces
479	>	* contention by only committing count updates upon these size
480	>	* checks.
481	>	*
482	>	* Maintaining API and serialization compatibility with previous
483	>	* versions of this class introduces several oddities. Mainly: We
484	>	* leave untouched but unused constructor arguments refering to
485	>	* concurrencyLevel. We accept a loadFactor constructor argument,
486	>	* but apply it only to initial table capacity (which is the only
487	>	* time that we can guarantee to honor it.) We also declare an
488	>	* unused "Segment" class that is instantiated in minimal form
489	>	* only when serializing.
490		*/
491
492		/* ---------------- Constants -------------- */
493
494		/**
495	<	* The default initial capacity for this table,
496	<	* used when not otherwise specified in a constructor.
495	>	* The largest possible table capacity.  This value must be
496	>	* exactly 1<<30 to stay within Java array allocation and indexing
497	>	* bounds for power of two table sizes, and is further required
498	>	* because the top two bits of 32bit hash fields are used for
499	>	* control purposes.
500		*/
501	<	static final int DEFAULT_INITIAL_CAPACITY = 16;
501	>	private static final int MAXIMUM_CAPACITY = 1 << 30;
502
503		/**
504	<	* The default load factor for this table, used when not
505	<	* otherwise specified in a constructor.
504	>	* The default initial table capacity.  Must be a power of 2
505	>	* (i.e., at least 1) and at most MAXIMUM_CAPACITY.
506		*/
507	<	static final float DEFAULT_LOAD_FACTOR = 0.75f;
507	>	private static final int DEFAULT_CAPACITY = 16;
508
509		/**
510	<	* The default concurrency level for this table, used when not
511	<	* otherwise specified in a constructor.
510	>	* The largest possible (non-power of two) array size.
511	>	* Needed by toArray and related methods.
512		*/
513	<	static final int DEFAULT_CONCURRENCY_LEVEL = 16;
513	>	static final int MAX_ARRAY_SIZE = Integer.MAX_VALUE - 8;
514
515		/**
516	<	* The maximum capacity, used if a higher value is implicitly
517	<	* specified by either of the constructors with arguments.  MUST
120	<	* be a power of two <= 1<<30 to ensure that entries are indexable
121	<	* using ints.
516	>	* The default concurrency level for this table. Unused but
517	>	* defined for compatibility with previous versions of this class.
518		*/
519	<	static final int MAXIMUM_CAPACITY = 1 << 30;
519	>	private static final int DEFAULT_CONCURRENCY_LEVEL = 16;
520
521		/**
522	<	* The minimum capacity for per-segment tables.  Must be a power
523	<	* of two, at least two to avoid immediate resizing on next use
524	<	* after lazy construction.
522	>	* The load factor for this table. Overrides of this value in
523	>	* constructors affect only the initial table capacity.  The
524	>	* actual floating point value isn't normally used -- it is
525	>	* simpler to use expressions such as {@code n - (n >>> 2)} for
526	>	* the associated resizing threshold.
527		*/
528	<	static final int MIN_SEGMENT_TABLE_CAPACITY = 2;
528	>	private static final float LOAD_FACTOR = 0.75f;
529
530		/**
531	<	* The maximum number of segments to allow; used to bound
532	<	* constructor arguments. Must be power of two less than 1 << 24.
531	>	* The buffer size for skipped bins during transfers. The
532	>	* value is arbitrary but should be large enough to avoid
533	>	* most locking stalls during resizes.
534		*/
535	<	static final int MAX_SEGMENTS = 1 << 16; // slightly conservative
535	>	private static final int TRANSFER_BUFFER_SIZE = 32;
536
537		/**
538	<	* Number of unsynchronized retries in size and containsValue
539	<	* methods before resorting to locking. This is used to avoid
540	<	* unbounded retries if tables undergo continuous modification
142	<	* which would make it impossible to obtain an accurate result.
538	>	* The bin count threshold for using a tree rather than list for a
539	>	* bin.  The value reflects the approximate break-even point for
540	>	* using tree-based operations.
541		*/
542	<	static final int RETRIES_BEFORE_LOCK = 2;
542	>	private static final int TREE_THRESHOLD = 8;
543	>
544	>	/*
545	>	* Encodings for special uses of Node hash fields. See above for
546	>	* explanation.
547	>	*/
548	>	static final int MOVED     = 0x80000000; // hash field for forwarding nodes
549	>	static final int LOCKED    = 0x40000000; // set/tested only as a bit
550	>	static final int WAITING   = 0xc0000000; // both bits set/tested together
551	>	static final int HASH_BITS = 0x3fffffff; // usable bits of normal node hash
552
553		/* ---------------- Fields -------------- */
554
555		/**
556	<	* Mask value for indexing into segments. The upper bits of a
557	<	* key's hash code are used to choose the segment.
556	>	* The array of bins. Lazily initialized upon first insertion.
557	>	* Size is always a power of two. Accessed directly by iterators.
558		*/
559	<	final int segmentMask;
559	>	transient volatile Node[] table;
560
561		/**
562	<	* Shift value for indexing within segments.
562	>	* The counter maintaining number of elements.
563		*/
564	<	final int segmentShift;
564	>	private transient final LongAdder counter;
565
566		/**
567	<	* The segments, each of which is a specialized hash table.
567	>	* Table initialization and resizing control.  When negative, the
568	>	* table is being initialized or resized. Otherwise, when table is
569	>	* null, holds the initial table size to use upon creation, or 0
570	>	* for default. After initialization, holds the next element count
571	>	* value upon which to resize the table.
572		*/
573	<	final Segment<K,V>[] segments;
573	>	private transient volatile int sizeCtl;
574	>
575	>	// views
576	>	private transient KeySetView<K,V> keySet;
577	>	private transient ValuesView<K,V> values;
578	>	private transient EntrySetView<K,V> entrySet;
579	>
580	>	/** For serialization compatibility. Null unless serialized; see below */
581	>	private Segment<K,V>[] segments;
582	>
583	>	/* ---------------- Table element access -------------- */
584	>
585	>	/*
586	>	* Volatile access methods are used for table elements as well as
587	>	* elements of in-progress next table while resizing.  Uses are
588	>	* null checked by callers, and implicitly bounds-checked, relying
589	>	* on the invariants that tab arrays have non-zero size, and all
590	>	* indices are masked with (tab.length - 1) which is never
591	>	* negative and always less than length. Note that, to be correct
592	>	* wrt arbitrary concurrency errors by users, bounds checks must
593	>	* operate on local variables, which accounts for some odd-looking
594	>	* inline assignments below.
595	>	*/
596	>
597	>	static final Node tabAt(Node[] tab, int i) { // used by Iter
598	>	return (Node)UNSAFE.getObjectVolatile(tab, ((long)i<<ASHIFT)+ABASE);
599	>	}
600	>
601	>	private static final boolean casTabAt(Node[] tab, int i, Node c, Node v) {
602	>	return UNSAFE.compareAndSwapObject(tab, ((long)i<<ASHIFT)+ABASE, c, v);
603	>	}
604	>
605	>	private static final void setTabAt(Node[] tab, int i, Node v) {
606	>	UNSAFE.putObjectVolatile(tab, ((long)i<<ASHIFT)+ABASE, v);
607	>	}
608
609	<	transient Set<K> keySet;
165	<	transient Set<Map.Entry<K,V>> entrySet;
166	<	transient Collection<V> values;
609	>	/* ---------------- Nodes -------------- */
610
611		/**
612	<	* ConcurrentHashMap list entry. Note that this is never exported
613	<	* out as a user-visible Map.Entry.
612	>	* Key-value entry. Note that this is never exported out as a
613	>	* user-visible Map.Entry (see MapEntry below). Nodes with a hash
614	>	* field of MOVED are special, and do not contain user keys or
615	>	* values.  Otherwise, keys are never null, and null val fields
616	>	* indicate that a node is in the process of being deleted or
617	>	* created. For purposes of read-only access, a key may be read
618	>	* before a val, but can only be used after checking val to be
619	>	* non-null.
620		*/
621	<	static final class HashEntry<K,V> {
622	<	final int hash;
623	<	final K key;
624	<	volatile V value;
625	<	volatile HashEntry<K,V> next;
621	>	static class Node {
622	>	volatile int hash;
623	>	final Object key;
624	>	volatile Object val;
625	>	volatile Node next;
626
627	<	HashEntry(int hash, K key, V value, HashEntry<K,V> next) {
627	>	Node(int hash, Object key, Object val, Node next) {
628		this.hash = hash;
629		this.key = key;
630	<	this.value = value;
630	>	this.val = val;
631		this.next = next;
632		}
633
634	+	/** CompareAndSet the hash field */
635	+	final boolean casHash(int cmp, int val) {
636	+	return UNSAFE.compareAndSwapInt(this, hashOffset, cmp, val);
637	+	}
638	+
639	+	/** The number of spins before blocking for a lock */
640	+	static final int MAX_SPINS =
641	+	Runtime.getRuntime().availableProcessors() > 1 ? 64 : 1;
642	+
643		/**
644	<	* Sets next field with volatile write semantics.  (See above
645	<	* about use of putOrderedObject.)
644	>	* Spins a while if LOCKED bit set and this node is the first
645	>	* of its bin, and then sets WAITING bits on hash field and
646	>	* blocks (once) if they are still set.  It is OK for this
647	>	* method to return even if lock is not available upon exit,
648	>	* which enables these simple single-wait mechanics.
649	>	*
650	>	* The corresponding signalling operation is performed within
651	>	* callers: Upon detecting that WAITING has been set when
652	>	* unlocking lock (via a failed CAS from non-waiting LOCKED
653	>	* state), unlockers acquire the sync lock and perform a
654	>	* notifyAll.
655	>	*
656	>	* The initial sanity check on tab and bounds is not currently
657	>	* necessary in the only usages of this method, but enables
658	>	* use in other future contexts.
659		*/
660	<	final void setNext(HashEntry<K,V> n) {
661	<	UNSAFE.putOrderedObject(this, nextOffset, n);
660	>	final void tryAwaitLock(Node[] tab, int i) {
661	>	if (tab != null && i >= 0 && i < tab.length) { // sanity check
662	>	int r = ThreadLocalRandom.current().nextInt(); // randomize spins
663	>	int spins = MAX_SPINS, h;
664	>	while (tabAt(tab, i) == this && ((h = hash) & LOCKED) != 0) {
665	>	if (spins >= 0) {
666	>	r ^= r << 1; r ^= r >>> 3; r ^= r << 10; // xorshift
667	>	if (r >= 0 && --spins == 0)
668	>	Thread.yield();  // yield before block
669	>	}
670	>	else if (casHash(h, h | WAITING)) {
671	>	synchronized (this) {
672	>	if (tabAt(tab, i) == this &&
673	>	(hash & WAITING) == WAITING) {
674	>	try {
675	>	wait();
676	>	} catch (InterruptedException ie) {
677	>	try {
678	>	Thread.currentThread().interrupt();
679	>	} catch (SecurityException ignore) {
680	>	}
681	>	}
682	>	}
683	>	else
684	>	notifyAll(); // possibly won race vs signaller
685	>	}
686	>	break;
687	>	}
688	>	}
689	>	}
690		}
691
692	<	// Unsafe mechanics
693	<	static final sun.misc.Unsafe UNSAFE;
694	<	static final long nextOffset;
692	>	// Unsafe mechanics for casHash
693	>	private static final sun.misc.Unsafe UNSAFE;
694	>	private static final long hashOffset;
695	>
696		static {
697		try {
698		UNSAFE = sun.misc.Unsafe.getUnsafe();
699	<	Class<?> k = HashEntry.class;
700	<	nextOffset = UNSAFE.objectFieldOffset
701	<	(k.getDeclaredField("next"));
699	>	Class<?> k = Node.class;
700	>	hashOffset = UNSAFE.objectFieldOffset
701	>	(k.getDeclaredField("hash"));
702		} catch (Exception e) {
703		throw new Error(e);
704		}
705		}
706		}
707
708	+	/* ---------------- TreeBins -------------- */
709	+
710		/**
711	<	* Gets the ith element of given table (if nonnull) with volatile
712	<	* read semantics. Note: This is manually integrated into a few
713	<	* performance-sensitive methods to reduce call overhead.
714	<	*/
715	<	@SuppressWarnings("unchecked")
716	<	static final <K,V> HashEntry<K,V> entryAt(HashEntry<K,V>[] tab, int i) {
717	<	return (tab == null) ? null :
718	<	(HashEntry<K,V>) UNSAFE.getObjectVolatile
719	<	(tab, ((long)i << TSHIFT) + TBASE);
720	<	}
721	<
722	<	/**
723	<	* Sets the ith element of given table, with volatile write
724	<	* semantics. (See above about use of putOrderedObject.)
223	<	*/
224	<	static final <K,V> void setEntryAt(HashEntry<K,V>[] tab, int i,
225	<	HashEntry<K,V> e) {
226	<	UNSAFE.putOrderedObject(tab, ((long)i << TSHIFT) + TBASE, e);
227	<	}
228	<
229	<	/**
230	<	* Applies a supplemental hash function to a given hashCode, which
231	<	* defends against poor quality hash functions.  This is critical
232	<	* because ConcurrentHashMap uses power-of-two length hash tables,
233	<	* that otherwise encounter collisions for hashCodes that do not
234	<	* differ in lower or upper bits.
235	<	*/
236	<	private static int hash(int h) {
237	<	// Spread bits to regularize both segment and index locations,
238	<	// using variant of single-word Wang/Jenkins hash.
239	<	h += (h <<  15) ^ 0xffffcd7d;
240	<	h ^= (h >>> 10);
241	<	h += (h <<   3);
242	<	h ^= (h >>>  6);
243	<	h += (h <<   2) + (h << 14);
244	<	return h ^ (h >>> 16);
245	<	}
246	<
247	<	/**
248	<	* Segments are specialized versions of hash tables.  This
249	<	* subclasses from ReentrantLock opportunistically, just to
250	<	* simplify some locking and avoid separate construction.
251	<	*/
252	<	static final class Segment<K,V> extends ReentrantLock implements Serializable {
253	<	/*
254	<	* Segments maintain a table of entry lists that are always
255	<	* kept in a consistent state, so can be read (via volatile
256	<	* reads of segments and tables) without locking.  This
257	<	* requires replicating nodes when necessary during table
258	<	* resizing, so the old lists can be traversed by readers
259	<	* still using old version of table.
260	<	*
261	<	* This class defines only mutative methods requiring locking.
262	<	* Except as noted, the methods of this class perform the
263	<	* per-segment versions of ConcurrentHashMap methods.  (Other
264	<	* methods are integrated directly into ConcurrentHashMap
265	<	* methods.) These mutative methods use a form of controlled
266	<	* spinning on contention via methods scanAndLock and
267	<	* scanAndLockForPut. These intersperse tryLocks with
268	<	* traversals to locate nodes.  The main benefit is to absorb
269	<	* cache misses (which are very common for hash tables) while
270	<	* obtaining locks so that traversal is faster once
271	<	* acquired. We do not actually use the found nodes since they
272	<	* must be re-acquired under lock anyway to ensure sequential
273	<	* consistency of updates (and in any case may be undetectably
274	<	* stale), but they will normally be much faster to re-locate.
275	<	* Also, scanAndLockForPut speculatively creates a fresh node
276	<	* to use in put if no node is found.
277	<	*/
711	>	* Nodes for use in TreeBins
712	>	*/
713	>	static final class TreeNode extends Node {
714	>	TreeNode parent;  // red-black tree links
715	>	TreeNode left;
716	>	TreeNode right;
717	>	TreeNode prev;    // needed to unlink next upon deletion
718	>	boolean red;
719	>
720	>	TreeNode(int hash, Object key, Object val, Node next, TreeNode parent) {
721	>	super(hash, key, val, next);
722	>	this.parent = parent;
723	>	}
724	>	}
725
726	+	/**
727	+	* A specialized form of red-black tree for use in bins
728	+	* whose size exceeds a threshold.
729	+	*
730	+	* TreeBins use a special form of comparison for search and
731	+	* related operations (which is the main reason we cannot use
732	+	* existing collections such as TreeMaps). TreeBins contain
733	+	* Comparable elements, but may contain others, as well as
734	+	* elements that are Comparable but not necessarily Comparable<T>
735	+	* for the same T, so we cannot invoke compareTo among them. To
736	+	* handle this, the tree is ordered primarily by hash value, then
737	+	* by getClass().getName() order, and then by Comparator order
738	+	* among elements of the same class.  On lookup at a node, if
739	+	* elements are not comparable or compare as 0, both left and
740	+	* right children may need to be searched in the case of tied hash
741	+	* values. (This corresponds to the full list search that would be
742	+	* necessary if all elements were non-Comparable and had tied
743	+	* hashes.)  The red-black balancing code is updated from
744	+	* pre-jdk-collections
745	+	* (http://gee.cs.oswego.edu/dl/classes/collections/RBCell.java)
746	+	* based in turn on Cormen, Leiserson, and Rivest "Introduction to
747	+	* Algorithms" (CLR).
748	+	*
749	+	* TreeBins also maintain a separate locking discipline than
750	+	* regular bins. Because they are forwarded via special MOVED
751	+	* nodes at bin heads (which can never change once established),
752	+	* we cannot use those nodes as locks. Instead, TreeBin
753	+	* extends AbstractQueuedSynchronizer to support a simple form of
754	+	* read-write lock. For update operations and table validation,
755	+	* the exclusive form of lock behaves in the same way as bin-head
756	+	* locks. However, lookups use shared read-lock mechanics to allow
757	+	* multiple readers in the absence of writers.  Additionally,
758	+	* these lookups do not ever block: While the lock is not
759	+	* available, they proceed along the slow traversal path (via
760	+	* next-pointers) until the lock becomes available or the list is
761	+	* exhausted, whichever comes first. (These cases are not fast,
762	+	* but maximize aggregate expected throughput.)  The AQS mechanics
763	+	* for doing this are straightforward.  The lock state is held as
764	+	* AQS getState().  Read counts are negative; the write count (1)
765	+	* is positive.  There are no signalling preferences among readers
766	+	* and writers. Since we don't need to export full Lock API, we
767	+	* just override the minimal AQS methods and use them directly.
768	+	*/
769	+	static final class TreeBin extends AbstractQueuedSynchronizer {
770		private static final long serialVersionUID = 2249069246763182397L;
771	+	transient TreeNode root;  // root of tree
772	+	transient TreeNode first; // head of next-pointer list
773
774	<	/**
775	<	* The maximum number of times to tryLock in a prescan before
776	<	* possibly blocking on acquire in preparation for a locked
777	<	* segment operation. On multiprocessors, using a bounded
778	<	* number of retries maintains cache acquired while locating
779	<	* nodes.
780	<	*/
781	<	static final int MAX_SCAN_RETRIES =
782	<	Runtime.getRuntime().availableProcessors() > 1 ? 64 : 1;
774	>	/* AQS overrides */
775	>	public final boolean isHeldExclusively() { return getState() > 0; }
776	>	public final boolean tryAcquire(int ignore) {
777	>	if (compareAndSetState(0, 1)) {
778	>	setExclusiveOwnerThread(Thread.currentThread());
779	>	return true;
780	>	}
781	>	return false;
782	>	}
783	>	public final boolean tryRelease(int ignore) {
784	>	setExclusiveOwnerThread(null);
785	>	setState(0);
786	>	return true;
787	>	}
788	>	public final int tryAcquireShared(int ignore) {
789	>	for (int c;;) {
790	>	if ((c = getState()) > 0)
791	>	return -1;
792	>	if (compareAndSetState(c, c -1))
793	>	return 1;
794	>	}
795	>	}
796	>	public final boolean tryReleaseShared(int ignore) {
797	>	int c;
798	>	do {} while (!compareAndSetState(c = getState(), c + 1));
799	>	return c == -1;
800	>	}
801	>
802	>	/** From CLR */
803	>	private void rotateLeft(TreeNode p) {
804	>	if (p != null) {
805	>	TreeNode r = p.right, pp, rl;
806	>	if ((rl = p.right = r.left) != null)
807	>	rl.parent = p;
808	>	if ((pp = r.parent = p.parent) == null)
809	>	root = r;
810	>	else if (pp.left == p)
811	>	pp.left = r;
812	>	else
813	>	pp.right = r;
814	>	r.left = p;
815	>	p.parent = r;
816	>	}
817	>	}
818
819	<	/**
820	<	* The per-segment table. Elements are accessed via
821	<	* entryAt/setEntryAt providing volatile semantics.
822	<	*/
823	<	transient volatile HashEntry<K,V>[] table;
819	>	/** From CLR */
820	>	private void rotateRight(TreeNode p) {
821	>	if (p != null) {
822	>	TreeNode l = p.left, pp, lr;
823	>	if ((lr = p.left = l.right) != null)
824	>	lr.parent = p;
825	>	if ((pp = l.parent = p.parent) == null)
826	>	root = l;
827	>	else if (pp.right == p)
828	>	pp.right = l;
829	>	else
830	>	pp.left = l;
831	>	l.right = p;
832	>	p.parent = l;
833	>	}
834	>	}
835
836		/**
837	<	* The number of elements. Accessed only either within locks
838	<	* or among other volatile reads that maintain visibility.
837	>	* Returns the TreeNode (or null if not found) for the given key
838	>	* starting at given root.
839		*/
840	<	transient int count;
840	>	@SuppressWarnings("unchecked") final TreeNode getTreeNode
841	>	(int h, Object k, TreeNode p) {
842	>	Class<?> c = k.getClass();
843	>	while (p != null) {
844	>	int dir, ph;  Object pk; Class<?> pc;
845	>	if ((ph = p.hash) == h) {
846	>	if ((pk = p.key) == k || k.equals(pk))
847	>	return p;
848	>	if (c != (pc = pk.getClass()) ||
849	>	!(k instanceof Comparable) ||
850	>	(dir = ((Comparable)k).compareTo((Comparable)pk)) == 0) {
851	>	if ((dir = (c == pc) ? 0 :
852	>	c.getName().compareTo(pc.getName())) == 0) {
853	>	TreeNode r = null, pl, pr; // check both sides
854	>	if ((pr = p.right) != null && h >= pr.hash &&
855	>	(r = getTreeNode(h, k, pr)) != null)
856	>	return r;
857	>	else if ((pl = p.left) != null && h <= pl.hash)
858	>	dir = -1;
859	>	else // nothing there
860	>	return null;
861	>	}
862	>	}
863	>	}
864	>	else
865	>	dir = (h < ph) ? -1 : 1;
866	>	p = (dir > 0) ? p.right : p.left;
867	>	}
868	>	return null;
869	>	}
870
871		/**
872	<	* The total number of mutative operations in this segment.
873	<	* Even though this may overflows 32 bits, it provides
874	<	* sufficient accuracy for stability checks in CHM isEmpty()
307	<	* and size() methods.  Accessed only either within locks or
308	<	* among other volatile reads that maintain visibility.
872	>	* Wrapper for getTreeNode used by CHM.get. Tries to obtain
873	>	* read-lock to call getTreeNode, but during failure to get
874	>	* lock, searches along next links.
875		*/
876	<	transient int modCount;
876	>	final Object getValue(int h, Object k) {
877	>	Node r = null;
878	>	int c = getState(); // Must read lock state first
879	>	for (Node e = first; e != null; e = e.next) {
880	>	if (c <= 0 && compareAndSetState(c, c - 1)) {
881	>	try {
882	>	r = getTreeNode(h, k, root);
883	>	} finally {
884	>	releaseShared(0);
885	>	}
886	>	break;
887	>	}
888	>	else if ((e.hash & HASH_BITS) == h && k.equals(e.key)) {
889	>	r = e;
890	>	break;
891	>	}
892	>	else
893	>	c = getState();
894	>	}
895	>	return r == null ? null : r.val;
896	>	}
897
898		/**
899	<	* The table is rehashed when its size exceeds this threshold.
900	<	* (The value of this field is always <tt>(int)(capacity *
315	<	* loadFactor)</tt>.)
899	>	* Finds or adds a node.
900	>	* @return null if added
901		*/
902	<	transient int threshold;
902	>	@SuppressWarnings("unchecked") final TreeNode putTreeNode
903	>	(int h, Object k, Object v) {
904	>	Class<?> c = k.getClass();
905	>	TreeNode pp = root, p = null;
906	>	int dir = 0;
907	>	while (pp != null) { // find existing node or leaf to insert at
908	>	int ph;  Object pk; Class<?> pc;
909	>	p = pp;
910	>	if ((ph = p.hash) == h) {
911	>	if ((pk = p.key) == k || k.equals(pk))
912	>	return p;
913	>	if (c != (pc = pk.getClass()) ||
914	>	!(k instanceof Comparable) ||
915	>	(dir = ((Comparable)k).compareTo((Comparable)pk)) == 0) {
916	>	TreeNode s = null, r = null, pr;
917	>	if ((dir = (c == pc) ? 0 :
918	>	c.getName().compareTo(pc.getName())) == 0) {
919	>	if ((pr = p.right) != null && h >= pr.hash &&
920	>	(r = getTreeNode(h, k, pr)) != null)
921	>	return r;
922	>	else // continue left
923	>	dir = -1;
924	>	}
925	>	else if ((pr = p.right) != null && h >= pr.hash)
926	>	s = pr;
927	>	if (s != null && (r = getTreeNode(h, k, s)) != null)
928	>	return r;
929	>	}
930	>	}
931	>	else
932	>	dir = (h < ph) ? -1 : 1;
933	>	pp = (dir > 0) ? p.right : p.left;
934	>	}
935	>
936	>	TreeNode f = first;
937	>	TreeNode x = first = new TreeNode(h, k, v, f, p);
938	>	if (p == null)
939	>	root = x;
940	>	else { // attach and rebalance; adapted from CLR
941	>	TreeNode xp, xpp;
942	>	if (f != null)
943	>	f.prev = x;
944	>	if (dir <= 0)
945	>	p.left = x;
946	>	else
947	>	p.right = x;
948	>	x.red = true;
949	>	while (x != null && (xp = x.parent) != null && xp.red &&
950	>	(xpp = xp.parent) != null) {
951	>	TreeNode xppl = xpp.left;
952	>	if (xp == xppl) {
953	>	TreeNode y = xpp.right;
954	>	if (y != null && y.red) {
955	>	y.red = false;
956	>	xp.red = false;
957	>	xpp.red = true;
958	>	x = xpp;
959	>	}
960	>	else {
961	>	if (x == xp.right) {
962	>	rotateLeft(x = xp);
963	>	xpp = (xp = x.parent) == null ? null : xp.parent;
964	>	}
965	>	if (xp != null) {
966	>	xp.red = false;
967	>	if (xpp != null) {
968	>	xpp.red = true;
969	>	rotateRight(xpp);
970	>	}
971	>	}
972	>	}
973	>	}
974	>	else {
975	>	TreeNode y = xppl;
976	>	if (y != null && y.red) {
977	>	y.red = false;
978	>	xp.red = false;
979	>	xpp.red = true;
980	>	x = xpp;
981	>	}
982	>	else {
983	>	if (x == xp.left) {
984	>	rotateRight(x = xp);
985	>	xpp = (xp = x.parent) == null ? null : xp.parent;
986	>	}
987	>	if (xp != null) {
988	>	xp.red = false;
989	>	if (xpp != null) {
990	>	xpp.red = true;
991	>	rotateLeft(xpp);
992	>	}
993	>	}
994	>	}
995	>	}
996	>	}
997	>	TreeNode r = root;
998	>	if (r != null && r.red)
999	>	r.red = false;
1000	>	}
1001	>	return null;
1002	>	}
1003
1004		/**
1005	<	* The load factor for the hash table.  Even though this value
1006	<	* is same for all segments, it is replicated to avoid needing
1007	<	* links to outer object.
1008	<	* @serial
1005	>	* Removes the given node, that must be present before this
1006	>	* call.  This is messier than typical red-black deletion code
1007	>	* because we cannot swap the contents of an interior node
1008	>	* with a leaf successor that is pinned by "next" pointers
1009	>	* that are accessible independently of lock. So instead we
1010	>	* swap the tree linkages.
1011		*/
1012	<	final float loadFactor;
1012	>	final void deleteTreeNode(TreeNode p) {
1013	>	TreeNode next = (TreeNode)p.next; // unlink traversal pointers
1014	>	TreeNode pred = p.prev;
1015	>	if (pred == null)
1016	>	first = next;
1017	>	else
1018	>	pred.next = next;
1019	>	if (next != null)
1020	>	next.prev = pred;
1021	>	TreeNode replacement;
1022	>	TreeNode pl = p.left;
1023	>	TreeNode pr = p.right;
1024	>	if (pl != null && pr != null) {
1025	>	TreeNode s = pr, sl;
1026	>	while ((sl = s.left) != null) // find successor
1027	>	s = sl;
1028	>	boolean c = s.red; s.red = p.red; p.red = c; // swap colors
1029	>	TreeNode sr = s.right;
1030	>	TreeNode pp = p.parent;
1031	>	if (s == pr) { // p was s's direct parent
1032	>	p.parent = s;
1033	>	s.right = p;
1034	>	}
1035	>	else {
1036	>	TreeNode sp = s.parent;
1037	>	if ((p.parent = sp) != null) {
1038	>	if (s == sp.left)
1039	>	sp.left = p;
1040	>	else
1041	>	sp.right = p;
1042	>	}
1043	>	if ((s.right = pr) != null)
1044	>	pr.parent = s;
1045	>	}
1046	>	p.left = null;
1047	>	if ((p.right = sr) != null)
1048	>	sr.parent = p;
1049	>	if ((s.left = pl) != null)
1050	>	pl.parent = s;
1051	>	if ((s.parent = pp) == null)
1052	>	root = s;
1053	>	else if (p == pp.left)
1054	>	pp.left = s;
1055	>	else
1056	>	pp.right = s;
1057	>	replacement = sr;
1058	>	}
1059	>	else
1060	>	replacement = (pl != null) ? pl : pr;
1061	>	TreeNode pp = p.parent;
1062	>	if (replacement == null) {
1063	>	if (pp == null) {
1064	>	root = null;
1065	>	return;
1066	>	}
1067	>	replacement = p;
1068	>	}
1069	>	else {
1070	>	replacement.parent = pp;
1071	>	if (pp == null)
1072	>	root = replacement;
1073	>	else if (p == pp.left)
1074	>	pp.left = replacement;
1075	>	else
1076	>	pp.right = replacement;
1077	>	p.left = p.right = p.parent = null;
1078	>	}
1079	>	if (!p.red) { // rebalance, from CLR
1080	>	TreeNode x = replacement;
1081	>	while (x != null) {
1082	>	TreeNode xp, xpl;
1083	>	if (x.red || (xp = x.parent) == null) {
1084	>	x.red = false;
1085	>	break;
1086	>	}
1087	>	if (x == (xpl = xp.left)) {
1088	>	TreeNode sib = xp.right;
1089	>	if (sib != null && sib.red) {
1090	>	sib.red = false;
1091	>	xp.red = true;
1092	>	rotateLeft(xp);
1093	>	sib = (xp = x.parent) == null ? null : xp.right;
1094	>	}
1095	>	if (sib == null)
1096	>	x = xp;
1097	>	else {
1098	>	TreeNode sl = sib.left, sr = sib.right;
1099	>	if ((sr == null || !sr.red) &&
1100	>	(sl == null || !sl.red)) {
1101	>	sib.red = true;
1102	>	x = xp;
1103	>	}
1104	>	else {
1105	>	if (sr == null || !sr.red) {
1106	>	if (sl != null)
1107	>	sl.red = false;
1108	>	sib.red = true;
1109	>	rotateRight(sib);
1110	>	sib = (xp = x.parent) == null ? null : xp.right;
1111	>	}
1112	>	if (sib != null) {
1113	>	sib.red = (xp == null) ? false : xp.red;
1114	>	if ((sr = sib.right) != null)
1115	>	sr.red = false;
1116	>	}
1117	>	if (xp != null) {
1118	>	xp.red = false;
1119	>	rotateLeft(xp);
1120	>	}
1121	>	x = root;
1122	>	}
1123	>	}
1124	>	}
1125	>	else { // symmetric
1126	>	TreeNode sib = xpl;
1127	>	if (sib != null && sib.red) {
1128	>	sib.red = false;
1129	>	xp.red = true;
1130	>	rotateRight(xp);
1131	>	sib = (xp = x.parent) == null ? null : xp.left;
1132	>	}
1133	>	if (sib == null)
1134	>	x = xp;
1135	>	else {
1136	>	TreeNode sl = sib.left, sr = sib.right;
1137	>	if ((sl == null || !sl.red) &&
1138	>	(sr == null || !sr.red)) {
1139	>	sib.red = true;
1140	>	x = xp;
1141	>	}
1142	>	else {
1143	>	if (sl == null || !sl.red) {
1144	>	if (sr != null)
1145	>	sr.red = false;
1146	>	sib.red = true;
1147	>	rotateLeft(sib);
1148	>	sib = (xp = x.parent) == null ? null : xp.left;
1149	>	}
1150	>	if (sib != null) {
1151	>	sib.red = (xp == null) ? false : xp.red;
1152	>	if ((sl = sib.left) != null)
1153	>	sl.red = false;
1154	>	}
1155	>	if (xp != null) {
1156	>	xp.red = false;
1157	>	rotateRight(xp);
1158	>	}
1159	>	x = root;
1160	>	}
1161	>	}
1162	>	}
1163	>	}
1164	>	}
1165	>	if (p == replacement && (pp = p.parent) != null) {
1166	>	if (p == pp.left) // detach pointers
1167	>	pp.left = null;
1168	>	else if (p == pp.right)
1169	>	pp.right = null;
1170	>	p.parent = null;
1171	>	}
1172	>	}
1173	>	}
1174	>
1175	>	/* ---------------- Collision reduction methods -------------- */
1176
1177	<	Segment(float lf, int threshold, HashEntry<K,V>[] tab) {
1178	<	this.loadFactor = lf;
1179	<	this.threshold = threshold;
1180	<	this.table = tab;
1177	>	/**
1178	>	* Spreads higher bits to lower, and also forces top 2 bits to 0.
1179	>	* Because the table uses power-of-two masking, sets of hashes
1180	>	* that vary only in bits above the current mask will always
1181	>	* collide. (Among known examples are sets of Float keys holding
1182	>	* consecutive whole numbers in small tables.)  To counter this,
1183	>	* we apply a transform that spreads the impact of higher bits
1184	>	* downward. There is a tradeoff between speed, utility, and
1185	>	* quality of bit-spreading. Because many common sets of hashes
1186	>	* are already reasonably distributed across bits (so don't benefit
1187	>	* from spreading), and because we use trees to handle large sets
1188	>	* of collisions in bins, we don't need excessively high quality.
1189	>	*/
1190	>	private static final int spread(int h) {
1191	>	h ^= (h >>> 18) ^ (h >>> 12);
1192	>	return (h ^ (h >>> 10)) & HASH_BITS;
1193	>	}
1194	>
1195	>	/**
1196	>	* Replaces a list bin with a tree bin. Call only when locked.
1197	>	* Fails to replace if the given key is non-comparable or table
1198	>	* is, or needs, resizing.
1199	>	*/
1200	>	private final void replaceWithTreeBin(Node[] tab, int index, Object key) {
1201	>	if ((key instanceof Comparable) &&
1202	>	(tab.length >= MAXIMUM_CAPACITY || counter.sum() < (long)sizeCtl)) {
1203	>	TreeBin t = new TreeBin();
1204	>	for (Node e = tabAt(tab, index); e != null; e = e.next)
1205	>	t.putTreeNode(e.hash & HASH_BITS, e.key, e.val);
1206	>	setTabAt(tab, index, new Node(MOVED, t, null, null));
1207		}
1208	+	}
1209
1210	<	final V put(K key, int hash, V value, boolean onlyIfAbsent) {
1211	<	HashEntry<K,V> node = tryLock() ? null :
1212	<	scanAndLockForPut(key, hash, value);
1213	<	V oldValue;
1214	<	try {
1215	<	HashEntry<K,V>[] tab = table;
1216	<	int index = (tab.length - 1) & hash;
1217	<	HashEntry<K,V> first = entryAt(tab, index);
1218	<	for (HashEntry<K,V> e = first;;) {
1219	<	if (e != null) {
1220	<	K k;
1221	<	if ((k = e.key) == key ||
1222	<	(e.hash == hash && key.equals(k))) {
1223	<	oldValue = e.value;
1224	<	if (!onlyIfAbsent) {
1225	<	e.value = value;
1226	<	++modCount;
1210	>	/* ---------------- Internal access and update methods -------------- */
1211	>
1212	>	/** Implementation for get and containsKey */
1213	>	private final Object internalGet(Object k) {
1214	>	int h = spread(k.hashCode());
1215	>	retry: for (Node[] tab = table; tab != null;) {
1216	>	Node e, p; Object ek, ev; int eh;      // locals to read fields once
1217	>	for (e = tabAt(tab, (tab.length - 1) & h); e != null; e = e.next) {
1218	>	if ((eh = e.hash) == MOVED) {
1219	>	if ((ek = e.key) instanceof TreeBin)  // search TreeBin
1220	>	return ((TreeBin)ek).getValue(h, k);
1221	>	else {                        // restart with new table
1222	>	tab = (Node[])ek;
1223	>	continue retry;
1224	>	}
1225	>	}
1226	>	else if ((eh & HASH_BITS) == h && (ev = e.val) != null &&
1227	>	((ek = e.key) == k || k.equals(ek)))
1228	>	return ev;
1229	>	}
1230	>	break;
1231	>	}
1232	>	return null;
1233	>	}
1234	>
1235	>	/**
1236	>	* Implementation for the four public remove/replace methods:
1237	>	* Replaces node value with v, conditional upon match of cv if
1238	>	* non-null.  If resulting value is null, delete.
1239	>	*/
1240	>	private final Object internalReplace(Object k, Object v, Object cv) {
1241	>	int h = spread(k.hashCode());
1242	>	Object oldVal = null;
1243	>	for (Node[] tab = table;;) {
1244	>	Node f; int i, fh; Object fk;
1245	>	if (tab == null ||
1246	>	(f = tabAt(tab, i = (tab.length - 1) & h)) == null)
1247	>	break;
1248	>	else if ((fh = f.hash) == MOVED) {
1249	>	if ((fk = f.key) instanceof TreeBin) {
1250	>	TreeBin t = (TreeBin)fk;
1251	>	boolean validated = false;
1252	>	boolean deleted = false;
1253	>	t.acquire(0);
1254	>	try {
1255	>	if (tabAt(tab, i) == f) {
1256	>	validated = true;
1257	>	TreeNode p = t.getTreeNode(h, k, t.root);
1258	>	if (p != null) {
1259	>	Object pv = p.val;
1260	>	if (cv == null || cv == pv || cv.equals(pv)) {
1261	>	oldVal = pv;
1262	>	if ((p.val = v) == null) {
1263	>	deleted = true;
1264	>	t.deleteTreeNode(p);
1265	>	}
1266	>	}
1267		}
351	–	break;
1268		}
1269	<	e = e.next;
1269	>	} finally {
1270	>	t.release(0);
1271		}
1272	<	else {
1273	<	if (node != null)
1274	<	node.setNext(first);
358	<	else
359	<	node = new HashEntry<K,V>(hash, key, value, first);
360	<	int c = count + 1;
361	<	if (c > threshold && tab.length < MAXIMUM_CAPACITY)
362	<	rehash(node);
363	<	else
364	<	setEntryAt(tab, index, node);
365	<	++modCount;
366	<	count = c;
367	<	oldValue = null;
1272	>	if (validated) {
1273	>	if (deleted)
1274	>	counter.add(-1L);
1275		break;
1276		}
1277		}
1278	<	} finally {
1279	<	unlock();
1278	>	else
1279	>	tab = (Node[])fk;
1280	>	}
1281	>	else if ((fh & HASH_BITS) != h && f.next == null) // precheck
1282	>	break;                          // rules out possible existence
1283	>	else if ((fh & LOCKED) != 0) {
1284	>	checkForResize();               // try resizing if can't get lock
1285	>	f.tryAwaitLock(tab, i);
1286	>	}
1287	>	else if (f.casHash(fh, fh | LOCKED)) {
1288	>	boolean validated = false;
1289	>	boolean deleted = false;
1290	>	try {
1291	>	if (tabAt(tab, i) == f) {
1292	>	validated = true;
1293	>	for (Node e = f, pred = null;;) {
1294	>	Object ek, ev;
1295	>	if ((e.hash & HASH_BITS) == h &&
1296	>	((ev = e.val) != null) &&
1297	>	((ek = e.key) == k || k.equals(ek))) {
1298	>	if (cv == null || cv == ev || cv.equals(ev)) {
1299	>	oldVal = ev;
1300	>	if ((e.val = v) == null) {
1301	>	deleted = true;
1302	>	Node en = e.next;
1303	>	if (pred != null)
1304	>	pred.next = en;
1305	>	else
1306	>	setTabAt(tab, i, en);
1307	>	}
1308	>	}
1309	>	break;
1310	>	}
1311	>	pred = e;
1312	>	if ((e = e.next) == null)
1313	>	break;
1314	>	}
1315	>	}
1316	>	} finally {
1317	>	if (!f.casHash(fh | LOCKED, fh)) {
1318	>	f.hash = fh;
1319	>	synchronized (f) { f.notifyAll(); };
1320	>	}
1321	>	}
1322	>	if (validated) {
1323	>	if (deleted)
1324	>	counter.add(-1L);
1325	>	break;
1326	>	}
1327		}
374	–	return oldValue;
1328		}
1329	+	return oldVal;
1330	+	}
1331
1332	<	/**
1333	<	* Doubles size of table and repacks entries, also adding the
1334	<	* given node to new table
1335	<	*/
1336	<	@SuppressWarnings("unchecked")
1337	<	private void rehash(HashEntry<K,V> node) {
1338	<	/*
1339	<	* Reclassify nodes in each list to new table.  Because we
1340	<	* are using power-of-two expansion, the elements from
1341	<	* each bin must either stay at same index, or move with a
1342	<	* power of two offset. We eliminate unnecessary node
1343	<	* creation by catching cases where old nodes can be
1344	<	* reused because their next fields won't change.
1345	<	* Statistically, at the default threshold, only about
1346	<	* one-sixth of them need cloning when a table
1347	<	* doubles. The nodes they replace will be garbage
1348	<	* collectable as soon as they are no longer referenced by
1349	<	* any reader thread that may be in the midst of
1350	<	* concurrently traversing table. Entry accesses use plain
1351	<	* array indexing because they are followed by volatile
1352	<	* table write.
1353	<	*/
1354	<	HashEntry<K,V>[] oldTable = table;
1355	<	int oldCapacity = oldTable.length;
1356	<	int newCapacity = oldCapacity << 1;
1357	<	threshold = (int)(newCapacity * loadFactor);
1358	<	HashEntry<K,V>[] newTable =
1359	<	(HashEntry<K,V>[]) new HashEntry<?,?>[newCapacity];
1360	<	int sizeMask = newCapacity - 1;
1361	<	for (int i = 0; i < oldCapacity ; i++) {
1362	<	HashEntry<K,V> e = oldTable[i];
1363	<	if (e != null) {
1364	<	HashEntry<K,V> next = e.next;
1365	<	int idx = e.hash & sizeMask;
1366	<	if (next == null)   //  Single node on list
1367	<	newTable[idx] = e;
1368	<	else { // Reuse consecutive sequence at same slot
1369	<	HashEntry<K,V> lastRun = e;
1370	<	int lastIdx = idx;
1371	<	for (HashEntry<K,V> last = next;
1372	<	last != null;
1373	<	last = last.next) {
1374	<	int k = last.hash & sizeMask;
1375	<	if (k != lastIdx) {
1376	<	lastIdx = k;
1377	<	lastRun = last;
1378	<	}
1379	<	}
1380	<	newTable[lastIdx] = lastRun;
1381	<	// Clone remaining nodes
1382	<	for (HashEntry<K,V> p = e; p != lastRun; p = p.next) {
1383	<	V v = p.value;
1384	<	int h = p.hash;
1385	<	int k = h & sizeMask;
1386	<	HashEntry<K,V> n = newTable[k];
1387	<	newTable[k] = new HashEntry<K,V>(h, p.key, v, n);
1388	<	}
1389	<	}
1390	<	}
1391	<	}
1392	<	int nodeIndex = node.hash & sizeMask; // add the new node
1393	<	node.setNext(newTable[nodeIndex]);
1394	<	newTable[nodeIndex] = node;
1395	<	table = newTable;
1396	<	}
1397	<
1398	<	/**
1399	<	* Scans for a node containing given key while trying to
1400	<	* acquire lock, creating and returning one if not found. Upon
1401	<	* return, guarantees that lock is held. Unlike in most
1402	<	* methods, calls to method equals are not screened: Since
1403	<	* traversal speed doesn't matter, we might as well help warm
1404	<	* up the associated code and accesses as well.
1405	<	*
1406	<	* @return a new node if key not found, else null
1407	<	*/
1408	<	private HashEntry<K,V> scanAndLockForPut(K key, int hash, V value) {
1409	<	HashEntry<K,V> first = entryForHash(this, hash);
1410	<	HashEntry<K,V> e = first;
1411	<	HashEntry<K,V> node = null;
1412	<	int retries = -1; // negative while locating node
1413	<	while (!tryLock()) {
1414	<	HashEntry<K,V> f; // to recheck first below
1415	<	if (retries < 0) {
1416	<	if (e == null) {
1417	<	if (node == null) // speculatively create node
1418	<	node = new HashEntry<K,V>(hash, key, value, null);
1419	<	retries = 0;
1332	>	/*
1333	>	* Internal versions of the six insertion methods, each a
1334	>	* little more complicated than the last. All have
1335	>	* the same basic structure as the first (internalPut):
1336	>	*  1. If table uninitialized, create
1337	>	*  2. If bin empty, try to CAS new node
1338	>	*  3. If bin stale, use new table
1339	>	*  4. if bin converted to TreeBin, validate and relay to TreeBin methods
1340	>	*  5. Lock and validate; if valid, scan and add or update
1341	>	*
1342	>	* The others interweave other checks and/or alternative actions:
1343	>	*  * Plain put checks for and performs resize after insertion.
1344	>	*  * putIfAbsent prescans for mapping without lock (and fails to add
1345	>	*    if present), which also makes pre-emptive resize checks worthwhile.
1346	>	*  * computeIfAbsent extends form used in putIfAbsent with additional
1347	>	*    mechanics to deal with, calls, potential exceptions and null
1348	>	*    returns from function call.
1349	>	*  * compute uses the same function-call mechanics, but without
1350	>	*    the prescans
1351	>	*  * merge acts as putIfAbsent in the absent case, but invokes the
1352	>	*    update function if present
1353	>	*  * putAll attempts to pre-allocate enough table space
1354	>	*    and more lazily performs count updates and checks.
1355	>	*
1356	>	* Someday when details settle down a bit more, it might be worth
1357	>	* some factoring to reduce sprawl.
1358	>	*/
1359	>
1360	>	/** Implementation for put */
1361	>	private final Object internalPut(Object k, Object v) {
1362	>	int h = spread(k.hashCode());
1363	>	int count = 0;
1364	>	for (Node[] tab = table;;) {
1365	>	int i; Node f; int fh; Object fk;
1366	>	if (tab == null)
1367	>	tab = initTable();
1368	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1369	>	if (casTabAt(tab, i, null, new Node(h, k, v, null)))
1370	>	break;                   // no lock when adding to empty bin
1371	>	}
1372	>	else if ((fh = f.hash) == MOVED) {
1373	>	if ((fk = f.key) instanceof TreeBin) {
1374	>	TreeBin t = (TreeBin)fk;
1375	>	Object oldVal = null;
1376	>	t.acquire(0);
1377	>	try {
1378	>	if (tabAt(tab, i) == f) {
1379	>	count = 2;
1380	>	TreeNode p = t.putTreeNode(h, k, v);
1381	>	if (p != null) {
1382	>	oldVal = p.val;
1383	>	p.val = v;
1384	>	}
1385	>	}
1386	>	} finally {
1387	>	t.release(0);
1388	>	}
1389	>	if (count != 0) {
1390	>	if (oldVal != null)
1391	>	return oldVal;
1392	>	break;
1393	>	}
1394	>	}
1395	>	else
1396	>	tab = (Node[])fk;
1397	>	}
1398	>	else if ((fh & LOCKED) != 0) {
1399	>	checkForResize();
1400	>	f.tryAwaitLock(tab, i);
1401	>	}
1402	>	else if (f.casHash(fh, fh | LOCKED)) {
1403	>	Object oldVal = null;
1404	>	try {                        // needed in case equals() throws
1405	>	if (tabAt(tab, i) == f) {
1406	>	count = 1;
1407	>	for (Node e = f;; ++count) {
1408	>	Object ek, ev;
1409	>	if ((e.hash & HASH_BITS) == h &&
1410	>	(ev = e.val) != null &&
1411	>	((ek = e.key) == k || k.equals(ek))) {
1412	>	oldVal = ev;
1413	>	e.val = v;
1414	>	break;
1415	>	}
1416	>	Node last = e;
1417	>	if ((e = e.next) == null) {
1418	>	last.next = new Node(h, k, v, null);
1419	>	if (count >= TREE_THRESHOLD)
1420	>	replaceWithTreeBin(tab, i, k);
1421	>	break;
1422	>	}
1423	>	}
1424	>	}
1425	>	} finally {                  // unlock and signal if needed
1426	>	if (!f.casHash(fh | LOCKED, fh)) {
1427	>	f.hash = fh;
1428	>	synchronized (f) { f.notifyAll(); };
1429		}
466	–	else if (key.equals(e.key))
467	–	retries = 0;
468	–	else
469	–	e = e.next;
1430		}
1431	<	else if (++retries > MAX_SCAN_RETRIES) {
1432	<	lock();
1431	>	if (count != 0) {
1432	>	if (oldVal != null)
1433	>	return oldVal;
1434	>	if (tab.length <= 64)
1435	>	count = 2;
1436		break;
1437		}
1438	<	else if ((retries & 1) == 0 &&
1439	<	(f = entryForHash(this, hash)) != first) {
1440	<	e = first = f; // re-traverse if entry changed
1441	<	retries = -1;
1438	>	}
1439	>	}
1440	>	counter.add(1L);
1441	>	if (count > 1)
1442	>	checkForResize();
1443	>	return null;
1444	>	}
1445	>
1446	>	/** Implementation for putIfAbsent */
1447	>	private final Object internalPutIfAbsent(Object k, Object v) {
1448	>	int h = spread(k.hashCode());
1449	>	int count = 0;
1450	>	for (Node[] tab = table;;) {
1451	>	int i; Node f; int fh; Object fk, fv;
1452	>	if (tab == null)
1453	>	tab = initTable();
1454	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1455	>	if (casTabAt(tab, i, null, new Node(h, k, v, null)))
1456	>	break;
1457	>	}
1458	>	else if ((fh = f.hash) == MOVED) {
1459	>	if ((fk = f.key) instanceof TreeBin) {
1460	>	TreeBin t = (TreeBin)fk;
1461	>	Object oldVal = null;
1462	>	t.acquire(0);
1463	>	try {
1464	>	if (tabAt(tab, i) == f) {
1465	>	count = 2;
1466	>	TreeNode p = t.putTreeNode(h, k, v);
1467	>	if (p != null)
1468	>	oldVal = p.val;
1469	>	}
1470	>	} finally {
1471	>	t.release(0);
1472	>	}
1473	>	if (count != 0) {
1474	>	if (oldVal != null)
1475	>	return oldVal;
1476	>	break;
1477	>	}
1478	>	}
1479	>	else
1480	>	tab = (Node[])fk;
1481	>	}
1482	>	else if ((fh & HASH_BITS) == h && (fv = f.val) != null &&
1483	>	((fk = f.key) == k || k.equals(fk)))
1484	>	return fv;
1485	>	else {
1486	>	Node g = f.next;
1487	>	if (g != null) { // at least 2 nodes -- search and maybe resize
1488	>	for (Node e = g;;) {
1489	>	Object ek, ev;
1490	>	if ((e.hash & HASH_BITS) == h && (ev = e.val) != null &&
1491	>	((ek = e.key) == k || k.equals(ek)))
1492	>	return ev;
1493	>	if ((e = e.next) == null) {
1494	>	checkForResize();
1495	>	break;
1496	>	}
1497	>	}
1498	>	}
1499	>	if (((fh = f.hash) & LOCKED) != 0) {
1500	>	checkForResize();
1501	>	f.tryAwaitLock(tab, i);
1502	>	}
1503	>	else if (tabAt(tab, i) == f && f.casHash(fh, fh | LOCKED)) {
1504	>	Object oldVal = null;
1505	>	try {
1506	>	if (tabAt(tab, i) == f) {
1507	>	count = 1;
1508	>	for (Node e = f;; ++count) {
1509	>	Object ek, ev;
1510	>	if ((e.hash & HASH_BITS) == h &&
1511	>	(ev = e.val) != null &&
1512	>	((ek = e.key) == k || k.equals(ek))) {
1513	>	oldVal = ev;
1514	>	break;
1515	>	}
1516	>	Node last = e;
1517	>	if ((e = e.next) == null) {
1518	>	last.next = new Node(h, k, v, null);
1519	>	if (count >= TREE_THRESHOLD)
1520	>	replaceWithTreeBin(tab, i, k);
1521	>	break;
1522	>	}
1523	>	}
1524	>	}
1525	>	} finally {
1526	>	if (!f.casHash(fh | LOCKED, fh)) {
1527	>	f.hash = fh;
1528	>	synchronized (f) { f.notifyAll(); };
1529	>	}
1530	>	}
1531	>	if (count != 0) {
1532	>	if (oldVal != null)
1533	>	return oldVal;
1534	>	if (tab.length <= 64)
1535	>	count = 2;
1536	>	break;
1537	>	}
1538		}
1539		}
481	–	return node;
1540		}
1541	+	counter.add(1L);
1542	+	if (count > 1)
1543	+	checkForResize();
1544	+	return null;
1545	+	}
1546
1547	<	/**
1548	<	* Scans for a node containing the given key while trying to
1549	<	* acquire lock for a remove or replace operation. Upon
1550	<	* return, guarantees that lock is held.  Note that we must
1551	<	* lock even if the key is not found, to ensure sequential
1552	<	* consistency of updates.
1553	<	*/
1554	<	private void scanAndLock(Object key, int hash) {
1555	<	// similar to but simpler than scanAndLockForPut
1556	<	HashEntry<K,V> first = entryForHash(this, hash);
1557	<	HashEntry<K,V> e = first;
1558	<	int retries = -1;
1559	<	while (!tryLock()) {
1560	<	HashEntry<K,V> f;
1561	<	if (retries < 0) {
1562	<	if (e == null || key.equals(e.key))
1563	<	retries = 0;
1564	<	else
1565	<	e = e.next;
1547	>	/** Implementation for computeIfAbsent */
1548	>	private final Object internalComputeIfAbsent(K k,
1549	>	Fun<? super K, ?> mf) {
1550	>	int h = spread(k.hashCode());
1551	>	Object val = null;
1552	>	int count = 0;
1553	>	for (Node[] tab = table;;) {
1554	>	Node f; int i, fh; Object fk, fv;
1555	>	if (tab == null)
1556	>	tab = initTable();
1557	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1558	>	Node node = new Node(fh = h | LOCKED, k, null, null);
1559	>	if (casTabAt(tab, i, null, node)) {
1560	>	count = 1;
1561	>	try {
1562	>	if ((val = mf.apply(k)) != null)
1563	>	node.val = val;
1564	>	} finally {
1565	>	if (val == null)
1566	>	setTabAt(tab, i, null);
1567	>	if (!node.casHash(fh, h)) {
1568	>	node.hash = h;
1569	>	synchronized (node) { node.notifyAll(); };
1570	>	}
1571	>	}
1572		}
1573	<	else if (++retries > MAX_SCAN_RETRIES) {
505	<	lock();
1573	>	if (count != 0)
1574		break;
1575	+	}
1576	+	else if ((fh = f.hash) == MOVED) {
1577	+	if ((fk = f.key) instanceof TreeBin) {
1578	+	TreeBin t = (TreeBin)fk;
1579	+	boolean added = false;
1580	+	t.acquire(0);
1581	+	try {
1582	+	if (tabAt(tab, i) == f) {
1583	+	count = 1;
1584	+	TreeNode p = t.getTreeNode(h, k, t.root);
1585	+	if (p != null)
1586	+	val = p.val;
1587	+	else if ((val = mf.apply(k)) != null) {
1588	+	added = true;
1589	+	count = 2;
1590	+	t.putTreeNode(h, k, val);
1591	+	}
1592	+	}
1593	+	} finally {
1594	+	t.release(0);
1595	+	}
1596	+	if (count != 0) {
1597	+	if (!added)
1598	+	return val;
1599	+	break;
1600	+	}
1601		}
1602	<	else if ((retries & 1) == 0 &&
1603	<	(f = entryForHash(this, hash)) != first) {
1604	<	e = first = f;
1605	<	retries = -1;
1602	>	else
1603	>	tab = (Node[])fk;
1604	>	}
1605	>	else if ((fh & HASH_BITS) == h && (fv = f.val) != null &&
1606	>	((fk = f.key) == k || k.equals(fk)))
1607	>	return fv;
1608	>	else {
1609	>	Node g = f.next;
1610	>	if (g != null) {
1611	>	for (Node e = g;;) {
1612	>	Object ek, ev;
1613	>	if ((e.hash & HASH_BITS) == h && (ev = e.val) != null &&
1614	>	((ek = e.key) == k || k.equals(ek)))
1615	>	return ev;
1616	>	if ((e = e.next) == null) {
1617	>	checkForResize();
1618	>	break;
1619	>	}
1620	>	}
1621	>	}
1622	>	if (((fh = f.hash) & LOCKED) != 0) {
1623	>	checkForResize();
1624	>	f.tryAwaitLock(tab, i);
1625	>	}
1626	>	else if (tabAt(tab, i) == f && f.casHash(fh, fh | LOCKED)) {
1627	>	boolean added = false;
1628	>	try {
1629	>	if (tabAt(tab, i) == f) {
1630	>	count = 1;
1631	>	for (Node e = f;; ++count) {
1632	>	Object ek, ev;
1633	>	if ((e.hash & HASH_BITS) == h &&
1634	>	(ev = e.val) != null &&
1635	>	((ek = e.key) == k || k.equals(ek))) {
1636	>	val = ev;
1637	>	break;
1638	>	}
1639	>	Node last = e;
1640	>	if ((e = e.next) == null) {
1641	>	if ((val = mf.apply(k)) != null) {
1642	>	added = true;
1643	>	last.next = new Node(h, k, val, null);
1644	>	if (count >= TREE_THRESHOLD)
1645	>	replaceWithTreeBin(tab, i, k);
1646	>	}
1647	>	break;
1648	>	}
1649	>	}
1650	>	}
1651	>	} finally {
1652	>	if (!f.casHash(fh | LOCKED, fh)) {
1653	>	f.hash = fh;
1654	>	synchronized (f) { f.notifyAll(); };
1655	>	}
1656	>	}
1657	>	if (count != 0) {
1658	>	if (!added)
1659	>	return val;
1660	>	if (tab.length <= 64)
1661	>	count = 2;
1662	>	break;
1663	>	}
1664		}
1665		}
1666		}
1667	<
1668	<	/**
1669	<	* Remove; match on key only if value null, else match both.
1670	<	*/
1671	<	final V remove(Object key, int hash, Object value) {
1672	<	if (!tryLock())
1673	<	scanAndLock(key, hash);
1674	<	V oldValue = null;
1675	<	try {
1676	<	HashEntry<K,V>[] tab = table;
1677	<	int index = (tab.length - 1) & hash;
1678	<	HashEntry<K,V> e = entryAt(tab, index);
1679	<	HashEntry<K,V> pred = null;
1680	<	while (e != null) {
1681	<	K k;
1682	<	HashEntry<K,V> next = e.next;
1683	<	if ((k = e.key) == key ||
1684	<	(e.hash == hash && key.equals(k))) {
1685	<	V v = e.value;
1686	<	if (value == null || value == v || value.equals(v)) {
1687	<	if (pred == null)
1688	<	setEntryAt(tab, index, next);
1689	<	else
1690	<	pred.setNext(next);
1691	<	++modCount;
1692	<	--count;
1693	<	oldValue = v;
1667	>	if (val != null) {
1668	>	counter.add(1L);
1669	>	if (count > 1)
1670	>	checkForResize();
1671	>	}
1672	>	return val;
1673	>	}
1674	>
1675	>	/** Implementation for compute */
1676	>	@SuppressWarnings("unchecked") private final Object internalCompute
1677	>	(K k, boolean onlyIfPresent, BiFun<? super K, ? super V, ? extends V> mf) {
1678	>	int h = spread(k.hashCode());
1679	>	Object val = null;
1680	>	int delta = 0;
1681	>	int count = 0;
1682	>	for (Node[] tab = table;;) {
1683	>	Node f; int i, fh; Object fk;
1684	>	if (tab == null)
1685	>	tab = initTable();
1686	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1687	>	if (onlyIfPresent)
1688	>	break;
1689	>	Node node = new Node(fh = h | LOCKED, k, null, null);
1690	>	if (casTabAt(tab, i, null, node)) {
1691	>	try {
1692	>	count = 1;
1693	>	if ((val = mf.apply(k, null)) != null) {
1694	>	node.val = val;
1695	>	delta = 1;
1696	>	}
1697	>	} finally {
1698	>	if (delta == 0)
1699	>	setTabAt(tab, i, null);
1700	>	if (!node.casHash(fh, h)) {
1701	>	node.hash = h;
1702	>	synchronized (node) { node.notifyAll(); };
1703		}
1704	+	}
1705	+	}
1706	+	if (count != 0)
1707	+	break;
1708	+	}
1709	+	else if ((fh = f.hash) == MOVED) {
1710	+	if ((fk = f.key) instanceof TreeBin) {
1711	+	TreeBin t = (TreeBin)fk;
1712	+	t.acquire(0);
1713	+	try {
1714	+	if (tabAt(tab, i) == f) {
1715	+	count = 1;
1716	+	TreeNode p = t.getTreeNode(h, k, t.root);
1717	+	Object pv = (p == null) ? null : p.val;
1718	+	if ((val = mf.apply(k, (V)pv)) != null) {
1719	+	if (p != null)
1720	+	p.val = val;
1721	+	else {
1722	+	count = 2;
1723	+	delta = 1;
1724	+	t.putTreeNode(h, k, val);
1725	+	}
1726	+	}
1727	+	else if (p != null) {
1728	+	delta = -1;
1729	+	t.deleteTreeNode(p);
1730	+	}
1731	+	}
1732	+	} finally {
1733	+	t.release(0);
1734	+	}
1735	+	if (count != 0)
1736		break;
1737	+	}
1738	+	else
1739	+	tab = (Node[])fk;
1740	+	}
1741	+	else if ((fh & LOCKED) != 0) {
1742	+	checkForResize();
1743	+	f.tryAwaitLock(tab, i);
1744	+	}
1745	+	else if (f.casHash(fh, fh | LOCKED)) {
1746	+	try {
1747	+	if (tabAt(tab, i) == f) {
1748	+	count = 1;
1749	+	for (Node e = f, pred = null;; ++count) {
1750	+	Object ek, ev;
1751	+	if ((e.hash & HASH_BITS) == h &&
1752	+	(ev = e.val) != null &&
1753	+	((ek = e.key) == k || k.equals(ek))) {
1754	+	val = mf.apply(k, (V)ev);
1755	+	if (val != null)
1756	+	e.val = val;
1757	+	else {
1758	+	delta = -1;
1759	+	Node en = e.next;
1760	+	if (pred != null)
1761	+	pred.next = en;
1762	+	else
1763	+	setTabAt(tab, i, en);
1764	+	}
1765	+	break;
1766	+	}
1767	+	pred = e;
1768	+	if ((e = e.next) == null) {
1769	+	if (!onlyIfPresent && (val = mf.apply(k, null)) != null) {
1770	+	pred.next = new Node(h, k, val, null);
1771	+	delta = 1;
1772	+	if (count >= TREE_THRESHOLD)
1773	+	replaceWithTreeBin(tab, i, k);
1774	+	}
1775	+	break;
1776	+	}
1777	+	}
1778	+	}
1779	+	} finally {
1780	+	if (!f.casHash(fh | LOCKED, fh)) {
1781	+	f.hash = fh;
1782	+	synchronized (f) { f.notifyAll(); };
1783		}
545	–	pred = e;
546	–	e = next;
1784		}
1785	<	} finally {
1786	<	unlock();
1785	>	if (count != 0) {
1786	>	if (tab.length <= 64)
1787	>	count = 2;
1788	>	break;
1789	>	}
1790		}
551	–	return oldValue;
1791		}
1792	<
1793	<	final boolean replace(K key, int hash, V oldValue, V newValue) {
1794	<	if (!tryLock())
1795	<	scanAndLock(key, hash);
1796	<	boolean replaced = false;
1797	<	try {
1798	<	HashEntry<K,V> e;
1799	<	for (e = entryForHash(this, hash); e != null; e = e.next) {
1800	<	K k;
1801	<	if ((k = e.key) == key ||
1802	<	(e.hash == hash && key.equals(k))) {
1803	<	if (oldValue.equals(e.value)) {
1804	<	e.value = newValue;
1805	<	++modCount;
1806	<	replaced = true;
1792	>	if (delta != 0) {
1793	>	counter.add((long)delta);
1794	>	if (count > 1)
1795	>	checkForResize();
1796	>	}
1797	>	return val;
1798	>	}
1799	>
1800	>	/** Implementation for merge */
1801	>	@SuppressWarnings("unchecked") private final Object internalMerge
1802	>	(K k, V v, BiFun<? super V, ? super V, ? extends V> mf) {
1803	>	int h = spread(k.hashCode());
1804	>	Object val = null;
1805	>	int delta = 0;
1806	>	int count = 0;
1807	>	for (Node[] tab = table;;) {
1808	>	int i; Node f; int fh; Object fk, fv;
1809	>	if (tab == null)
1810	>	tab = initTable();
1811	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1812	>	if (casTabAt(tab, i, null, new Node(h, k, v, null))) {
1813	>	delta = 1;
1814	>	val = v;
1815	>	break;
1816	>	}
1817	>	}
1818	>	else if ((fh = f.hash) == MOVED) {
1819	>	if ((fk = f.key) instanceof TreeBin) {
1820	>	TreeBin t = (TreeBin)fk;
1821	>	t.acquire(0);
1822	>	try {
1823	>	if (tabAt(tab, i) == f) {
1824	>	count = 1;
1825	>	TreeNode p = t.getTreeNode(h, k, t.root);
1826	>	val = (p == null) ? v : mf.apply((V)p.val, v);
1827	>	if (val != null) {
1828	>	if (p != null)
1829	>	p.val = val;
1830	>	else {
1831	>	count = 2;
1832	>	delta = 1;
1833	>	t.putTreeNode(h, k, val);
1834	>	}
1835	>	}
1836	>	else if (p != null) {
1837	>	delta = -1;
1838	>	t.deleteTreeNode(p);
1839	>	}
1840		}
1841	+	} finally {
1842	+	t.release(0);
1843	+	}
1844	+	if (count != 0)
1845		break;
1846	+	}
1847	+	else
1848	+	tab = (Node[])fk;
1849	+	}
1850	+	else if ((fh & LOCKED) != 0) {
1851	+	checkForResize();
1852	+	f.tryAwaitLock(tab, i);
1853	+	}
1854	+	else if (f.casHash(fh, fh | LOCKED)) {
1855	+	try {
1856	+	if (tabAt(tab, i) == f) {
1857	+	count = 1;
1858	+	for (Node e = f, pred = null;; ++count) {
1859	+	Object ek, ev;
1860	+	if ((e.hash & HASH_BITS) == h &&
1861	+	(ev = e.val) != null &&
1862	+	((ek = e.key) == k || k.equals(ek))) {
1863	+	val = mf.apply(v, (V)ev);
1864	+	if (val != null)
1865	+	e.val = val;
1866	+	else {
1867	+	delta = -1;
1868	+	Node en = e.next;
1869	+	if (pred != null)
1870	+	pred.next = en;
1871	+	else
1872	+	setTabAt(tab, i, en);
1873	+	}
1874	+	break;
1875	+	}
1876	+	pred = e;
1877	+	if ((e = e.next) == null) {
1878	+	val = v;
1879	+	pred.next = new Node(h, k, val, null);
1880	+	delta = 1;
1881	+	if (count >= TREE_THRESHOLD)
1882	+	replaceWithTreeBin(tab, i, k);
1883	+	break;
1884	+	}
1885	+	}
1886	+	}
1887	+	} finally {
1888	+	if (!f.casHash(fh | LOCKED, fh)) {
1889	+	f.hash = fh;
1890	+	synchronized (f) { f.notifyAll(); };
1891		}
1892		}
1893	<	} finally {
1894	<	unlock();
1893	>	if (count != 0) {
1894	>	if (tab.length <= 64)
1895	>	count = 2;
1896	>	break;
1897	>	}
1898		}
575	–	return replaced;
1899		}
1900	+	if (delta != 0) {
1901	+	counter.add((long)delta);
1902	+	if (count > 1)
1903	+	checkForResize();
1904	+	}
1905	+	return val;
1906	+	}
1907	+
1908	+	/** Implementation for putAll */
1909	+	private final void internalPutAll(Map<?, ?> m) {
1910	+	tryPresize(m.size());
1911	+	long delta = 0L;     // number of uncommitted additions
1912	+	boolean npe = false; // to throw exception on exit for nulls
1913	+	try {                // to clean up counts on other exceptions
1914	+	for (Map.Entry<?, ?> entry : m.entrySet()) {
1915	+	Object k, v;
1916	+	if (entry == null || (k = entry.getKey()) == null ||
1917	+	(v = entry.getValue()) == null) {
1918	+	npe = true;
1919	+	break;
1920	+	}
1921	+	int h = spread(k.hashCode());
1922	+	for (Node[] tab = table;;) {
1923	+	int i; Node f; int fh; Object fk;
1924	+	if (tab == null)
1925	+	tab = initTable();
1926	+	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null){
1927	+	if (casTabAt(tab, i, null, new Node(h, k, v, null))) {
1928	+	++delta;
1929	+	break;
1930	+	}
1931	+	}
1932	+	else if ((fh = f.hash) == MOVED) {
1933	+	if ((fk = f.key) instanceof TreeBin) {
1934	+	TreeBin t = (TreeBin)fk;
1935	+	boolean validated = false;
1936	+	t.acquire(0);
1937	+	try {
1938	+	if (tabAt(tab, i) == f) {
1939	+	validated = true;
1940	+	TreeNode p = t.getTreeNode(h, k, t.root);
1941	+	if (p != null)
1942	+	p.val = v;
1943	+	else {
1944	+	t.putTreeNode(h, k, v);
1945	+	++delta;
1946	+	}
1947	+	}
1948	+	} finally {
1949	+	t.release(0);
1950	+	}
1951	+	if (validated)
1952	+	break;
1953	+	}
1954	+	else
1955	+	tab = (Node[])fk;
1956	+	}
1957	+	else if ((fh & LOCKED) != 0) {
1958	+	counter.add(delta);
1959	+	delta = 0L;
1960	+	checkForResize();
1961	+	f.tryAwaitLock(tab, i);
1962	+	}
1963	+	else if (f.casHash(fh, fh | LOCKED)) {
1964	+	int count = 0;
1965	+	try {
1966	+	if (tabAt(tab, i) == f) {
1967	+	count = 1;
1968	+	for (Node e = f;; ++count) {
1969	+	Object ek, ev;
1970	+	if ((e.hash & HASH_BITS) == h &&
1971	+	(ev = e.val) != null &&
1972	+	((ek = e.key) == k || k.equals(ek))) {
1973	+	e.val = v;
1974	+	break;
1975	+	}
1976	+	Node last = e;
1977	+	if ((e = e.next) == null) {
1978	+	++delta;
1979	+	last.next = new Node(h, k, v, null);
1980	+	if (count >= TREE_THRESHOLD)
1981	+	replaceWithTreeBin(tab, i, k);
1982	+	break;
1983	+	}
1984	+	}
1985	+	}
1986	+	} finally {
1987	+	if (!f.casHash(fh | LOCKED, fh)) {
1988	+	f.hash = fh;
1989	+	synchronized (f) { f.notifyAll(); };
1990	+	}
1991	+	}
1992	+	if (count != 0) {
1993	+	if (count > 1) {
1994	+	counter.add(delta);
1995	+	delta = 0L;
1996	+	checkForResize();
1997	+	}
1998	+	break;
1999	+	}
2000	+	}
2001	+	}
2002	+	}
2003	+	} finally {
2004	+	if (delta != 0)
2005	+	counter.add(delta);
2006	+	}
2007	+	if (npe)
2008	+	throw new NullPointerException();
2009	+	}
2010
2011	<	final V replace(K key, int hash, V value) {
2012	<	if (!tryLock())
2013	<	scanAndLock(key, hash);
2014	<	V oldValue = null;
2015	<	try {
2016	<	HashEntry<K,V> e;
2017	<	for (e = entryForHash(this, hash); e != null; e = e.next) {
2018	<	K k;
2019	<	if ((k = e.key) == key ||
2020	<	(e.hash == hash && key.equals(k))) {
2021	<	oldValue = e.value;
2022	<	e.value = value;
2023	<	++modCount;
2024	<	break;
2011	>	/* ---------------- Table Initialization and Resizing -------------- */
2012	>
2013	>	/**
2014	>	* Returns a power of two table size for the given desired capacity.
2015	>	* See Hackers Delight, sec 3.2
2016	>	*/
2017	>	private static final int tableSizeFor(int c) {
2018	>	int n = c - 1;
2019	>	n |= n >>> 1;
2020	>	n |= n >>> 2;
2021	>	n |= n >>> 4;
2022	>	n |= n >>> 8;
2023	>	n |= n >>> 16;
2024	>	return (n < 0) ? 1 : (n >= MAXIMUM_CAPACITY) ? MAXIMUM_CAPACITY : n + 1;
2025	>	}
2026	>
2027	>	/**
2028	>	* Initializes table, using the size recorded in sizeCtl.
2029	>	*/
2030	>	private final Node[] initTable() {
2031	>	Node[] tab; int sc;
2032	>	while ((tab = table) == null) {
2033	>	if ((sc = sizeCtl) < 0)
2034	>	Thread.yield(); // lost initialization race; just spin
2035	>	else if (UNSAFE.compareAndSwapInt(this, sizeCtlOffset, sc, -1)) {
2036	>	try {
2037	>	if ((tab = table) == null) {
2038	>	int n = (sc > 0) ? sc : DEFAULT_CAPACITY;
2039	>	tab = table = new Node[n];
2040	>	sc = n - (n >>> 2);
2041		}
2042	+	} finally {
2043	+	sizeCtl = sc;
2044		}
2045	<	} finally {
595	<	unlock();
2045	>	break;
2046		}
597	–	return oldValue;
2047		}
2048	+	return tab;
2049	+	}
2050
2051	<	final void clear() {
2052	<	lock();
2051	>	/**
2052	>	* If table is too small and not already resizing, creates next
2053	>	* table and transfers bins.  Rechecks occupancy after a transfer
2054	>	* to see if another resize is already needed because resizings
2055	>	* are lagging additions.
2056	>	*/
2057	>	private final void checkForResize() {
2058	>	Node[] tab; int n, sc;
2059	>	while ((tab = table) != null &&
2060	>	(n = tab.length) < MAXIMUM_CAPACITY &&
2061	>	(sc = sizeCtl) >= 0 && counter.sum() >= (long)sc &&
2062	>	UNSAFE.compareAndSwapInt(this, sizeCtlOffset, sc, -1)) {
2063		try {
2064	<	HashEntry<K,V>[] tab = table;
2065	<	for (int i = 0; i < tab.length ; i++)
2066	<	setEntryAt(tab, i, null);
2067	<	++modCount;
607	<	count = 0;
2064	>	if (tab == table) {
2065	>	table = rebuild(tab);
2066	>	sc = (n << 1) - (n >>> 1);
2067	>	}
2068		} finally {
2069	<	unlock();
2069	>	sizeCtl = sc;
2070		}
2071		}
2072		}
2073
614	–	// Accessing segments
615	–
2074		/**
2075	<	* Gets the jth element of given segment array (if nonnull) with
2076	<	* volatile element access semantics via Unsafe. (The null check
2077	<	* can trigger harmlessly only during deserialization.) Note:
620	<	* because each element of segments array is set only once (using
621	<	* fully ordered writes), some performance-sensitive methods rely
622	<	* on this method only as a recheck upon null reads.
2075	>	* Tries to presize table to accommodate the given number of elements.
2076	>	*
2077	>	* @param size number of elements (doesn't need to be perfectly accurate)
2078		*/
2079	<	@SuppressWarnings("unchecked")
2080	<	static final <K,V> Segment<K,V> segmentAt(Segment<K,V>[] ss, int j) {
2081	<	long u = (j << SSHIFT) + SBASE;
2082	<	return ss == null ? null :
2083	<	(Segment<K,V>) UNSAFE.getObjectVolatile(ss, u);
2079	>	private final void tryPresize(int size) {
2080	>	int c = (size >= (MAXIMUM_CAPACITY >>> 1)) ? MAXIMUM_CAPACITY :
2081	>	tableSizeFor(size + (size >>> 1) + 1);
2082	>	int sc;
2083	>	while ((sc = sizeCtl) >= 0) {
2084	>	Node[] tab = table; int n;
2085	>	if (tab == null || (n = tab.length) == 0) {
2086	>	n = (sc > c) ? sc : c;
2087	>	if (UNSAFE.compareAndSwapInt(this, sizeCtlOffset, sc, -1)) {
2088	>	try {
2089	>	if (table == tab) {
2090	>	table = new Node[n];
2091	>	sc = n - (n >>> 2);
2092	>	}
2093	>	} finally {
2094	>	sizeCtl = sc;
2095	>	}
2096	>	}
2097	>	}
2098	>	else if (c <= sc || n >= MAXIMUM_CAPACITY)
2099	>	break;
2100	>	else if (UNSAFE.compareAndSwapInt(this, sizeCtlOffset, sc, -1)) {
2101	>	try {
2102	>	if (table == tab) {
2103	>	table = rebuild(tab);
2104	>	sc = (n << 1) - (n >>> 1);
2105	>	}
2106	>	} finally {
2107	>	sizeCtl = sc;
2108	>	}
2109	>	}
2110	>	}
2111		}
2112
2113	<	/**
2114	<	* Returns the segment for the given index, creating it and
2115	<	* recording in segment table (via CAS) if not already present.
2113	>	/*
2114	>	* Moves and/or copies the nodes in each bin to new table. See
2115	>	* above for explanation.
2116		*
2117	<	* @param k the index
636	<	* @return the segment
2117	>	* @return the new table
2118		*/
2119	<	@SuppressWarnings("unchecked")
2120	<	private Segment<K,V> ensureSegment(int k) {
2121	<	final Segment<K,V>[] ss = this.segments;
2122	<	long u = (k << SSHIFT) + SBASE; // raw offset
2123	<	Segment<K,V> seg;
2124	<	if ((seg = (Segment<K,V>)UNSAFE.getObjectVolatile(ss, u)) == null) {
2125	<	Segment<K,V> proto = ss[0]; // use segment 0 as prototype
2126	<	int cap = proto.table.length;
2127	<	float lf = proto.loadFactor;
2128	<	int threshold = (int)(cap * lf);
2129	<	HashEntry<K,V>[] tab = (HashEntry<K,V>[])new HashEntry<?,?>[cap];
2130	<	if ((seg = (Segment<K,V>)UNSAFE.getObjectVolatile(ss, u))
2131	<	== null) { // recheck
2132	<	Segment<K,V> s = new Segment<K,V>(lf, threshold, tab);
2133	<	while ((seg = (Segment<K,V>)UNSAFE.getObjectVolatile(ss, u))
2134	<	== null) {
2135	<	if (UNSAFE.compareAndSwapObject(ss, u, null, seg = s))
2136	<	break;
2119	>	private static final Node[] rebuild(Node[] tab) {
2120	>	int n = tab.length;
2121	>	Node[] nextTab = new Node[n << 1];
2122	>	Node fwd = new Node(MOVED, nextTab, null, null);
2123	>	int[] buffer = null;       // holds bins to revisit; null until needed
2124	>	Node rev = null;           // reverse forwarder; null until needed
2125	>	int nbuffered = 0;         // the number of bins in buffer list
2126	>	int bufferIndex = 0;       // buffer index of current buffered bin
2127	>	int bin = n - 1;           // current non-buffered bin or -1 if none
2128	>
2129	>	for (int i = bin;;) {      // start upwards sweep
2130	>	int fh; Node f;
2131	>	if ((f = tabAt(tab, i)) == null) {
2132	>	if (bin >= 0) {    // Unbuffered; no lock needed (or available)
2133	>	if (!casTabAt(tab, i, f, fwd))
2134	>	continue;
2135	>	}
2136	>	else {             // transiently use a locked forwarding node
2137	>	Node g = new Node(MOVED|LOCKED, nextTab, null, null);
2138	>	if (!casTabAt(tab, i, f, g))
2139	>	continue;
2140	>	setTabAt(nextTab, i, null);
2141	>	setTabAt(nextTab, i + n, null);
2142	>	setTabAt(tab, i, fwd);
2143	>	if (!g.casHash(MOVED|LOCKED, MOVED)) {
2144	>	g.hash = MOVED;
2145	>	synchronized (g) { g.notifyAll(); }
2146	>	}
2147	>	}
2148	>	}
2149	>	else if ((fh = f.hash) == MOVED) {
2150	>	Object fk = f.key;
2151	>	if (fk instanceof TreeBin) {
2152	>	TreeBin t = (TreeBin)fk;
2153	>	boolean validated = false;
2154	>	t.acquire(0);
2155	>	try {
2156	>	if (tabAt(tab, i) == f) {
2157	>	validated = true;
2158	>	splitTreeBin(nextTab, i, t);
2159	>	setTabAt(tab, i, fwd);
2160	>	}
2161	>	} finally {
2162	>	t.release(0);
2163	>	}
2164	>	if (!validated)
2165	>	continue;
2166	>	}
2167	>	}
2168	>	else if ((fh & LOCKED) == 0 && f.casHash(fh, fh|LOCKED)) {
2169	>	boolean validated = false;
2170	>	try {              // split to lo and hi lists; copying as needed
2171	>	if (tabAt(tab, i) == f) {
2172	>	validated = true;
2173	>	splitBin(nextTab, i, f);
2174	>	setTabAt(tab, i, fwd);
2175	>	}
2176	>	} finally {
2177	>	if (!f.casHash(fh | LOCKED, fh)) {
2178	>	f.hash = fh;
2179	>	synchronized (f) { f.notifyAll(); };
2180	>	}
2181		}
2182	+	if (!validated)
2183	+	continue;
2184		}
2185	+	else {
2186	+	if (buffer == null) // initialize buffer for revisits
2187	+	buffer = new int[TRANSFER_BUFFER_SIZE];
2188	+	if (bin < 0 && bufferIndex > 0) {
2189	+	int j = buffer[--bufferIndex];
2190	+	buffer[bufferIndex] = i;
2191	+	i = j;         // swap with another bin
2192	+	continue;
2193	+	}
2194	+	if (bin < 0 || nbuffered >= TRANSFER_BUFFER_SIZE) {
2195	+	f.tryAwaitLock(tab, i);
2196	+	continue;      // no other options -- block
2197	+	}
2198	+	if (rev == null)   // initialize reverse-forwarder
2199	+	rev = new Node(MOVED, tab, null, null);
2200	+	if (tabAt(tab, i) != f || (f.hash & LOCKED) == 0)
2201	+	continue;      // recheck before adding to list
2202	+	buffer[nbuffered++] = i;
2203	+	setTabAt(nextTab, i, rev);     // install place-holders
2204	+	setTabAt(nextTab, i + n, rev);
2205	+	}
2206	+
2207	+	if (bin > 0)
2208	+	i = --bin;
2209	+	else if (buffer != null && nbuffered > 0) {
2210	+	bin = -1;
2211	+	i = buffer[bufferIndex = --nbuffered];
2212	+	}
2213	+	else
2214	+	return nextTab;
2215		}
659	–	return seg;
2216		}
2217
662	–	// Hash-based segment and entry accesses
663	–
2218		/**
2219	<	* Gets the segment for the given hash code.
2219	>	* Splits a normal bin with list headed by e into lo and hi parts;
2220	>	* installs in given table.
2221		*/
2222	<	@SuppressWarnings("unchecked")
2223	<	private Segment<K,V> segmentForHash(int h) {
2224	<	long u = (((h >>> segmentShift) & segmentMask) << SSHIFT) + SBASE;
2225	<	return (Segment<K,V>) UNSAFE.getObjectVolatile(segments, u);
2222	>	private static void splitBin(Node[] nextTab, int i, Node e) {
2223	>	int bit = nextTab.length >>> 1; // bit to split on
2224	>	int runBit = e.hash & bit;
2225	>	Node lastRun = e, lo = null, hi = null;
2226	>	for (Node p = e.next; p != null; p = p.next) {
2227	>	int b = p.hash & bit;
2228	>	if (b != runBit) {
2229	>	runBit = b;
2230	>	lastRun = p;
2231	>	}
2232	>	}
2233	>	if (runBit == 0)
2234	>	lo = lastRun;
2235	>	else
2236	>	hi = lastRun;
2237	>	for (Node p = e; p != lastRun; p = p.next) {
2238	>	int ph = p.hash & HASH_BITS;
2239	>	Object pk = p.key, pv = p.val;
2240	>	if ((ph & bit) == 0)
2241	>	lo = new Node(ph, pk, pv, lo);
2242	>	else
2243	>	hi = new Node(ph, pk, pv, hi);
2244	>	}
2245	>	setTabAt(nextTab, i, lo);
2246	>	setTabAt(nextTab, i + bit, hi);
2247		}
2248
2249		/**
2250	<	* Gets the table entry for the given segment and hash code.
2251	<	*/
2252	<	@SuppressWarnings("unchecked")
2253	<	static final <K,V> HashEntry<K,V> entryForHash(Segment<K,V> seg, int h) {
2254	<	HashEntry<K,V>[] tab;
2255	<	return (seg == null || (tab = seg.table) == null) ? null :
2256	<	(HashEntry<K,V>) UNSAFE.getObjectVolatile
2257	<	(tab, ((long)(((tab.length - 1) & h)) << TSHIFT) + TBASE);
2250	>	* Splits a tree bin into lo and hi parts; installs in given table.
2251	>	*/
2252	>	private static void splitTreeBin(Node[] nextTab, int i, TreeBin t) {
2253	>	int bit = nextTab.length >>> 1;
2254	>	TreeBin lt = new TreeBin();
2255	>	TreeBin ht = new TreeBin();
2256	>	int lc = 0, hc = 0;
2257	>	for (Node e = t.first; e != null; e = e.next) {
2258	>	int h = e.hash & HASH_BITS;
2259	>	Object k = e.key, v = e.val;
2260	>	if ((h & bit) == 0) {
2261	>	++lc;
2262	>	lt.putTreeNode(h, k, v);
2263	>	}
2264	>	else {
2265	>	++hc;
2266	>	ht.putTreeNode(h, k, v);
2267	>	}
2268	>	}
2269	>	Node ln, hn; // throw away trees if too small
2270	>	if (lc <= (TREE_THRESHOLD >>> 1)) {
2271	>	ln = null;
2272	>	for (Node p = lt.first; p != null; p = p.next)
2273	>	ln = new Node(p.hash, p.key, p.val, ln);
2274	>	}
2275	>	else
2276	>	ln = new Node(MOVED, lt, null, null);
2277	>	setTabAt(nextTab, i, ln);
2278	>	if (hc <= (TREE_THRESHOLD >>> 1)) {
2279	>	hn = null;
2280	>	for (Node p = ht.first; p != null; p = p.next)
2281	>	hn = new Node(p.hash, p.key, p.val, hn);
2282	>	}
2283	>	else
2284	>	hn = new Node(MOVED, ht, null, null);
2285	>	setTabAt(nextTab, i + bit, hn);
2286		}
2287
2288	<	/* ---------------- Public operations -------------- */
2288	>	/**
2289	>	* Implementation for clear. Steps through each bin, removing all
2290	>	* nodes.
2291	>	*/
2292	>	private final void internalClear() {
2293	>	long delta = 0L; // negative number of deletions
2294	>	int i = 0;
2295	>	Node[] tab = table;
2296	>	while (tab != null && i < tab.length) {
2297	>	int fh; Object fk;
2298	>	Node f = tabAt(tab, i);
2299	>	if (f == null)
2300	>	++i;
2301	>	else if ((fh = f.hash) == MOVED) {
2302	>	if ((fk = f.key) instanceof TreeBin) {
2303	>	TreeBin t = (TreeBin)fk;
2304	>	t.acquire(0);
2305	>	try {
2306	>	if (tabAt(tab, i) == f) {
2307	>	for (Node p = t.first; p != null; p = p.next) {
2308	>	if (p.val != null) { // (currently always true)
2309	>	p.val = null;
2310	>	--delta;
2311	>	}
2312	>	}
2313	>	t.first = null;
2314	>	t.root = null;
2315	>	++i;
2316	>	}
2317	>	} finally {
2318	>	t.release(0);
2319	>	}
2320	>	}
2321	>	else
2322	>	tab = (Node[])fk;
2323	>	}
2324	>	else if ((fh & LOCKED) != 0) {
2325	>	counter.add(delta); // opportunistically update count
2326	>	delta = 0L;
2327	>	f.tryAwaitLock(tab, i);
2328	>	}
2329	>	else if (f.casHash(fh, fh | LOCKED)) {
2330	>	try {
2331	>	if (tabAt(tab, i) == f) {
2332	>	for (Node e = f; e != null; e = e.next) {
2333	>	if (e.val != null) {  // (currently always true)
2334	>	e.val = null;
2335	>	--delta;
2336	>	}
2337	>	}
2338	>	setTabAt(tab, i, null);
2339	>	++i;
2340	>	}
2341	>	} finally {
2342	>	if (!f.casHash(fh | LOCKED, fh)) {
2343	>	f.hash = fh;
2344	>	synchronized (f) { f.notifyAll(); };
2345	>	}
2346	>	}
2347	>	}
2348	>	}
2349	>	if (delta != 0)
2350	>	counter.add(delta);
2351	>	}
2352	>
2353	>	/* ----------------Table Traversal -------------- */
2354
2355		/**
2356	<	* Creates a new, empty map with the specified initial
2357	<	* capacity, load factor and concurrency level.
2356	>	* Encapsulates traversal for methods such as containsValue; also
2357	>	* serves as a base class for other iterators and bulk tasks.
2358		*
2359	<	* @param initialCapacity the initial capacity. The implementation
2360	<	* performs internal sizing to accommodate this many elements.
2361	<	* @param loadFactor  the load factor threshold, used to control resizing.
2362	<	* Resizing may be performed when the average number of elements per
2363	<	* bin exceeds this threshold.
2364	<	* @param concurrencyLevel the estimated number of concurrently
2365	<	* updating threads. The implementation performs internal sizing
2366	<	* to try to accommodate this many threads.
2367	<	* @throws IllegalArgumentException if the initial capacity is
2368	<	* negative or the load factor or concurrencyLevel are
2369	<	* nonpositive.
2359	>	* At each step, the iterator snapshots the key ("nextKey") and
2360	>	* value ("nextVal") of a valid node (i.e., one that, at point of
2361	>	* snapshot, has a non-null user value). Because val fields can
2362	>	* change (including to null, indicating deletion), field nextVal
2363	>	* might not be accurate at point of use, but still maintains the
2364	>	* weak consistency property of holding a value that was once
2365	>	* valid. To support iterator.remove, the nextKey field is not
2366	>	* updated (nulled out) when the iterator cannot advance.
2367	>	*
2368	>	* Internal traversals directly access these fields, as in:
2369	>	* {@code while (it.advance() != null) { process(it.nextKey); }}
2370	>	*
2371	>	* Exported iterators must track whether the iterator has advanced
2372	>	* (in hasNext vs next) (by setting/checking/nulling field
2373	>	* nextVal), and then extract key, value, or key-value pairs as
2374	>	* return values of next().
2375	>	*
2376	>	* The iterator visits once each still-valid node that was
2377	>	* reachable upon iterator construction. It might miss some that
2378	>	* were added to a bin after the bin was visited, which is OK wrt
2379	>	* consistency guarantees. Maintaining this property in the face
2380	>	* of possible ongoing resizes requires a fair amount of
2381	>	* bookkeeping state that is difficult to optimize away amidst
2382	>	* volatile accesses.  Even so, traversal maintains reasonable
2383	>	* throughput.
2384	>	*
2385	>	* Normally, iteration proceeds bin-by-bin traversing lists.
2386	>	* However, if the table has been resized, then all future steps
2387	>	* must traverse both the bin at the current index as well as at
2388	>	* (index + baseSize); and so on for further resizings. To
2389	>	* paranoically cope with potential sharing by users of iterators
2390	>	* across threads, iteration terminates if a bounds checks fails
2391	>	* for a table read.
2392	>	*
2393	>	* This class extends CountedCompleter to streamline parallel
2394	>	* iteration in bulk operations. This adds only a few fields of
2395	>	* space overhead, which is small enough in cases where it is not
2396	>	* needed to not worry about it.  Because CountedCompleter is
2397	>	* Serializable, but iterators need not be, we need to add warning
2398	>	* suppressions.
2399	>	*/
2400	>	@SuppressWarnings("serial") static class Traverser<K,V,R> extends CountedCompleter<R> {
2401	>	final ConcurrentHashMap<K, V> map;
2402	>	Node next;           // the next entry to use
2403	>	Object nextKey;      // cached key field of next
2404	>	Object nextVal;      // cached val field of next
2405	>	Node[] tab;          // current table; updated if resized
2406	>	int index;           // index of bin to use next
2407	>	int baseIndex;       // current index of initial table
2408	>	int baseLimit;       // index bound for initial table
2409	>	int baseSize;        // initial table size
2410	>	int batch;           // split control
2411	>
2412	>	/** Creates iterator for all entries in the table. */
2413	>	Traverser(ConcurrentHashMap<K, V> map) {
2414	>	this.map = map;
2415	>	}
2416	>
2417	>	/** Creates iterator for split() methods and task constructors */
2418	>	Traverser(ConcurrentHashMap<K,V> map, Traverser<K,V,?> it, int batch) {
2419	>	super(it);
2420	>	this.batch = batch;
2421	>	if ((this.map = map) != null && it != null) { // split parent
2422	>	Node[] t;
2423	>	if ((t = it.tab) == null &&
2424	>	(t = it.tab = map.table) != null)
2425	>	it.baseLimit = it.baseSize = t.length;
2426	>	this.tab = t;
2427	>	this.baseSize = it.baseSize;
2428	>	int hi = this.baseLimit = it.baseLimit;
2429	>	it.baseLimit = this.index = this.baseIndex =
2430	>	(hi + it.baseIndex + 1) >>> 1;
2431	>	}
2432	>	}
2433	>
2434	>	/**
2435	>	* Advances next; returns nextVal or null if terminated.
2436	>	* See above for explanation.
2437	>	*/
2438	>	final Object advance() {
2439	>	Node e = next;
2440	>	Object ev = null;
2441	>	outer: do {
2442	>	if (e != null)                  // advance past used/skipped node
2443	>	e = e.next;
2444	>	while (e == null) {             // get to next non-null bin
2445	>	ConcurrentHashMap<K, V> m;
2446	>	Node[] t; int b, i, n; Object ek; // checks must use locals
2447	>	if ((t = tab) != null)
2448	>	n = t.length;
2449	>	else if ((m = map) != null && (t = tab = m.table) != null)
2450	>	n = baseLimit = baseSize = t.length;
2451	>	else
2452	>	break outer;
2453	>	if ((b = baseIndex) >= baseLimit ||
2454	>	(i = index) < 0 || i >= n)
2455	>	break outer;
2456	>	if ((e = tabAt(t, i)) != null && e.hash == MOVED) {
2457	>	if ((ek = e.key) instanceof TreeBin)
2458	>	e = ((TreeBin)ek).first;
2459	>	else {
2460	>	tab = (Node[])ek;
2461	>	continue;           // restarts due to null val
2462	>	}
2463	>	}                           // visit upper slots if present
2464	>	index = (i += baseSize) < n ? i : (baseIndex = b + 1);
2465	>	}
2466	>	nextKey = e.key;
2467	>	} while ((ev = e.val) == null);    // skip deleted or special nodes
2468	>	next = e;
2469	>	return nextVal = ev;
2470	>	}
2471	>
2472	>	public final void remove() {
2473	>	Object k = nextKey;
2474	>	if (k == null && (advance() == null || (k = nextKey) == null))
2475	>	throw new IllegalStateException();
2476	>	map.internalReplace(k, null, null);
2477	>	}
2478	>
2479	>	public final boolean hasNext() {
2480	>	return nextVal != null || advance() != null;
2481	>	}
2482	>
2483	>	public final boolean hasMoreElements() { return hasNext(); }
2484	>
2485	>	public void compute() { } // default no-op CountedCompleter body
2486	>
2487	>	/**
2488	>	* Returns a batch value > 0 if this task should (and must) be
2489	>	* split, if so, adding to pending count, and in any case
2490	>	* updating batch value. The initial batch value is approx
2491	>	* exp2 of the number of times (minus one) to split task by
2492	>	* two before executing leaf action. This value is faster to
2493	>	* compute and more convenient to use as a guide to splitting
2494	>	* than is the depth, since it is used while dividing by two
2495	>	* anyway.
2496	>	*/
2497	>	final int preSplit() {
2498	>	ConcurrentHashMap<K, V> m; int b; Node[] t;  ForkJoinPool pool;
2499	>	if ((b = batch) < 0 && (m = map) != null) { // force initialization
2500	>	if ((t = tab) == null && (t = tab = m.table) != null)
2501	>	baseLimit = baseSize = t.length;
2502	>	if (t != null) {
2503	>	long n = m.counter.sum();
2504	>	int par = ((pool = getPool()) == null) ?
2505	>	ForkJoinPool.getCommonPoolParallelism() :
2506	>	pool.getParallelism();
2507	>	int sp = par << 3; // slack of 8
2508	>	b = (n <= 0L) ? 0 : (n < (long)sp) ? (int)n : sp;
2509	>	}
2510	>	}
2511	>	b = (b <= 1 || baseIndex == baseLimit) ? 0 : (b >>> 1);
2512	>	if ((batch = b) > 0)
2513	>	addToPendingCount(1);
2514	>	return b;
2515	>	}
2516	>
2517	>	}
2518	>
2519	>	/* ---------------- Public operations -------------- */
2520	>
2521	>	/**
2522	>	* Creates a new, empty map with the default initial table size (16).
2523		*/
2524	<	@SuppressWarnings("unchecked")
2525	<	public ConcurrentHashMap(int initialCapacity,
704	<	float loadFactor, int concurrencyLevel) {
705	<	if (!(loadFactor > 0) || initialCapacity < 0 || concurrencyLevel <= 0)
706	<	throw new IllegalArgumentException();
707	<	if (concurrencyLevel > MAX_SEGMENTS)
708	<	concurrencyLevel = MAX_SEGMENTS;
709	<	// Find power-of-two sizes best matching arguments
710	<	int sshift = 0;
711	<	int ssize = 1;
712	<	while (ssize < concurrencyLevel) {
713	<	++sshift;
714	<	ssize <<= 1;
715	<	}
716	<	this.segmentShift = 32 - sshift;
717	<	this.segmentMask = ssize - 1;
718	<	if (initialCapacity > MAXIMUM_CAPACITY)
719	<	initialCapacity = MAXIMUM_CAPACITY;
720	<	int c = initialCapacity / ssize;
721	<	if (c * ssize < initialCapacity)
722	<	++c;
723	<	int cap = MIN_SEGMENT_TABLE_CAPACITY;
724	<	while (cap < c)
725	<	cap <<= 1;
726	<	// create segments and segments[0]
727	<	Segment<K,V> s0 =
728	<	new Segment<K,V>(loadFactor, (int)(cap * loadFactor),
729	<	(HashEntry<K,V>[])new HashEntry<?,?>[cap]);
730	<	Segment<K,V>[] ss = (Segment<K,V>[])new Segment<?,?>[ssize];
731	<	UNSAFE.putOrderedObject(ss, SBASE, s0); // ordered write of segments[0]
732	<	this.segments = ss;
2524	>	public ConcurrentHashMap() {
2525	>	this.counter = new LongAdder();
2526		}
2527
2528		/**
2529	<	* Creates a new, empty map with the specified initial capacity
2530	<	* and load factor and with the default concurrencyLevel (16).
2529	>	* Creates a new, empty map with an initial table size
2530	>	* accommodating the specified number of elements without the need
2531	>	* to dynamically resize.
2532		*
2533		* @param initialCapacity The implementation performs internal
2534		* sizing to accommodate this many elements.
2535	<	* @param loadFactor  the load factor threshold, used to control resizing.
2536	<	* Resizing may be performed when the average number of elements per
2537	<	* bin exceeds this threshold.
2535	>	* @throws IllegalArgumentException if the initial capacity of
2536	>	* elements is negative
2537	>	*/
2538	>	public ConcurrentHashMap(int initialCapacity) {
2539	>	if (initialCapacity < 0)
2540	>	throw new IllegalArgumentException();
2541	>	int cap = ((initialCapacity >= (MAXIMUM_CAPACITY >>> 1)) ?
2542	>	MAXIMUM_CAPACITY :
2543	>	tableSizeFor(initialCapacity + (initialCapacity >>> 1) + 1));
2544	>	this.counter = new LongAdder();
2545	>	this.sizeCtl = cap;
2546	>	}
2547	>
2548	>	/**
2549	>	* Creates a new map with the same mappings as the given map.
2550	>	*
2551	>	* @param m the map
2552	>	*/
2553	>	public ConcurrentHashMap(Map<? extends K, ? extends V> m) {
2554	>	this.counter = new LongAdder();
2555	>	this.sizeCtl = DEFAULT_CAPACITY;
2556	>	internalPutAll(m);
2557	>	}
2558	>
2559	>	/**
2560	>	* Creates a new, empty map with an initial table size based on
2561	>	* the given number of elements ({@code initialCapacity}) and
2562	>	* initial table density ({@code loadFactor}).
2563	>	*
2564	>	* @param initialCapacity the initial capacity. The implementation
2565	>	* performs internal sizing to accommodate this many elements,
2566	>	* given the specified load factor.
2567	>	* @param loadFactor the load factor (table density) for
2568	>	* establishing the initial table size
2569		* @throws IllegalArgumentException if the initial capacity of
2570		* elements is negative or the load factor is nonpositive
2571		*
2572		* @since 1.6
2573		*/
2574		public ConcurrentHashMap(int initialCapacity, float loadFactor) {
2575	<	this(initialCapacity, loadFactor, DEFAULT_CONCURRENCY_LEVEL);
2575	>	this(initialCapacity, loadFactor, 1);
2576		}
2577
2578		/**
2579	<	* Creates a new, empty map with the specified initial capacity,
2580	<	* and with default load factor (0.75) and concurrencyLevel (16).
2579	>	* Creates a new, empty map with an initial table size based on
2580	>	* the given number of elements ({@code initialCapacity}), table
2581	>	* density ({@code loadFactor}), and number of concurrently
2582	>	* updating threads ({@code concurrencyLevel}).
2583		*
2584		* @param initialCapacity the initial capacity. The implementation
2585	<	* performs internal sizing to accommodate this many elements.
2586	<	* @throws IllegalArgumentException if the initial capacity of
2587	<	* elements is negative.
2585	>	* performs internal sizing to accommodate this many elements,
2586	>	* given the specified load factor.
2587	>	* @param loadFactor the load factor (table density) for
2588	>	* establishing the initial table size
2589	>	* @param concurrencyLevel the estimated number of concurrently
2590	>	* updating threads. The implementation may use this value as
2591	>	* a sizing hint.
2592	>	* @throws IllegalArgumentException if the initial capacity is
2593	>	* negative or the load factor or concurrencyLevel are
2594	>	* nonpositive
2595		*/
2596	<	public ConcurrentHashMap(int initialCapacity) {
2597	<	this(initialCapacity, DEFAULT_LOAD_FACTOR, DEFAULT_CONCURRENCY_LEVEL);
2596	>	public ConcurrentHashMap(int initialCapacity,
2597	>	float loadFactor, int concurrencyLevel) {
2598	>	if (!(loadFactor > 0.0f) || initialCapacity < 0 || concurrencyLevel <= 0)
2599	>	throw new IllegalArgumentException();
2600	>	if (initialCapacity < concurrencyLevel)   // Use at least as many bins
2601	>	initialCapacity = concurrencyLevel;   // as estimated threads
2602	>	long size = (long)(1.0 + (long)initialCapacity / loadFactor);
2603	>	int cap = (size >= (long)MAXIMUM_CAPACITY) ?
2604	>	MAXIMUM_CAPACITY : tableSizeFor((int)size);
2605	>	this.counter = new LongAdder();
2606	>	this.sizeCtl = cap;
2607		}
2608
2609		/**
2610	<	* Creates a new, empty map with a default initial capacity (16),
2611	<	* load factor (0.75) and concurrencyLevel (16).
2610	>	* Creates a new {@link Set} backed by a ConcurrentHashMap
2611	>	* from the given type to {@code Boolean.TRUE}.
2612	>	*
2613	>	* @return the new set
2614		*/
2615	<	public ConcurrentHashMap() {
2616	<	this(DEFAULT_INITIAL_CAPACITY, DEFAULT_LOAD_FACTOR, DEFAULT_CONCURRENCY_LEVEL);
2615	>	public static <K> KeySetView<K,Boolean> newKeySet() {
2616	>	return new KeySetView<K,Boolean>(new ConcurrentHashMap<K,Boolean>(),
2617	>	Boolean.TRUE);
2618		}
2619
2620		/**
2621	<	* Creates a new map with the same mappings as the given map.
2622	<	* The map is created with a capacity of 1.5 times the number
777	<	* of mappings in the given map or 16 (whichever is greater),
778	<	* and a default load factor (0.75) and concurrencyLevel (16).
2621	>	* Creates a new {@link Set} backed by a ConcurrentHashMap
2622	>	* from the given type to {@code Boolean.TRUE}.
2623		*
2624	<	* @param m the map
2624	>	* @param initialCapacity The implementation performs internal
2625	>	* sizing to accommodate this many elements.
2626	>	* @throws IllegalArgumentException if the initial capacity of
2627	>	* elements is negative
2628	>	* @return the new set
2629		*/
2630	<	public ConcurrentHashMap(Map<? extends K, ? extends V> m) {
2631	<	this(Math.max((int) (m.size() / DEFAULT_LOAD_FACTOR) + 1,
2632	<	DEFAULT_INITIAL_CAPACITY),
785	<	DEFAULT_LOAD_FACTOR, DEFAULT_CONCURRENCY_LEVEL);
786	<	putAll(m);
2630	>	public static <K> KeySetView<K,Boolean> newKeySet(int initialCapacity) {
2631	>	return new KeySetView<K,Boolean>(new ConcurrentHashMap<K,Boolean>(initialCapacity),
2632	>	Boolean.TRUE);
2633		}
2634
2635		/**
2636	<	* Returns <tt>true</tt> if this map contains no key-value mappings.
791	<	*
792	<	* @return <tt>true</tt> if this map contains no key-value mappings
2636	>	* {@inheritDoc}
2637		*/
2638		public boolean isEmpty() {
2639	<	/*
796	<	* Sum per-segment modCounts to avoid mis-reporting when
797	<	* elements are concurrently added and removed in one segment
798	<	* while checking another, in which case the table was never
799	<	* actually empty at any point. (The sum ensures accuracy up
800	<	* through at least 1<<31 per-segment modifications before
801	<	* recheck.)  Methods size() and containsValue() use similar
802	<	* constructions for stability checks.
803	<	*/
804	<	long sum = 0L;
805	<	final Segment<K,V>[] segments = this.segments;
806	<	for (int j = 0; j < segments.length; ++j) {
807	<	Segment<K,V> seg = segmentAt(segments, j);
808	<	if (seg != null) {
809	<	if (seg.count != 0)
810	<	return false;
811	<	sum += seg.modCount;
812	<	}
813	<	}
814	<	if (sum != 0L) { // recheck unless no modifications
815	<	for (int j = 0; j < segments.length; ++j) {
816	<	Segment<K,V> seg = segmentAt(segments, j);
817	<	if (seg != null) {
818	<	if (seg.count != 0)
819	<	return false;
820	<	sum -= seg.modCount;
821	<	}
822	<	}
823	<	if (sum != 0L)
824	<	return false;
825	<	}
826	<	return true;
2639	>	return counter.sum() <= 0L; // ignore transient negative values
2640		}
2641
2642		/**
2643	<	* Returns the number of key-value mappings in this map.  If the
831	<	* map contains more than <tt>Integer.MAX_VALUE</tt> elements, returns
832	<	* <tt>Integer.MAX_VALUE</tt>.
833	<	*
834	<	* @return the number of key-value mappings in this map
2643	>	* {@inheritDoc}
2644		*/
2645		public int size() {
2646	<	// Try a few times to get accurate count. On failure due to
2647	<	// continuous async changes in table, resort to locking.
2648	<	final Segment<K,V>[] segments = this.segments;
2649	<	final int segmentCount = segments.length;
2650	<
842	<	long previousSum = 0L;
843	<	for (int retries = -1; retries < RETRIES_BEFORE_LOCK; retries++) {
844	<	long sum = 0L;    // sum of modCounts
845	<	long size = 0L;
846	<	for (int i = 0; i < segmentCount; i++) {
847	<	Segment<K,V> segment = segmentAt(segments, i);
848	<	if (segment != null) {
849	<	sum += segment.modCount;
850	<	size += segment.count;
851	<	}
852	<	}
853	<	if (sum == previousSum)
854	<	return ((size >>> 31) == 0) ? (int) size : Integer.MAX_VALUE;
855	<	previousSum = sum;
856	<	}
2646	>	long n = counter.sum();
2647	>	return ((n < 0L) ? 0 :
2648	>	(n > (long)Integer.MAX_VALUE) ? Integer.MAX_VALUE :
2649	>	(int)n);
2650	>	}
2651
2652	<	long size = 0L;
2653	<	for (int i = 0; i < segmentCount; i++) {
2654	<	Segment<K,V> segment = ensureSegment(i);
2655	<	segment.lock();
2656	<	size += segment.count;
2657	<	}
2658	<	for (int i = 0; i < segmentCount; i++)
2659	<	segments[i].unlock();
2660	<	return ((size >>> 31) == 0) ? (int) size : Integer.MAX_VALUE;
2652	>	/**
2653	>	* Returns the number of mappings. This method should be used
2654	>	* instead of {@link #size} because a ConcurrentHashMap may
2655	>	* contain more mappings than can be represented as an int. The
2656	>	* value returned is an estimate; the actual count may differ if
2657	>	* there are concurrent insertions or removals.
2658	>	*
2659	>	* @return the number of mappings
2660	>	*/
2661	>	public long mappingCount() {
2662	>	long n = counter.sum();
2663	>	return (n < 0L) ? 0L : n; // ignore transient negative values
2664		}
2665
2666		/**
#	Line 877 | Line 2674 | public class ConcurrentHashMap<K, V> ext
2674		*
2675		* @throws NullPointerException if the specified key is null
2676		*/
2677	<	@SuppressWarnings("unchecked")
2678	<	public V get(Object key) {
2679	<	Segment<K,V> s; // manually integrate access methods to reduce overhead
2680	<	HashEntry<K,V>[] tab;
2681	<	int h = hash(key.hashCode());
2682	<	long u = (((h >>> segmentShift) & segmentMask) << SSHIFT) + SBASE;
2683	<	if ((s = (Segment<K,V>)UNSAFE.getObjectVolatile(segments, u)) != null &&
2684	<	(tab = s.table) != null) {
2685	<	for (HashEntry<K,V> e = (HashEntry<K,V>) UNSAFE.getObjectVolatile
2686	<	(tab, ((long)(((tab.length - 1) & h)) << TSHIFT) + TBASE);
2687	<	e != null; e = e.next) {
2688	<	K k;
2689	<	if ((k = e.key) == key || (e.hash == h && key.equals(k)))
2690	<	return e.value;
2691	<	}
2692	<	}
2693	<	return null;
2677	>	@SuppressWarnings("unchecked") public V get(Object key) {
2678	>	if (key == null)
2679	>	throw new NullPointerException();
2680	>	return (V)internalGet(key);
2681	>	}
2682	>
2683	>	/**
2684	>	* Returns the value to which the specified key is mapped,
2685	>	* or the given defaultValue if this map contains no mapping for the key.
2686	>	*
2687	>	* @param key the key
2688	>	* @param defaultValue the value to return if this map contains
2689	>	* no mapping for the given key
2690	>	* @return the mapping for the key, if present; else the defaultValue
2691	>	* @throws NullPointerException if the specified key is null
2692	>	*/
2693	>	@SuppressWarnings("unchecked") public V getValueOrDefault(Object key, V defaultValue) {
2694	>	if (key == null)
2695	>	throw new NullPointerException();
2696	>	V v = (V) internalGet(key);
2697	>	return v == null ? defaultValue : v;
2698		}
2699
2700		/**
2701		* Tests if the specified object is a key in this table.
2702		*
2703		* @param  key   possible key
2704	<	* @return <tt>true</tt> if and only if the specified object
2704	>	* @return {@code true} if and only if the specified object
2705		*         is a key in this table, as determined by the
2706	<	*         <tt>equals</tt> method; <tt>false</tt> otherwise.
2706	>	*         {@code equals} method; {@code false} otherwise
2707		* @throws NullPointerException if the specified key is null
2708		*/
908	–	@SuppressWarnings("unchecked")
2709		public boolean containsKey(Object key) {
2710	<	Segment<K,V> s; // same as get() except no need for volatile value read
2711	<	HashEntry<K,V>[] tab;
2712	<	int h = hash(key.hashCode());
913	<	long u = (((h >>> segmentShift) & segmentMask) << SSHIFT) + SBASE;
914	<	if ((s = (Segment<K,V>)UNSAFE.getObjectVolatile(segments, u)) != null &&
915	<	(tab = s.table) != null) {
916	<	for (HashEntry<K,V> e = (HashEntry<K,V>) UNSAFE.getObjectVolatile
917	<	(tab, ((long)(((tab.length - 1) & h)) << TSHIFT) + TBASE);
918	<	e != null; e = e.next) {
919	<	K k;
920	<	if ((k = e.key) == key || (e.hash == h && key.equals(k)))
921	<	return true;
922	<	}
923	<	}
924	<	return false;
2710	>	if (key == null)
2711	>	throw new NullPointerException();
2712	>	return internalGet(key) != null;
2713		}
2714
2715		/**
2716	<	* Returns <tt>true</tt> if this map maps one or more keys to the
2717	<	* specified value. Note: This method requires a full internal
2718	<	* traversal of the hash table, and so is much slower than
931	<	* method <tt>containsKey</tt>.
2716	>	* Returns {@code true} if this map maps one or more keys to the
2717	>	* specified value. Note: This method may require a full traversal
2718	>	* of the map, and is much slower than method {@code containsKey}.
2719		*
2720		* @param value value whose presence in this map is to be tested
2721	<	* @return <tt>true</tt> if this map maps one or more keys to the
2721	>	* @return {@code true} if this map maps one or more keys to the
2722		*         specified value
2723		* @throws NullPointerException if the specified value is null
2724		*/
2725		public boolean containsValue(Object value) {
939	–	// Same idea as size()
2726		if (value == null)
2727		throw new NullPointerException();
2728	<	final Segment<K,V>[] segments = this.segments;
2729	<	long previousSum = 0L;
2730	<	int lockCount = 0;
2731	<	try {
2732	<	for (int retries = -1; ; retries++) {
947	<	long sum = 0L;    // sum of modCounts
948	<	for (int j = 0; j < segments.length; j++) {
949	<	Segment<K,V> segment;
950	<	if (retries == RETRIES_BEFORE_LOCK) {
951	<	segment = ensureSegment(j);
952	<	segment.lock();
953	<	lockCount++;
954	<	} else {
955	<	segment = segmentAt(segments, j);
956	<	if (segment == null)
957	<	continue;
958	<	}
959	<	HashEntry<K,V>[] tab = segment.table;
960	<	if (tab != null) {
961	<	for (int i = 0 ; i < tab.length; i++) {
962	<	HashEntry<K,V> e;
963	<	for (e = entryAt(tab, i); e != null; e = e.next) {
964	<	V v = e.value;
965	<	if (v != null && value.equals(v))
966	<	return true;
967	<	}
968	<	}
969	<	sum += segment.modCount;
970	<	}
971	<	}
972	<	if ((retries >= 0 && sum == previousSum) || lockCount > 0)
973	<	return false;
974	<	previousSum = sum;
975	<	}
976	<	} finally {
977	<	for (int j = 0; j < lockCount; j++)
978	<	segments[j].unlock();
2728	>	Object v;
2729	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
2730	>	while ((v = it.advance()) != null) {
2731	>	if (v == value || value.equals(v))
2732	>	return true;
2733		}
2734	+	return false;
2735		}
2736
2737		/**
#	Line 988 | Line 2743 | public class ConcurrentHashMap<K, V> ext
2743		* Java Collections framework.
2744		*
2745		* @param  value a value to search for
2746	<	* @return <tt>true</tt> if and only if some key maps to the
2747	<	*         <tt>value</tt> argument in this table as
2748	<	*         determined by the <tt>equals</tt> method;
2749	<	*         <tt>false</tt> otherwise
2746	>	* @return {@code true} if and only if some key maps to the
2747	>	*         {@code value} argument in this table as
2748	>	*         determined by the {@code equals} method;
2749	>	*         {@code false} otherwise
2750		* @throws NullPointerException if the specified value is null
2751		*/
2752		public boolean contains(Object value) {
#	Line 1002 | Line 2757 | public class ConcurrentHashMap<K, V> ext
2757		* Maps the specified key to the specified value in this table.
2758		* Neither the key nor the value can be null.
2759		*
2760	<	* <p> The value can be retrieved by calling the <tt>get</tt> method
2760	>	* <p>The value can be retrieved by calling the {@code get} method
2761		* with a key that is equal to the original key.
2762		*
2763		* @param key key with which the specified value is to be associated
2764		* @param value value to be associated with the specified key
2765	<	* @return the previous value associated with <tt>key</tt>, or
2766	<	*         <tt>null</tt> if there was no mapping for <tt>key</tt>
2765	>	* @return the previous value associated with {@code key}, or
2766	>	*         {@code null} if there was no mapping for {@code key}
2767		* @throws NullPointerException if the specified key or value is null
2768		*/
2769	<	@SuppressWarnings("unchecked")
2770	<	public V put(K key, V value) {
1016	<	Segment<K,V> s;
1017	<	if (value == null)
2769	>	@SuppressWarnings("unchecked") public V put(K key, V value) {
2770	>	if (key == null || value == null)
2771		throw new NullPointerException();
2772	<	int hash = hash(key.hashCode());
1020	<	int j = (hash >>> segmentShift) & segmentMask;
1021	<	if ((s = (Segment<K,V>)UNSAFE.getObject          // nonvolatile; recheck
1022	<	(segments, (j << SSHIFT) + SBASE)) == null) //  in ensureSegment
1023	<	s = ensureSegment(j);
1024	<	return s.put(key, hash, value, false);
2772	>	return (V)internalPut(key, value);
2773		}
2774
2775		/**
2776		* {@inheritDoc}
2777		*
2778		* @return the previous value associated with the specified key,
2779	<	*         or <tt>null</tt> if there was no mapping for the key
2779	>	*         or {@code null} if there was no mapping for the key
2780		* @throws NullPointerException if the specified key or value is null
2781		*/
2782	<	@SuppressWarnings("unchecked")
2783	<	public V putIfAbsent(K key, V value) {
1036	<	Segment<K,V> s;
1037	<	if (value == null)
2782	>	@SuppressWarnings("unchecked") public V putIfAbsent(K key, V value) {
2783	>	if (key == null || value == null)
2784		throw new NullPointerException();
2785	<	int hash = hash(key.hashCode());
1040	<	int j = (hash >>> segmentShift) & segmentMask;
1041	<	if ((s = (Segment<K,V>)UNSAFE.getObject
1042	<	(segments, (j << SSHIFT) + SBASE)) == null)
1043	<	s = ensureSegment(j);
1044	<	return s.put(key, hash, value, true);
2785	>	return (V)internalPutIfAbsent(key, value);
2786		}
2787
2788		/**
#	Line 1052 | Line 2793 | public class ConcurrentHashMap<K, V> ext
2793		* @param m mappings to be stored in this map
2794		*/
2795		public void putAll(Map<? extends K, ? extends V> m) {
2796	<	for (Map.Entry<? extends K, ? extends V> e : m.entrySet())
2797	<	put(e.getKey(), e.getValue());
2796	>	internalPutAll(m);
2797	>	}
2798	>
2799	>	/**
2800	>	* If the specified key is not already associated with a value,
2801	>	* computes its value using the given mappingFunction and enters
2802	>	* it into the map unless null.  This is equivalent to
2803	>	* <pre> {@code
2804	>	* if (map.containsKey(key))
2805	>	*   return map.get(key);
2806	>	* value = mappingFunction.apply(key);
2807	>	* if (value != null)
2808	>	*   map.put(key, value);
2809	>	* return value;}</pre>
2810	>	*
2811	>	* except that the action is performed atomically.  If the
2812	>	* function returns {@code null} no mapping is recorded. If the
2813	>	* function itself throws an (unchecked) exception, the exception
2814	>	* is rethrown to its caller, and no mapping is recorded.  Some
2815	>	* attempted update operations on this map by other threads may be
2816	>	* blocked while computation is in progress, so the computation
2817	>	* should be short and simple, and must not attempt to update any
2818	>	* other mappings of this Map. The most appropriate usage is to
2819	>	* construct a new object serving as an initial mapped value, or
2820	>	* memoized result, as in:
2821	>	*
2822	>	*  <pre> {@code
2823	>	* map.computeIfAbsent(key, new Fun<K, V>() {
2824	>	*   public V map(K k) { return new Value(f(k)); }});}</pre>
2825	>	*
2826	>	* @param key key with which the specified value is to be associated
2827	>	* @param mappingFunction the function to compute a value
2828	>	* @return the current (existing or computed) value associated with
2829	>	*         the specified key, or null if the computed value is null
2830	>	* @throws NullPointerException if the specified key or mappingFunction
2831	>	*         is null
2832	>	* @throws IllegalStateException if the computation detectably
2833	>	*         attempts a recursive update to this map that would
2834	>	*         otherwise never complete
2835	>	* @throws RuntimeException or Error if the mappingFunction does so,
2836	>	*         in which case the mapping is left unestablished
2837	>	*/
2838	>	@SuppressWarnings("unchecked") public V computeIfAbsent
2839	>	(K key, Fun<? super K, ? extends V> mappingFunction) {
2840	>	if (key == null || mappingFunction == null)
2841	>	throw new NullPointerException();
2842	>	return (V)internalComputeIfAbsent(key, mappingFunction);
2843	>	}
2844	>
2845	>	/**
2846	>	* If the given key is present, computes a new mapping value given a key and
2847	>	* its current mapped value. This is equivalent to
2848	>	*  <pre> {@code
2849	>	*   if (map.containsKey(key)) {
2850	>	*     value = remappingFunction.apply(key, map.get(key));
2851	>	*     if (value != null)
2852	>	*       map.put(key, value);
2853	>	*     else
2854	>	*       map.remove(key);
2855	>	*   }
2856	>	* }</pre>
2857	>	*
2858	>	* except that the action is performed atomically.  If the
2859	>	* function returns {@code null}, the mapping is removed.  If the
2860	>	* function itself throws an (unchecked) exception, the exception
2861	>	* is rethrown to its caller, and the current mapping is left
2862	>	* unchanged.  Some attempted update operations on this map by
2863	>	* other threads may be blocked while computation is in progress,
2864	>	* so the computation should be short and simple, and must not
2865	>	* attempt to update any other mappings of this Map.
2866	>	*
2867	>	* @param key key with which the specified value is to be associated
2868	>	* @param remappingFunction the function to compute a value
2869	>	* @return the new value associated with the specified key, or null if none
2870	>	* @throws NullPointerException if the specified key or remappingFunction
2871	>	*         is null
2872	>	* @throws IllegalStateException if the computation detectably
2873	>	*         attempts a recursive update to this map that would
2874	>	*         otherwise never complete
2875	>	* @throws RuntimeException or Error if the remappingFunction does so,
2876	>	*         in which case the mapping is unchanged
2877	>	*/
2878	>	@SuppressWarnings("unchecked") public V computeIfPresent
2879	>	(K key, BiFun<? super K, ? super V, ? extends V> remappingFunction) {
2880	>	if (key == null || remappingFunction == null)
2881	>	throw new NullPointerException();
2882	>	return (V)internalCompute(key, true, remappingFunction);
2883	>	}
2884	>
2885	>	/**
2886	>	* Computes a new mapping value given a key and
2887	>	* its current mapped value (or {@code null} if there is no current
2888	>	* mapping). This is equivalent to
2889	>	*  <pre> {@code
2890	>	*   value = remappingFunction.apply(key, map.get(key));
2891	>	*   if (value != null)
2892	>	*     map.put(key, value);
2893	>	*   else
2894	>	*     map.remove(key);
2895	>	* }</pre>
2896	>	*
2897	>	* except that the action is performed atomically.  If the
2898	>	* function returns {@code null}, the mapping is removed.  If the
2899	>	* function itself throws an (unchecked) exception, the exception
2900	>	* is rethrown to its caller, and the current mapping is left
2901	>	* unchanged.  Some attempted update operations on this map by
2902	>	* other threads may be blocked while computation is in progress,
2903	>	* so the computation should be short and simple, and must not
2904	>	* attempt to update any other mappings of this Map. For example,
2905	>	* to either create or append new messages to a value mapping:
2906	>	*
2907	>	* <pre> {@code
2908	>	* Map<Key, String> map = ...;
2909	>	* final String msg = ...;
2910	>	* map.compute(key, new BiFun<Key, String, String>() {
2911	>	*   public String apply(Key k, String v) {
2912	>	*    return (v == null) ? msg : v + msg;});}}</pre>
2913	>	*
2914	>	* @param key key with which the specified value is to be associated
2915	>	* @param remappingFunction the function to compute a value
2916	>	* @return the new value associated with the specified key, or null if none
2917	>	* @throws NullPointerException if the specified key or remappingFunction
2918	>	*         is null
2919	>	* @throws IllegalStateException if the computation detectably
2920	>	*         attempts a recursive update to this map that would
2921	>	*         otherwise never complete
2922	>	* @throws RuntimeException or Error if the remappingFunction does so,
2923	>	*         in which case the mapping is unchanged
2924	>	*/
2925	>	@SuppressWarnings("unchecked") public V compute
2926	>	(K key, BiFun<? super K, ? super V, ? extends V> remappingFunction) {
2927	>	if (key == null || remappingFunction == null)
2928	>	throw new NullPointerException();
2929	>	return (V)internalCompute(key, false, remappingFunction);
2930	>	}
2931	>
2932	>	/**
2933	>	* If the specified key is not already associated
2934	>	* with a value, associate it with the given value.
2935	>	* Otherwise, replace the value with the results of
2936	>	* the given remapping function. This is equivalent to:
2937	>	*  <pre> {@code
2938	>	*   if (!map.containsKey(key))
2939	>	*     map.put(value);
2940	>	*   else {
2941	>	*     newValue = remappingFunction.apply(map.get(key), value);
2942	>	*     if (value != null)
2943	>	*       map.put(key, value);
2944	>	*     else
2945	>	*       map.remove(key);
2946	>	*   }
2947	>	* }</pre>
2948	>	* except that the action is performed atomically.  If the
2949	>	* function returns {@code null}, the mapping is removed.  If the
2950	>	* function itself throws an (unchecked) exception, the exception
2951	>	* is rethrown to its caller, and the current mapping is left
2952	>	* unchanged.  Some attempted update operations on this map by
2953	>	* other threads may be blocked while computation is in progress,
2954	>	* so the computation should be short and simple, and must not
2955	>	* attempt to update any other mappings of this Map.
2956	>	*/
2957	>	@SuppressWarnings("unchecked") public V merge
2958	>	(K key, V value, BiFun<? super V, ? super V, ? extends V> remappingFunction) {
2959	>	if (key == null || value == null || remappingFunction == null)
2960	>	throw new NullPointerException();
2961	>	return (V)internalMerge(key, value, remappingFunction);
2962		}
2963
2964		/**
#	Line 1061 | Line 2966 | public class ConcurrentHashMap<K, V> ext
2966		* This method does nothing if the key is not in the map.
2967		*
2968		* @param  key the key that needs to be removed
2969	<	* @return the previous value associated with <tt>key</tt>, or
2970	<	*         <tt>null</tt> if there was no mapping for <tt>key</tt>
2969	>	* @return the previous value associated with {@code key}, or
2970	>	*         {@code null} if there was no mapping for {@code key}
2971		* @throws NullPointerException if the specified key is null
2972		*/
2973	<	public V remove(Object key) {
2974	<	int hash = hash(key.hashCode());
2975	<	Segment<K,V> s = segmentForHash(hash);
2976	<	return s == null ? null : s.remove(key, hash, null);
2973	>	@SuppressWarnings("unchecked") public V remove(Object key) {
2974	>	if (key == null)
2975	>	throw new NullPointerException();
2976	>	return (V)internalReplace(key, null, null);
2977		}
2978
2979		/**
#	Line 1077 | Line 2982 | public class ConcurrentHashMap<K, V> ext
2982		* @throws NullPointerException if the specified key is null
2983		*/
2984		public boolean remove(Object key, Object value) {
2985	<	int hash = hash(key.hashCode());
2986	<	Segment<K,V> s;
2987	<	return value != null && (s = segmentForHash(hash)) != null &&
2988	<	s.remove(key, hash, value) != null;
2985	>	if (key == null)
2986	>	throw new NullPointerException();
2987	>	if (value == null)
2988	>	return false;
2989	>	return internalReplace(key, null, value) != null;
2990		}
2991
2992		/**
#	Line 1089 | Line 2995 | public class ConcurrentHashMap<K, V> ext
2995		* @throws NullPointerException if any of the arguments are null
2996		*/
2997		public boolean replace(K key, V oldValue, V newValue) {
2998	<	int hash = hash(key.hashCode());
1093	<	if (oldValue == null || newValue == null)
2998	>	if (key == null || oldValue == null || newValue == null)
2999		throw new NullPointerException();
3000	<	Segment<K,V> s = segmentForHash(hash);
1096	<	return s != null && s.replace(key, hash, oldValue, newValue);
3000	>	return internalReplace(key, newValue, oldValue) != null;
3001		}
3002
3003		/**
3004		* {@inheritDoc}
3005		*
3006		* @return the previous value associated with the specified key,
3007	<	*         or <tt>null</tt> if there was no mapping for the key
3007	>	*         or {@code null} if there was no mapping for the key
3008		* @throws NullPointerException if the specified key or value is null
3009		*/
3010	<	public V replace(K key, V value) {
3011	<	int hash = hash(key.hashCode());
1108	<	if (value == null)
3010	>	@SuppressWarnings("unchecked") public V replace(K key, V value) {
3011	>	if (key == null || value == null)
3012		throw new NullPointerException();
3013	<	Segment<K,V> s = segmentForHash(hash);
1111	<	return s == null ? null : s.replace(key, hash, value);
3013	>	return (V)internalReplace(key, value, null);
3014		}
3015
3016		/**
3017		* Removes all of the mappings from this map.
3018		*/
3019		public void clear() {
3020	<	final Segment<K,V>[] segments = this.segments;
1119	<	for (int j = 0; j < segments.length; ++j) {
1120	<	Segment<K,V> s = segmentAt(segments, j);
1121	<	if (s != null)
1122	<	s.clear();
1123	<	}
3020	>	internalClear();
3021		}
3022
3023		/**
3024		* Returns a {@link Set} view of the keys contained in this map.
3025		* The set is backed by the map, so changes to the map are
3026	<	* reflected in the set, and vice-versa.  The set supports element
1130	<	* removal, which removes the corresponding mapping from this map,
1131	<	* via the <tt>Iterator.remove</tt>, <tt>Set.remove</tt>,
1132	<	* <tt>removeAll</tt>, <tt>retainAll</tt>, and <tt>clear</tt>
1133	<	* operations.  It does not support the <tt>add</tt> or
1134	<	* <tt>addAll</tt> operations.
3026	>	* reflected in the set, and vice-versa.
3027		*
3028	<	* <p>The view's <tt>iterator</tt> is a "weakly consistent" iterator
3029	<	* that will never throw {@link ConcurrentModificationException},
3030	<	* and guarantees to traverse elements as they existed upon
3031	<	* construction of the iterator, and may (but is not guaranteed to)
3032	<	* reflect any modifications subsequent to construction.
3028	>	* @return the set view
3029	>	*/
3030	>	public KeySetView<K,V> keySet() {
3031	>	KeySetView<K,V> ks = keySet;
3032	>	return (ks != null) ? ks : (keySet = new KeySetView<K,V>(this, null));
3033	>	}
3034	>
3035	>	/**
3036	>	* Returns a {@link Set} view of the keys in this map, using the
3037	>	* given common mapped value for any additions (i.e., {@link
3038	>	* Collection#add} and {@link Collection#addAll}). This is of
3039	>	* course only appropriate if it is acceptable to use the same
3040	>	* value for all additions from this view.
3041	>	*
3042	>	* @param mappedValue the mapped value to use for any
3043	>	* additions.
3044	>	* @return the set view
3045	>	* @throws NullPointerException if the mappedValue is null
3046		*/
3047	<	public Set<K> keySet() {
3048	<	Set<K> ks = keySet;
3049	<	return (ks != null) ? ks : (keySet = new KeySet());
3047	>	public KeySetView<K,V> keySet(V mappedValue) {
3048	>	if (mappedValue == null)
3049	>	throw new NullPointerException();
3050	>	return new KeySetView<K,V>(this, mappedValue);
3051		}
3052
3053		/**
3054		* Returns a {@link Collection} view of the values contained in this map.
3055		* The collection is backed by the map, so changes to the map are
3056	<	* reflected in the collection, and vice-versa.  The collection
1151	<	* supports element removal, which removes the corresponding
1152	<	* mapping from this map, via the <tt>Iterator.remove</tt>,
1153	<	* <tt>Collection.remove</tt>, <tt>removeAll</tt>,
1154	<	* <tt>retainAll</tt>, and <tt>clear</tt> operations.  It does not
1155	<	* support the <tt>add</tt> or <tt>addAll</tt> operations.
1156	<	*
1157	<	* <p>The view's <tt>iterator</tt> is a "weakly consistent" iterator
1158	<	* that will never throw {@link ConcurrentModificationException},
1159	<	* and guarantees to traverse elements as they existed upon
1160	<	* construction of the iterator, and may (but is not guaranteed to)
1161	<	* reflect any modifications subsequent to construction.
3056	>	* reflected in the collection, and vice-versa.
3057		*/
3058	<	public Collection<V> values() {
3059	<	Collection<V> vs = values;
3060	<	return (vs != null) ? vs : (values = new Values());
3058	>	public ValuesView<K,V> values() {
3059	>	ValuesView<K,V> vs = values;
3060	>	return (vs != null) ? vs : (values = new ValuesView<K,V>(this));
3061		}
3062
3063		/**
#	Line 1170 | Line 3065 | public class ConcurrentHashMap<K, V> ext
3065		* The set is backed by the map, so changes to the map are
3066		* reflected in the set, and vice-versa.  The set supports element
3067		* removal, which removes the corresponding mapping from the map,
3068	<	* via the <tt>Iterator.remove</tt>, <tt>Set.remove</tt>,
3069	<	* <tt>removeAll</tt>, <tt>retainAll</tt>, and <tt>clear</tt>
3070	<	* operations.  It does not support the <tt>add</tt> or
3071	<	* <tt>addAll</tt> operations.
3068	>	* via the {@code Iterator.remove}, {@code Set.remove},
3069	>	* {@code removeAll}, {@code retainAll}, and {@code clear}
3070	>	* operations.  It does not support the {@code add} or
3071	>	* {@code addAll} operations.
3072		*
3073	<	* <p>The view's <tt>iterator</tt> is a "weakly consistent" iterator
3073	>	* <p>The view's {@code iterator} is a "weakly consistent" iterator
3074		* that will never throw {@link ConcurrentModificationException},
3075		* and guarantees to traverse elements as they existed upon
3076		* construction of the iterator, and may (but is not guaranteed to)
3077		* reflect any modifications subsequent to construction.
3078		*/
3079		public Set<Map.Entry<K,V>> entrySet() {
3080	<	Set<Map.Entry<K,V>> es = entrySet;
3081	<	return (es != null) ? es : (entrySet = new EntrySet());
3080	>	EntrySetView<K,V> es = entrySet;
3081	>	return (es != null) ? es : (entrySet = new EntrySetView<K,V>(this));
3082		}
3083
3084		/**
#	Line 1193 | Line 3088 | public class ConcurrentHashMap<K, V> ext
3088		* @see #keySet()
3089		*/
3090		public Enumeration<K> keys() {
3091	<	return new KeyIterator();
3091	>	return new KeyIterator<K,V>(this);
3092		}
3093
3094		/**
#	Line 1203 | Line 3098 | public class ConcurrentHashMap<K, V> ext
3098		* @see #values()
3099		*/
3100		public Enumeration<V> elements() {
3101	<	return new ValueIterator();
3101	>	return new ValueIterator<K,V>(this);
3102		}
3103
3104	<	/* ---------------- Iterator Support -------------- */
3104	>	/**
3105	>	* Returns a partitionable iterator of the keys in this map.
3106	>	*
3107	>	* @return a partitionable iterator of the keys in this map
3108	>	*/
3109	>	public Spliterator<K> keySpliterator() {
3110	>	return new KeyIterator<K,V>(this);
3111	>	}
3112
3113	<	abstract class HashIterator {
3114	<	int nextSegmentIndex;
3115	<	int nextTableIndex;
3116	<	HashEntry<K,V>[] currentTable;
3117	<	HashEntry<K, V> nextEntry;
3118	<	HashEntry<K, V> lastReturned;
3113	>	/**
3114	>	* Returns a partitionable iterator of the values in this map.
3115	>	*
3116	>	* @return a partitionable iterator of the values in this map
3117	>	*/
3118	>	public Spliterator<V> valueSpliterator() {
3119	>	return new ValueIterator<K,V>(this);
3120	>	}
3121
3122	<	HashIterator() {
3123	<	nextSegmentIndex = segments.length - 1;
3124	<	nextTableIndex = -1;
3125	<	advance();
3126	<	}
3122	>	/**
3123	>	* Returns a partitionable iterator of the entries in this map.
3124	>	*
3125	>	* @return a partitionable iterator of the entries in this map
3126	>	*/
3127	>	public Spliterator<Map.Entry<K,V>> entrySpliterator() {
3128	>	return new EntryIterator<K,V>(this);
3129	>	}
3130
3131	<	/**
3132	<	* Sets nextEntry to first node of next non-empty table
3133	<	* (in backwards order, to simplify checks).
3134	<	*/
3135	<	final void advance() {
3131	>	/**
3132	>	* Returns the hash code value for this {@link Map}, i.e.,
3133	>	* the sum of, for each key-value pair in the map,
3134	>	* {@code key.hashCode() ^ value.hashCode()}.
3135	>	*
3136	>	* @return the hash code value for this map
3137	>	*/
3138	>	public int hashCode() {
3139	>	int h = 0;
3140	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
3141	>	Object v;
3142	>	while ((v = it.advance()) != null) {
3143	>	h += it.nextKey.hashCode() ^ v.hashCode();
3144	>	}
3145	>	return h;
3146	>	}
3147	>
3148	>	/**
3149	>	* Returns a string representation of this map.  The string
3150	>	* representation consists of a list of key-value mappings (in no
3151	>	* particular order) enclosed in braces ("{@code {}}").  Adjacent
3152	>	* mappings are separated by the characters {@code ", "} (comma
3153	>	* and space).  Each key-value mapping is rendered as the key
3154	>	* followed by an equals sign ("{@code =}") followed by the
3155	>	* associated value.
3156	>	*
3157	>	* @return a string representation of this map
3158	>	*/
3159	>	public String toString() {
3160	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
3161	>	StringBuilder sb = new StringBuilder();
3162	>	sb.append('{');
3163	>	Object v;
3164	>	if ((v = it.advance()) != null) {
3165		for (;;) {
3166	<	if (nextTableIndex >= 0) {
3167	<	if ((nextEntry = entryAt(currentTable,
3168	<	nextTableIndex--)) != null)
3169	<	break;
3170	<	}
1235	<	else if (nextSegmentIndex >= 0) {
1236	<	Segment<K,V> seg = segmentAt(segments, nextSegmentIndex--);
1237	<	if (seg != null && (currentTable = seg.table) != null)
1238	<	nextTableIndex = currentTable.length - 1;
1239	<	}
1240	<	else
3166	>	Object k = it.nextKey;
3167	>	sb.append(k == this ? "(this Map)" : k);
3168	>	sb.append('=');
3169	>	sb.append(v == this ? "(this Map)" : v);
3170	>	if ((v = it.advance()) == null)
3171		break;
3172	+	sb.append(',').append(' ');
3173		}
3174		}
3175	+	return sb.append('}').toString();
3176	+	}
3177
3178	<	final HashEntry<K,V> nextEntry() {
3179	<	HashEntry<K,V> e = nextEntry;
3180	<	if (e == null)
3181	<	throw new NoSuchElementException();
3182	<	lastReturned = e; // cannot assign until after null check
3183	<	if ((nextEntry = e.next) == null)
3184	<	advance();
3185	<	return e;
3178	>	/**
3179	>	* Compares the specified object with this map for equality.
3180	>	* Returns {@code true} if the given object is a map with the same
3181	>	* mappings as this map.  This operation may return misleading
3182	>	* results if either map is concurrently modified during execution
3183	>	* of this method.
3184	>	*
3185	>	* @param o object to be compared for equality with this map
3186	>	* @return {@code true} if the specified object is equal to this map
3187	>	*/
3188	>	public boolean equals(Object o) {
3189	>	if (o != this) {
3190	>	if (!(o instanceof Map))
3191	>	return false;
3192	>	Map<?,?> m = (Map<?,?>) o;
3193	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
3194	>	Object val;
3195	>	while ((val = it.advance()) != null) {
3196	>	Object v = m.get(it.nextKey);
3197	>	if (v == null || (v != val && !v.equals(val)))
3198	>	return false;
3199	>	}
3200	>	for (Map.Entry<?,?> e : m.entrySet()) {
3201	>	Object mk, mv, v;
3202	>	if ((mk = e.getKey()) == null ||
3203	>	(mv = e.getValue()) == null ||
3204	>	(v = internalGet(mk)) == null ||
3205	>	(mv != v && !mv.equals(v)))
3206	>	return false;
3207	>	}
3208		}
3209	+	return true;
3210	+	}
3211
3212	<	public final boolean hasNext() { return nextEntry != null; }
1256	<	public final boolean hasMoreElements() { return nextEntry != null; }
3212	>	/* ----------------Iterators -------------- */
3213
3214	<	public final void remove() {
3215	<	if (lastReturned == null)
3214	>	@SuppressWarnings("serial") static final class KeyIterator<K,V> extends Traverser<K,V,Object>
3215	>	implements Spliterator<K>, Enumeration<K> {
3216	>	KeyIterator(ConcurrentHashMap<K, V> map) { super(map); }
3217	>	KeyIterator(ConcurrentHashMap<K, V> map, Traverser<K,V,Object> it) {
3218	>	super(map, it, -1);
3219	>	}
3220	>	public KeyIterator<K,V> split() {
3221	>	if (nextKey != null)
3222		throw new IllegalStateException();
3223	<	ConcurrentHashMap.this.remove(lastReturned.key);
1262	<	lastReturned = null;
3223	>	return new KeyIterator<K,V>(map, this);
3224		}
3225	+	@SuppressWarnings("unchecked") public final K next() {
3226	+	if (nextVal == null && advance() == null)
3227	+	throw new NoSuchElementException();
3228	+	Object k = nextKey;
3229	+	nextVal = null;
3230	+	return (K) k;
3231	+	}
3232	+
3233	+	public final K nextElement() { return next(); }
3234		}
3235
3236	<	final class KeyIterator
3237	<	extends HashIterator
3238	<	implements Iterator<K>, Enumeration<K>
3239	<	{
3240	<	public final K next()        { return super.nextEntry().key; }
3241	<	public final K nextElement() { return super.nextEntry().key; }
3236	>	@SuppressWarnings("serial") static final class ValueIterator<K,V> extends Traverser<K,V,Object>
3237	>	implements Spliterator<V>, Enumeration<V> {
3238	>	ValueIterator(ConcurrentHashMap<K, V> map) { super(map); }
3239	>	ValueIterator(ConcurrentHashMap<K, V> map, Traverser<K,V,Object> it) {
3240	>	super(map, it, -1);
3241	>	}
3242	>	public ValueIterator<K,V> split() {
3243	>	if (nextKey != null)
3244	>	throw new IllegalStateException();
3245	>	return new ValueIterator<K,V>(map, this);
3246	>	}
3247	>
3248	>	@SuppressWarnings("unchecked") public final V next() {
3249	>	Object v;
3250	>	if ((v = nextVal) == null && (v = advance()) == null)
3251	>	throw new NoSuchElementException();
3252	>	nextVal = null;
3253	>	return (V) v;
3254	>	}
3255	>
3256	>	public final V nextElement() { return next(); }
3257		}
3258
3259	<	final class ValueIterator
3260	<	extends HashIterator
3261	<	implements Iterator<V>, Enumeration<V>
3262	<	{
3263	<	public final V next()        { return super.nextEntry().value; }
3264	<	public final V nextElement() { return super.nextEntry().value; }
3259	>	@SuppressWarnings("serial") static final class EntryIterator<K,V> extends Traverser<K,V,Object>
3260	>	implements Spliterator<Map.Entry<K,V>> {
3261	>	EntryIterator(ConcurrentHashMap<K, V> map) { super(map); }
3262	>	EntryIterator(ConcurrentHashMap<K, V> map, Traverser<K,V,Object> it) {
3263	>	super(map, it, -1);
3264	>	}
3265	>	public EntryIterator<K,V> split() {
3266	>	if (nextKey != null)
3267	>	throw new IllegalStateException();
3268	>	return new EntryIterator<K,V>(map, this);
3269	>	}
3270	>
3271	>	@SuppressWarnings("unchecked") public final Map.Entry<K,V> next() {
3272	>	Object v;
3273	>	if ((v = nextVal) == null && (v = advance()) == null)
3274	>	throw new NoSuchElementException();
3275	>	Object k = nextKey;
3276	>	nextVal = null;
3277	>	return new MapEntry<K,V>((K)k, (V)v, map);
3278	>	}
3279		}
3280
3281		/**
3282	<	* Custom Entry class used by EntryIterator.next(), that relays
1284	<	* setValue changes to the underlying map.
3282	>	* Exported Entry for iterators
3283		*/
3284	<	final class WriteThroughEntry
3285	<	extends AbstractMap.SimpleEntry<K,V>
3286	<	{
3287	<	static final long serialVersionUID = 7249069246763182397L;
3288	<
3289	<	WriteThroughEntry(K k, V v) {
3290	<	super(k,v);
3284	>	static final class MapEntry<K,V> implements Map.Entry<K, V> {
3285	>	final K key; // non-null
3286	>	V val;       // non-null
3287	>	final ConcurrentHashMap<K, V> map;
3288	>	MapEntry(K key, V val, ConcurrentHashMap<K, V> map) {
3289	>	this.key = key;
3290	>	this.val = val;
3291	>	this.map = map;
3292	>	}
3293	>	public final K getKey()       { return key; }
3294	>	public final V getValue()     { return val; }
3295	>	public final int hashCode()   { return key.hashCode() ^ val.hashCode(); }
3296	>	public final String toString(){ return key + "=" + val; }
3297	>
3298	>	public final boolean equals(Object o) {
3299	>	Object k, v; Map.Entry<?,?> e;
3300	>	return ((o instanceof Map.Entry) &&
3301	>	(k = (e = (Map.Entry<?,?>)o).getKey()) != null &&
3302	>	(v = e.getValue()) != null &&
3303	>	(k == key || k.equals(key)) &&
3304	>	(v == val || v.equals(val)));
3305		}
3306
3307		/**
3308		* Sets our entry's value and writes through to the map. The
3309	<	* value to return is somewhat arbitrary here. Since a
3310	<	* WriteThroughEntry does not necessarily track asynchronous
3311	<	* changes, the most recent "previous" value could be
3312	<	* different from what we return (or could even have been
3313	<	* removed in which case the put will re-establish). We do not
1302	<	* and cannot guarantee more.
3309	>	* value to return is somewhat arbitrary here. Since we do not
3310	>	* necessarily track asynchronous changes, the most recent
3311	>	* "previous" value could be different from what we return (or
3312	>	* could even have been removed in which case the put will
3313	>	* re-establish). We do not and cannot guarantee more.
3314		*/
3315	<	public V setValue(V value) {
3315	>	public final V setValue(V value) {
3316		if (value == null) throw new NullPointerException();
3317	<	V v = super.setValue(value);
3318	<	ConcurrentHashMap.this.put(getKey(), value);
3317	>	V v = val;
3318	>	val = value;
3319	>	map.put(key, value);
3320		return v;
3321		}
3322		}
3323
3324	<	final class EntryIterator
3325	<	extends HashIterator
3326	<	implements Iterator<Entry<K,V>>
3327	<	{
3328	<	public Map.Entry<K,V> next() {
3329	<	HashEntry<K,V> e = super.nextEntry();
3330	<	return new WriteThroughEntry(e.key, e.value);
3324	>	/**
3325	>	* Returns exportable snapshot entry for the given key and value
3326	>	* when write-through can't or shouldn't be used.
3327	>	*/
3328	>	static <K,V> AbstractMap.SimpleEntry<K,V> entryFor(K k, V v) {
3329	>	return new AbstractMap.SimpleEntry<K,V>(k, v);
3330	>	}
3331	>
3332	>	/* ---------------- Serialization Support -------------- */
3333	>
3334	>	/**
3335	>	* Stripped-down version of helper class used in previous version,
3336	>	* declared for the sake of serialization compatibility
3337	>	*/
3338	>	static class Segment<K,V> implements Serializable {
3339	>	private static final long serialVersionUID = 2249069246763182397L;
3340	>	final float loadFactor;
3341	>	Segment(float lf) { this.loadFactor = lf; }
3342	>	}
3343	>
3344	>	/**
3345	>	* Saves the state of the {@code ConcurrentHashMap} instance to a
3346	>	* stream (i.e., serializes it).
3347	>	* @param s the stream
3348	>	* @serialData
3349	>	* the key (Object) and value (Object)
3350	>	* for each key-value mapping, followed by a null pair.
3351	>	* The key-value mappings are emitted in no particular order.
3352	>	*/
3353	>	@SuppressWarnings("unchecked") private void writeObject(java.io.ObjectOutputStream s)
3354	>	throws java.io.IOException {
3355	>	if (segments == null) { // for serialization compatibility
3356	>	segments = (Segment<K,V>[])
3357	>	new Segment<?,?>[DEFAULT_CONCURRENCY_LEVEL];
3358	>	for (int i = 0; i < segments.length; ++i)
3359	>	segments[i] = new Segment<K,V>(LOAD_FACTOR);
3360	>	}
3361	>	s.defaultWriteObject();
3362	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
3363	>	Object v;
3364	>	while ((v = it.advance()) != null) {
3365	>	s.writeObject(it.nextKey);
3366	>	s.writeObject(v);
3367	>	}
3368	>	s.writeObject(null);
3369	>	s.writeObject(null);
3370	>	segments = null; // throw away
3371	>	}
3372	>
3373	>	/**
3374	>	* Reconstitutes the instance from a stream (that is, deserializes it).
3375	>	* @param s the stream
3376	>	*/
3377	>	@SuppressWarnings("unchecked") private void readObject(java.io.ObjectInputStream s)
3378	>	throws java.io.IOException, ClassNotFoundException {
3379	>	s.defaultReadObject();
3380	>	this.segments = null; // unneeded
3381	>	// initialize transient final field
3382	>	UNSAFE.putObjectVolatile(this, counterOffset, new LongAdder());
3383	>
3384	>	// Create all nodes, then place in table once size is known
3385	>	long size = 0L;
3386	>	Node p = null;
3387	>	for (;;) {
3388	>	K k = (K) s.readObject();
3389	>	V v = (V) s.readObject();
3390	>	if (k != null && v != null) {
3391	>	int h = spread(k.hashCode());
3392	>	p = new Node(h, k, v, p);
3393	>	++size;
3394	>	}
3395	>	else
3396	>	break;
3397	>	}
3398	>	if (p != null) {
3399	>	boolean init = false;
3400	>	int n;
3401	>	if (size >= (long)(MAXIMUM_CAPACITY >>> 1))
3402	>	n = MAXIMUM_CAPACITY;
3403	>	else {
3404	>	int sz = (int)size;
3405	>	n = tableSizeFor(sz + (sz >>> 1) + 1);
3406	>	}
3407	>	int sc = sizeCtl;
3408	>	boolean collide = false;
3409	>	if (n > sc &&
3410	>	UNSAFE.compareAndSwapInt(this, sizeCtlOffset, sc, -1)) {
3411	>	try {
3412	>	if (table == null) {
3413	>	init = true;
3414	>	Node[] tab = new Node[n];
3415	>	int mask = n - 1;
3416	>	while (p != null) {
3417	>	int j = p.hash & mask;
3418	>	Node next = p.next;
3419	>	Node q = p.next = tabAt(tab, j);
3420	>	setTabAt(tab, j, p);
3421	>	if (!collide && q != null && q.hash == p.hash)
3422	>	collide = true;
3423	>	p = next;
3424	>	}
3425	>	table = tab;
3426	>	counter.add(size);
3427	>	sc = n - (n >>> 2);
3428	>	}
3429	>	} finally {
3430	>	sizeCtl = sc;
3431	>	}
3432	>	if (collide) { // rescan and convert to TreeBins
3433	>	Node[] tab = table;
3434	>	for (int i = 0; i < tab.length; ++i) {
3435	>	int c = 0;
3436	>	for (Node e = tabAt(tab, i); e != null; e = e.next) {
3437	>	if (++c > TREE_THRESHOLD &&
3438	>	(e.key instanceof Comparable)) {
3439	>	replaceWithTreeBin(tab, i, e.key);
3440	>	break;
3441	>	}
3442	>	}
3443	>	}
3444	>	}
3445	>	}
3446	>	if (!init) { // Can only happen if unsafely published.
3447	>	while (p != null) {
3448	>	internalPut(p.key, p.val);
3449	>	p = p.next;
3450	>	}
3451	>	}
3452	>	}
3453	>	}
3454	>
3455	>
3456	>	// -------------------------------------------------------
3457	>
3458	>	// Sams
3459	>	/** Interface describing a void action of one argument */
3460	>	public interface Action<A> { void apply(A a); }
3461	>	/** Interface describing a void action of two arguments */
3462	>	public interface BiAction<A,B> { void apply(A a, B b); }
3463	>	/** Interface describing a function of one argument */
3464	>	public interface Fun<A,T> { T apply(A a); }
3465	>	/** Interface describing a function of two arguments */
3466	>	public interface BiFun<A,B,T> { T apply(A a, B b); }
3467	>	/** Interface describing a function of no arguments */
3468	>	public interface Generator<T> { T apply(); }
3469	>	/** Interface describing a function mapping its argument to a double */
3470	>	public interface ObjectToDouble<A> { double apply(A a); }
3471	>	/** Interface describing a function mapping its argument to a long */
3472	>	public interface ObjectToLong<A> { long apply(A a); }
3473	>	/** Interface describing a function mapping its argument to an int */
3474	>	public interface ObjectToInt<A> {int apply(A a); }
3475	>	/** Interface describing a function mapping two arguments to a double */
3476	>	public interface ObjectByObjectToDouble<A,B> { double apply(A a, B b); }
3477	>	/** Interface describing a function mapping two arguments to a long */
3478	>	public interface ObjectByObjectToLong<A,B> { long apply(A a, B b); }
3479	>	/** Interface describing a function mapping two arguments to an int */
3480	>	public interface ObjectByObjectToInt<A,B> {int apply(A a, B b); }
3481	>	/** Interface describing a function mapping a double to a double */
3482	>	public interface DoubleToDouble { double apply(double a); }
3483	>	/** Interface describing a function mapping a long to a long */
3484	>	public interface LongToLong { long apply(long a); }
3485	>	/** Interface describing a function mapping an int to an int */
3486	>	public interface IntToInt { int apply(int a); }
3487	>	/** Interface describing a function mapping two doubles to a double */
3488	>	public interface DoubleByDoubleToDouble { double apply(double a, double b); }
3489	>	/** Interface describing a function mapping two longs to a long */
3490	>	public interface LongByLongToLong { long apply(long a, long b); }
3491	>	/** Interface describing a function mapping two ints to an int */
3492	>	public interface IntByIntToInt { int apply(int a, int b); }
3493	>
3494	>
3495	>	// -------------------------------------------------------
3496	>
3497	>	/**
3498	>	* Performs the given action for each (key, value).
3499	>	*
3500	>	* @param action the action
3501	>	*/
3502	>	public void forEach(BiAction<K,V> action) {
3503	>	ForkJoinTasks.forEach
3504	>	(this, action).invoke();
3505	>	}
3506	>
3507	>	/**
3508	>	* Performs the given action for each non-null transformation
3509	>	* of each (key, value).
3510	>	*
3511	>	* @param transformer a function returning the transformation
3512	>	* for an element, or null of there is no transformation (in
3513	>	* which case the action is not applied).
3514	>	* @param action the action
3515	>	*/
3516	>	public <U> void forEach(BiFun<? super K, ? super V, ? extends U> transformer,
3517	>	Action<U> action) {
3518	>	ForkJoinTasks.forEach
3519	>	(this, transformer, action).invoke();
3520	>	}
3521	>
3522	>	/**
3523	>	* Returns a non-null result from applying the given search
3524	>	* function on each (key, value), or null if none.  Upon
3525	>	* success, further element processing is suppressed and the
3526	>	* results of any other parallel invocations of the search
3527	>	* function are ignored.
3528	>	*
3529	>	* @param searchFunction a function returning a non-null
3530	>	* result on success, else null
3531	>	* @return a non-null result from applying the given search
3532	>	* function on each (key, value), or null if none
3533	>	*/
3534	>	public <U> U search(BiFun<? super K, ? super V, ? extends U> searchFunction) {
3535	>	return ForkJoinTasks.search
3536	>	(this, searchFunction).invoke();
3537	>	}
3538	>
3539	>	/**
3540	>	* Returns the result of accumulating the given transformation
3541	>	* of all (key, value) pairs using the given reducer to
3542	>	* combine values, or null if none.
3543	>	*
3544	>	* @param transformer a function returning the transformation
3545	>	* for an element, or null of there is no transformation (in
3546	>	* which case it is not combined).
3547	>	* @param reducer a commutative associative combining function
3548	>	* @return the result of accumulating the given transformation
3549	>	* of all (key, value) pairs
3550	>	*/
3551	>	public <U> U reduce(BiFun<? super K, ? super V, ? extends U> transformer,
3552	>	BiFun<? super U, ? super U, ? extends U> reducer) {
3553	>	return ForkJoinTasks.reduce
3554	>	(this, transformer, reducer).invoke();
3555	>	}
3556	>
3557	>	/**
3558	>	* Returns the result of accumulating the given transformation
3559	>	* of all (key, value) pairs using the given reducer to
3560	>	* combine values, and the given basis as an identity value.
3561	>	*
3562	>	* @param transformer a function returning the transformation
3563	>	* for an element
3564	>	* @param basis the identity (initial default value) for the reduction
3565	>	* @param reducer a commutative associative combining function
3566	>	* @return the result of accumulating the given transformation
3567	>	* of all (key, value) pairs
3568	>	*/
3569	>	public double reduceToDouble(ObjectByObjectToDouble<? super K, ? super V> transformer,
3570	>	double basis,
3571	>	DoubleByDoubleToDouble reducer) {
3572	>	return ForkJoinTasks.reduceToDouble
3573	>	(this, transformer, basis, reducer).invoke();
3574	>	}
3575	>
3576	>	/**
3577	>	* Returns the result of accumulating the given transformation
3578	>	* of all (key, value) pairs using the given reducer to
3579	>	* combine values, and the given basis as an identity value.
3580	>	*
3581	>	* @param transformer a function returning the transformation
3582	>	* for an element
3583	>	* @param basis the identity (initial default value) for the reduction
3584	>	* @param reducer a commutative associative combining function
3585	>	* @return the result of accumulating the given transformation
3586	>	* of all (key, value) pairs
3587	>	*/
3588	>	public long reduceToLong(ObjectByObjectToLong<? super K, ? super V> transformer,
3589	>	long basis,
3590	>	LongByLongToLong reducer) {
3591	>	return ForkJoinTasks.reduceToLong
3592	>	(this, transformer, basis, reducer).invoke();
3593	>	}
3594	>
3595	>	/**
3596	>	* Returns the result of accumulating the given transformation
3597	>	* of all (key, value) pairs using the given reducer to
3598	>	* combine values, and the given basis as an identity value.
3599	>	*
3600	>	* @param transformer a function returning the transformation
3601	>	* for an element
3602	>	* @param basis the identity (initial default value) for the reduction
3603	>	* @param reducer a commutative associative combining function
3604	>	* @return the result of accumulating the given transformation
3605	>	* of all (key, value) pairs
3606	>	*/
3607	>	public int reduceToInt(ObjectByObjectToInt<? super K, ? super V> transformer,
3608	>	int basis,
3609	>	IntByIntToInt reducer) {
3610	>	return ForkJoinTasks.reduceToInt
3611	>	(this, transformer, basis, reducer).invoke();
3612	>	}
3613	>
3614	>	/**
3615	>	* Performs the given action for each key.
3616	>	*
3617	>	* @param action the action
3618	>	*/
3619	>	public void forEachKey(Action<K> action) {
3620	>	ForkJoinTasks.forEachKey
3621	>	(this, action).invoke();
3622	>	}
3623	>
3624	>	/**
3625	>	* Performs the given action for each non-null transformation
3626	>	* of each key.
3627	>	*
3628	>	* @param transformer a function returning the transformation
3629	>	* for an element, or null of there is no transformation (in
3630	>	* which case the action is not applied).
3631	>	* @param action the action
3632	>	*/
3633	>	public <U> void forEachKey(Fun<? super K, ? extends U> transformer,
3634	>	Action<U> action) {
3635	>	ForkJoinTasks.forEachKey
3636	>	(this, transformer, action).invoke();
3637	>	}
3638	>
3639	>	/**
3640	>	* Returns a non-null result from applying the given search
3641	>	* function on each key, or null if none. Upon success,
3642	>	* further element processing is suppressed and the results of
3643	>	* any other parallel invocations of the search function are
3644	>	* ignored.
3645	>	*
3646	>	* @param searchFunction a function returning a non-null
3647	>	* result on success, else null
3648	>	* @return a non-null result from applying the given search
3649	>	* function on each key, or null if none
3650	>	*/
3651	>	public <U> U searchKeys(Fun<? super K, ? extends U> searchFunction) {
3652	>	return ForkJoinTasks.searchKeys
3653	>	(this, searchFunction).invoke();
3654	>	}
3655	>
3656	>	/**
3657	>	* Returns the result of accumulating all keys using the given
3658	>	* reducer to combine values, or null if none.
3659	>	*
3660	>	* @param reducer a commutative associative combining function
3661	>	* @return the result of accumulating all keys using the given
3662	>	* reducer to combine values, or null if none
3663	>	*/
3664	>	public K reduceKeys(BiFun<? super K, ? super K, ? extends K> reducer) {
3665	>	return ForkJoinTasks.reduceKeys
3666	>	(this, reducer).invoke();
3667	>	}
3668	>
3669	>	/**
3670	>	* Returns the result of accumulating the given transformation
3671	>	* of all keys using the given reducer to combine values, or
3672	>	* null if none.
3673	>	*
3674	>	* @param transformer a function returning the transformation
3675	>	* for an element, or null of there is no transformation (in
3676	>	* which case it is not combined).
3677	>	* @param reducer a commutative associative combining function
3678	>	* @return the result of accumulating the given transformation
3679	>	* of all keys
3680	>	*/
3681	>	public <U> U reduceKeys(Fun<? super K, ? extends U> transformer,
3682	>	BiFun<? super U, ? super U, ? extends U> reducer) {
3683	>	return ForkJoinTasks.reduceKeys
3684	>	(this, transformer, reducer).invoke();
3685	>	}
3686	>
3687	>	/**
3688	>	* Returns the result of accumulating the given transformation
3689	>	* of all keys using the given reducer to combine values, and
3690	>	* the given basis as an identity value.
3691	>	*
3692	>	* @param transformer a function returning the transformation
3693	>	* for an element
3694	>	* @param basis the identity (initial default value) for the reduction
3695	>	* @param reducer a commutative associative combining function
3696	>	* @return  the result of accumulating the given transformation
3697	>	* of all keys
3698	>	*/
3699	>	public double reduceKeysToDouble(ObjectToDouble<? super K> transformer,
3700	>	double basis,
3701	>	DoubleByDoubleToDouble reducer) {
3702	>	return ForkJoinTasks.reduceKeysToDouble
3703	>	(this, transformer, basis, reducer).invoke();
3704	>	}
3705	>
3706	>	/**
3707	>	* Returns the result of accumulating the given transformation
3708	>	* of all keys using the given reducer to combine values, and
3709	>	* the given basis as an identity value.
3710	>	*
3711	>	* @param transformer a function returning the transformation
3712	>	* for an element
3713	>	* @param basis the identity (initial default value) for the reduction
3714	>	* @param reducer a commutative associative combining function
3715	>	* @return the result of accumulating the given transformation
3716	>	* of all keys
3717	>	*/
3718	>	public long reduceKeysToLong(ObjectToLong<? super K> transformer,
3719	>	long basis,
3720	>	LongByLongToLong reducer) {
3721	>	return ForkJoinTasks.reduceKeysToLong
3722	>	(this, transformer, basis, reducer).invoke();
3723	>	}
3724	>
3725	>	/**
3726	>	* Returns the result of accumulating the given transformation
3727	>	* of all keys using the given reducer to combine values, and
3728	>	* the given basis as an identity value.
3729	>	*
3730	>	* @param transformer a function returning the transformation
3731	>	* for an element
3732	>	* @param basis the identity (initial default value) for the reduction
3733	>	* @param reducer a commutative associative combining function
3734	>	* @return the result of accumulating the given transformation
3735	>	* of all keys
3736	>	*/
3737	>	public int reduceKeysToInt(ObjectToInt<? super K> transformer,
3738	>	int basis,
3739	>	IntByIntToInt reducer) {
3740	>	return ForkJoinTasks.reduceKeysToInt
3741	>	(this, transformer, basis, reducer).invoke();
3742	>	}
3743	>
3744	>	/**
3745	>	* Performs the given action for each value.
3746	>	*
3747	>	* @param action the action
3748	>	*/
3749	>	public void forEachValue(Action<V> action) {
3750	>	ForkJoinTasks.forEachValue
3751	>	(this, action).invoke();
3752	>	}
3753	>
3754	>	/**
3755	>	* Performs the given action for each non-null transformation
3756	>	* of each value.
3757	>	*
3758	>	* @param transformer a function returning the transformation
3759	>	* for an element, or null of there is no transformation (in
3760	>	* which case the action is not applied).
3761	>	*/
3762	>	public <U> void forEachValue(Fun<? super V, ? extends U> transformer,
3763	>	Action<U> action) {
3764	>	ForkJoinTasks.forEachValue
3765	>	(this, transformer, action).invoke();
3766	>	}
3767	>
3768	>	/**
3769	>	* Returns a non-null result from applying the given search
3770	>	* function on each value, or null if none.  Upon success,
3771	>	* further element processing is suppressed and the results of
3772	>	* any other parallel invocations of the search function are
3773	>	* ignored.
3774	>	*
3775	>	* @param searchFunction a function returning a non-null
3776	>	* result on success, else null
3777	>	* @return a non-null result from applying the given search
3778	>	* function on each value, or null if none
3779	>	*
3780	>	*/
3781	>	public <U> U searchValues(Fun<? super V, ? extends U> searchFunction) {
3782	>	return ForkJoinTasks.searchValues
3783	>	(this, searchFunction).invoke();
3784	>	}
3785	>
3786	>	/**
3787	>	* Returns the result of accumulating all values using the
3788	>	* given reducer to combine values, or null if none.
3789	>	*
3790	>	* @param reducer a commutative associative combining function
3791	>	* @return  the result of accumulating all values
3792	>	*/
3793	>	public V reduceValues(BiFun<? super V, ? super V, ? extends V> reducer) {
3794	>	return ForkJoinTasks.reduceValues
3795	>	(this, reducer).invoke();
3796	>	}
3797	>
3798	>	/**
3799	>	* Returns the result of accumulating the given transformation
3800	>	* of all values using the given reducer to combine values, or
3801	>	* null if none.
3802	>	*
3803	>	* @param transformer a function returning the transformation
3804	>	* for an element, or null of there is no transformation (in
3805	>	* which case it is not combined).
3806	>	* @param reducer a commutative associative combining function
3807	>	* @return the result of accumulating the given transformation
3808	>	* of all values
3809	>	*/
3810	>	public <U> U reduceValues(Fun<? super V, ? extends U> transformer,
3811	>	BiFun<? super U, ? super U, ? extends U> reducer) {
3812	>	return ForkJoinTasks.reduceValues
3813	>	(this, transformer, reducer).invoke();
3814	>	}
3815	>
3816	>	/**
3817	>	* Returns the result of accumulating the given transformation
3818	>	* of all values using the given reducer to combine values,
3819	>	* and the given basis as an identity value.
3820	>	*
3821	>	* @param transformer a function returning the transformation
3822	>	* for an element
3823	>	* @param basis the identity (initial default value) for the reduction
3824	>	* @param reducer a commutative associative combining function
3825	>	* @return the result of accumulating the given transformation
3826	>	* of all values
3827	>	*/
3828	>	public double reduceValuesToDouble(ObjectToDouble<? super V> transformer,
3829	>	double basis,
3830	>	DoubleByDoubleToDouble reducer) {
3831	>	return ForkJoinTasks.reduceValuesToDouble
3832	>	(this, transformer, basis, reducer).invoke();
3833	>	}
3834	>
3835	>	/**
3836	>	* Returns the result of accumulating the given transformation
3837	>	* of all values using the given reducer to combine values,
3838	>	* and the given basis as an identity value.
3839	>	*
3840	>	* @param transformer a function returning the transformation
3841	>	* for an element
3842	>	* @param basis the identity (initial default value) for the reduction
3843	>	* @param reducer a commutative associative combining function
3844	>	* @return the result of accumulating the given transformation
3845	>	* of all values
3846	>	*/
3847	>	public long reduceValuesToLong(ObjectToLong<? super V> transformer,
3848	>	long basis,
3849	>	LongByLongToLong reducer) {
3850	>	return ForkJoinTasks.reduceValuesToLong
3851	>	(this, transformer, basis, reducer).invoke();
3852	>	}
3853	>
3854	>	/**
3855	>	* Returns the result of accumulating the given transformation
3856	>	* of all values using the given reducer to combine values,
3857	>	* and the given basis as an identity value.
3858	>	*
3859	>	* @param transformer a function returning the transformation
3860	>	* for an element
3861	>	* @param basis the identity (initial default value) for the reduction
3862	>	* @param reducer a commutative associative combining function
3863	>	* @return the result of accumulating the given transformation
3864	>	* of all values
3865	>	*/
3866	>	public int reduceValuesToInt(ObjectToInt<? super V> transformer,
3867	>	int basis,
3868	>	IntByIntToInt reducer) {
3869	>	return ForkJoinTasks.reduceValuesToInt
3870	>	(this, transformer, basis, reducer).invoke();
3871	>	}
3872	>
3873	>	/**
3874	>	* Performs the given action for each entry.
3875	>	*
3876	>	* @param action the action
3877	>	*/
3878	>	public void forEachEntry(Action<Map.Entry<K,V>> action) {
3879	>	ForkJoinTasks.forEachEntry
3880	>	(this, action).invoke();
3881	>	}
3882	>
3883	>	/**
3884	>	* Performs the given action for each non-null transformation
3885	>	* of each entry.
3886	>	*
3887	>	* @param transformer a function returning the transformation
3888	>	* for an element, or null of there is no transformation (in
3889	>	* which case the action is not applied).
3890	>	* @param action the action
3891	>	*/
3892	>	public <U> void forEachEntry(Fun<Map.Entry<K,V>, ? extends U> transformer,
3893	>	Action<U> action) {
3894	>	ForkJoinTasks.forEachEntry
3895	>	(this, transformer, action).invoke();
3896	>	}
3897	>
3898	>	/**
3899	>	* Returns a non-null result from applying the given search
3900	>	* function on each entry, or null if none.  Upon success,
3901	>	* further element processing is suppressed and the results of
3902	>	* any other parallel invocations of the search function are
3903	>	* ignored.
3904	>	*
3905	>	* @param searchFunction a function returning a non-null
3906	>	* result on success, else null
3907	>	* @return a non-null result from applying the given search
3908	>	* function on each entry, or null if none
3909	>	*/
3910	>	public <U> U searchEntries(Fun<Map.Entry<K,V>, ? extends U> searchFunction) {
3911	>	return ForkJoinTasks.searchEntries
3912	>	(this, searchFunction).invoke();
3913	>	}
3914	>
3915	>	/**
3916	>	* Returns the result of accumulating all entries using the
3917	>	* given reducer to combine values, or null if none.
3918	>	*
3919	>	* @param reducer a commutative associative combining function
3920	>	* @return the result of accumulating all entries
3921	>	*/
3922	>	public Map.Entry<K,V> reduceEntries(BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer) {
3923	>	return ForkJoinTasks.reduceEntries
3924	>	(this, reducer).invoke();
3925	>	}
3926	>
3927	>	/**
3928	>	* Returns the result of accumulating the given transformation
3929	>	* of all entries using the given reducer to combine values,
3930	>	* or null if none.
3931	>	*
3932	>	* @param transformer a function returning the transformation
3933	>	* for an element, or null of there is no transformation (in
3934	>	* which case it is not combined).
3935	>	* @param reducer a commutative associative combining function
3936	>	* @return the result of accumulating the given transformation
3937	>	* of all entries
3938	>	*/
3939	>	public <U> U reduceEntries(Fun<Map.Entry<K,V>, ? extends U> transformer,
3940	>	BiFun<? super U, ? super U, ? extends U> reducer) {
3941	>	return ForkJoinTasks.reduceEntries
3942	>	(this, transformer, reducer).invoke();
3943	>	}
3944	>
3945	>	/**
3946	>	* Returns the result of accumulating the given transformation
3947	>	* of all entries using the given reducer to combine values,
3948	>	* and the given basis as an identity value.
3949	>	*
3950	>	* @param transformer a function returning the transformation
3951	>	* for an element
3952	>	* @param basis the identity (initial default value) for the reduction
3953	>	* @param reducer a commutative associative combining function
3954	>	* @return the result of accumulating the given transformation
3955	>	* of all entries
3956	>	*/
3957	>	public double reduceEntriesToDouble(ObjectToDouble<Map.Entry<K,V>> transformer,
3958	>	double basis,
3959	>	DoubleByDoubleToDouble reducer) {
3960	>	return ForkJoinTasks.reduceEntriesToDouble
3961	>	(this, transformer, basis, reducer).invoke();
3962	>	}
3963	>
3964	>	/**
3965	>	* Returns the result of accumulating the given transformation
3966	>	* of all entries using the given reducer to combine values,
3967	>	* and the given basis as an identity value.
3968	>	*
3969	>	* @param transformer a function returning the transformation
3970	>	* for an element
3971	>	* @param basis the identity (initial default value) for the reduction
3972	>	* @param reducer a commutative associative combining function
3973	>	* @return  the result of accumulating the given transformation
3974	>	* of all entries
3975	>	*/
3976	>	public long reduceEntriesToLong(ObjectToLong<Map.Entry<K,V>> transformer,
3977	>	long basis,
3978	>	LongByLongToLong reducer) {
3979	>	return ForkJoinTasks.reduceEntriesToLong
3980	>	(this, transformer, basis, reducer).invoke();
3981	>	}
3982	>
3983	>	/**
3984	>	* Returns the result of accumulating the given transformation
3985	>	* of all entries using the given reducer to combine values,
3986	>	* and the given basis as an identity value.
3987	>	*
3988	>	* @param transformer a function returning the transformation
3989	>	* for an element
3990	>	* @param basis the identity (initial default value) for the reduction
3991	>	* @param reducer a commutative associative combining function
3992	>	* @return the result of accumulating the given transformation
3993	>	* of all entries
3994	>	*/
3995	>	public int reduceEntriesToInt(ObjectToInt<Map.Entry<K,V>> transformer,
3996	>	int basis,
3997	>	IntByIntToInt reducer) {
3998	>	return ForkJoinTasks.reduceEntriesToInt
3999	>	(this, transformer, basis, reducer).invoke();
4000	>	}
4001	>
4002	>	/* ----------------Views -------------- */
4003	>
4004	>	/**
4005	>	* Base class for views.
4006	>	*/
4007	>	static abstract class CHMView<K, V> {
4008	>	final ConcurrentHashMap<K, V> map;
4009	>	CHMView(ConcurrentHashMap<K, V> map)  { this.map = map; }
4010	>
4011	>	/**
4012	>	* Returns the map backing this view.
4013	>	*
4014	>	* @return the map backing this view
4015	>	*/
4016	>	public ConcurrentHashMap<K,V> getMap() { return map; }
4017	>
4018	>	public final int size()                 { return map.size(); }
4019	>	public final boolean isEmpty()          { return map.isEmpty(); }
4020	>	public final void clear()               { map.clear(); }
4021	>
4022	>	// implementations below rely on concrete classes supplying these
4023	>	abstract public Iterator<?> iterator();
4024	>	abstract public boolean contains(Object o);
4025	>	abstract public boolean remove(Object o);
4026	>
4027	>	private static final String oomeMsg = "Required array size too large";
4028	>
4029	>	public final Object[] toArray() {
4030	>	long sz = map.mappingCount();
4031	>	if (sz > (long)(MAX_ARRAY_SIZE))
4032	>	throw new OutOfMemoryError(oomeMsg);
4033	>	int n = (int)sz;
4034	>	Object[] r = new Object[n];
4035	>	int i = 0;
4036	>	Iterator<?> it = iterator();
4037	>	while (it.hasNext()) {
4038	>	if (i == n) {
4039	>	if (n >= MAX_ARRAY_SIZE)
4040	>	throw new OutOfMemoryError(oomeMsg);
4041	>	if (n >= MAX_ARRAY_SIZE - (MAX_ARRAY_SIZE >>> 1) - 1)
4042	>	n = MAX_ARRAY_SIZE;
4043	>	else
4044	>	n += (n >>> 1) + 1;
4045	>	r = Arrays.copyOf(r, n);
4046	>	}
4047	>	r[i++] = it.next();
4048	>	}
4049	>	return (i == n) ? r : Arrays.copyOf(r, i);
4050	>	}
4051	>
4052	>	@SuppressWarnings("unchecked") public final <T> T[] toArray(T[] a) {
4053	>	long sz = map.mappingCount();
4054	>	if (sz > (long)(MAX_ARRAY_SIZE))
4055	>	throw new OutOfMemoryError(oomeMsg);
4056	>	int m = (int)sz;
4057	>	T[] r = (a.length >= m) ? a :
4058	>	(T[])java.lang.reflect.Array
4059	>	.newInstance(a.getClass().getComponentType(), m);
4060	>	int n = r.length;
4061	>	int i = 0;
4062	>	Iterator<?> it = iterator();
4063	>	while (it.hasNext()) {
4064	>	if (i == n) {
4065	>	if (n >= MAX_ARRAY_SIZE)
4066	>	throw new OutOfMemoryError(oomeMsg);
4067	>	if (n >= MAX_ARRAY_SIZE - (MAX_ARRAY_SIZE >>> 1) - 1)
4068	>	n = MAX_ARRAY_SIZE;
4069	>	else
4070	>	n += (n >>> 1) + 1;
4071	>	r = Arrays.copyOf(r, n);
4072	>	}
4073	>	r[i++] = (T)it.next();
4074	>	}
4075	>	if (a == r && i < n) {
4076	>	r[i] = null; // null-terminate
4077	>	return r;
4078	>	}
4079	>	return (i == n) ? r : Arrays.copyOf(r, i);
4080	>	}
4081	>
4082	>	public final int hashCode() {
4083	>	int h = 0;
4084	>	for (Iterator<?> it = iterator(); it.hasNext();)
4085	>	h += it.next().hashCode();
4086	>	return h;
4087	>	}
4088	>
4089	>	public final String toString() {
4090	>	StringBuilder sb = new StringBuilder();
4091	>	sb.append('[');
4092	>	Iterator<?> it = iterator();
4093	>	if (it.hasNext()) {
4094	>	for (;;) {
4095	>	Object e = it.next();
4096	>	sb.append(e == this ? "(this Collection)" : e);
4097	>	if (!it.hasNext())
4098	>	break;
4099	>	sb.append(',').append(' ');
4100	>	}
4101	>	}
4102	>	return sb.append(']').toString();
4103	>	}
4104	>
4105	>	public final boolean containsAll(Collection<?> c) {
4106	>	if (c != this) {
4107	>	for (Iterator<?> it = c.iterator(); it.hasNext();) {
4108	>	Object e = it.next();
4109	>	if (e == null || !contains(e))
4110	>	return false;
4111	>	}
4112	>	}
4113	>	return true;
4114	>	}
4115	>
4116	>	public final boolean removeAll(Collection<?> c) {
4117	>	boolean modified = false;
4118	>	for (Iterator<?> it = iterator(); it.hasNext();) {
4119	>	if (c.contains(it.next())) {
4120	>	it.remove();
4121	>	modified = true;
4122	>	}
4123	>	}
4124	>	return modified;
4125	>	}
4126	>
4127	>	public final boolean retainAll(Collection<?> c) {
4128	>	boolean modified = false;
4129	>	for (Iterator<?> it = iterator(); it.hasNext();) {
4130	>	if (!c.contains(it.next())) {
4131	>	it.remove();
4132	>	modified = true;
4133	>	}
4134	>	}
4135	>	return modified;
4136		}
4137	+
4138		}
4139
4140	<	final class KeySet extends AbstractSet<K> {
4141	<	public Iterator<K> iterator() {
4142	<	return new KeyIterator();
4140	>	/**
4141	>	* A view of a ConcurrentHashMap as a {@link Set} of keys, in
4142	>	* which additions may optionally be enabled by mapping to a
4143	>	* common value.  This class cannot be directly instantiated. See
4144	>	* {@link #keySet}, {@link #keySet(Object)}, {@link #newKeySet()},
4145	>	* {@link #newKeySet(int)}.
4146	>	*/
4147	>	public static class KeySetView<K,V> extends CHMView<K,V> implements Set<K>, java.io.Serializable {
4148	>	private static final long serialVersionUID = 7249069246763182397L;
4149	>	private final V value;
4150	>	KeySetView(ConcurrentHashMap<K, V> map, V value) {  // non-public
4151	>	super(map);
4152	>	this.value = value;
4153	>	}
4154	>
4155	>	/**
4156	>	* Returns the default mapped value for additions,
4157	>	* or {@code null} if additions are not supported.
4158	>	*
4159	>	* @return the default mapped value for additions, or {@code null}
4160	>	* if not supported.
4161	>	*/
4162	>	public V getMappedValue() { return value; }
4163	>
4164	>	// implement Set API
4165	>
4166	>	public boolean contains(Object o) { return map.containsKey(o); }
4167	>	public boolean remove(Object o)   { return map.remove(o) != null; }
4168	>
4169	>	/**
4170	>	* Returns a "weakly consistent" iterator that will never
4171	>	* throw {@link ConcurrentModificationException}, and
4172	>	* guarantees to traverse elements as they existed upon
4173	>	* construction of the iterator, and may (but is not
4174	>	* guaranteed to) reflect any modifications subsequent to
4175	>	* construction.
4176	>	*
4177	>	* @return an iterator over the keys of this map
4178	>	*/
4179	>	public Iterator<K> iterator()     { return new KeyIterator<K,V>(map); }
4180	>	public boolean add(K e) {
4181	>	V v;
4182	>	if ((v = value) == null)
4183	>	throw new UnsupportedOperationException();
4184	>	if (e == null)
4185	>	throw new NullPointerException();
4186	>	return map.internalPutIfAbsent(e, v) == null;
4187	>	}
4188	>	public boolean addAll(Collection<? extends K> c) {
4189	>	boolean added = false;
4190	>	V v;
4191	>	if ((v = value) == null)
4192	>	throw new UnsupportedOperationException();
4193	>	for (K e : c) {
4194	>	if (e == null)
4195	>	throw new NullPointerException();
4196	>	if (map.internalPutIfAbsent(e, v) == null)
4197	>	added = true;
4198	>	}
4199	>	return added;
4200	>	}
4201	>	public boolean equals(Object o) {
4202	>	Set<?> c;
4203	>	return ((o instanceof Set) &&
4204	>	((c = (Set<?>)o) == this ||
4205	>	(containsAll(c) && c.containsAll(this))));
4206	>	}
4207	>
4208	>	/**
4209	>	* Performs the given action for each key.
4210	>	*
4211	>	* @param action the action
4212	>	*/
4213	>	public void forEach(Action<K> action) {
4214	>	ForkJoinTasks.forEachKey
4215	>	(map, action).invoke();
4216		}
4217	<	public int size() {
4218	<	return ConcurrentHashMap.this.size();
4217	>
4218	>	/**
4219	>	* Performs the given action for each non-null transformation
4220	>	* of each key.
4221	>	*
4222	>	* @param transformer a function returning the transformation
4223	>	* for an element, or null of there is no transformation (in
4224	>	* which case the action is not applied).
4225	>	* @param action the action
4226	>	*/
4227	>	public <U> void forEach(Fun<? super K, ? extends U> transformer,
4228	>	Action<U> action) {
4229	>	ForkJoinTasks.forEachKey
4230	>	(map, transformer, action).invoke();
4231		}
4232	<	public boolean isEmpty() {
4233	<	return ConcurrentHashMap.this.isEmpty();
4232	>
4233	>	/**
4234	>	* Returns a non-null result from applying the given search
4235	>	* function on each key, or null if none. Upon success,
4236	>	* further element processing is suppressed and the results of
4237	>	* any other parallel invocations of the search function are
4238	>	* ignored.
4239	>	*
4240	>	* @param searchFunction a function returning a non-null
4241	>	* result on success, else null
4242	>	* @return a non-null result from applying the given search
4243	>	* function on each key, or null if none
4244	>	*/
4245	>	public <U> U search(Fun<? super K, ? extends U> searchFunction) {
4246	>	return ForkJoinTasks.searchKeys
4247	>	(map, searchFunction).invoke();
4248	>	}
4249	>
4250	>	/**
4251	>	* Returns the result of accumulating all keys using the given
4252	>	* reducer to combine values, or null if none.
4253	>	*
4254	>	* @param reducer a commutative associative combining function
4255	>	* @return the result of accumulating all keys using the given
4256	>	* reducer to combine values, or null if none
4257	>	*/
4258	>	public K reduce(BiFun<? super K, ? super K, ? extends K> reducer) {
4259	>	return ForkJoinTasks.reduceKeys
4260	>	(map, reducer).invoke();
4261		}
4262	<	public boolean contains(Object o) {
4263	<	return ConcurrentHashMap.this.containsKey(o);
4262	>
4263	>	/**
4264	>	* Returns the result of accumulating the given transformation
4265	>	* of all keys using the given reducer to combine values, and
4266	>	* the given basis as an identity value.
4267	>	*
4268	>	* @param transformer a function returning the transformation
4269	>	* for an element
4270	>	* @param basis the identity (initial default value) for the reduction
4271	>	* @param reducer a commutative associative combining function
4272	>	* @return  the result of accumulating the given transformation
4273	>	* of all keys
4274	>	*/
4275	>	public double reduceToDouble(ObjectToDouble<? super K> transformer,
4276	>	double basis,
4277	>	DoubleByDoubleToDouble reducer) {
4278	>	return ForkJoinTasks.reduceKeysToDouble
4279	>	(map, transformer, basis, reducer).invoke();
4280		}
4281	<	public boolean remove(Object o) {
4282	<	return ConcurrentHashMap.this.remove(o) != null;
4281	>
4282	>
4283	>	/**
4284	>	* Returns the result of accumulating the given transformation
4285	>	* of all keys using the given reducer to combine values, and
4286	>	* the given basis as an identity value.
4287	>	*
4288	>	* @param transformer a function returning the transformation
4289	>	* for an element
4290	>	* @param basis the identity (initial default value) for the reduction
4291	>	* @param reducer a commutative associative combining function
4292	>	* @return the result of accumulating the given transformation
4293	>	* of all keys
4294	>	*/
4295	>	public long reduceToLong(ObjectToLong<? super K> transformer,
4296	>	long basis,
4297	>	LongByLongToLong reducer) {
4298	>	return ForkJoinTasks.reduceKeysToLong
4299	>	(map, transformer, basis, reducer).invoke();
4300		}
4301	<	public void clear() {
4302	<	ConcurrentHashMap.this.clear();
4301	>
4302	>	/**
4303	>	* Returns the result of accumulating the given transformation
4304	>	* of all keys using the given reducer to combine values, and
4305	>	* the given basis as an identity value.
4306	>	*
4307	>	* @param transformer a function returning the transformation
4308	>	* for an element
4309	>	* @param basis the identity (initial default value) for the reduction
4310	>	* @param reducer a commutative associative combining function
4311	>	* @return the result of accumulating the given transformation
4312	>	* of all keys
4313	>	*/
4314	>	public int reduceToInt(ObjectToInt<? super K> transformer,
4315	>	int basis,
4316	>	IntByIntToInt reducer) {
4317	>	return ForkJoinTasks.reduceKeysToInt
4318	>	(map, transformer, basis, reducer).invoke();
4319		}
4320	+
4321		}
4322
4323	<	final class Values extends AbstractCollection<V> {
4324	<	public Iterator<V> iterator() {
4325	<	return new ValueIterator();
4323	>	/**
4324	>	* A view of a ConcurrentHashMap as a {@link Collection} of
4325	>	* values, in which additions are disabled. This class cannot be
4326	>	* directly instantiated. See {@link #values},
4327	>	*
4328	>	* <p>The view's {@code iterator} is a "weakly consistent" iterator
4329	>	* that will never throw {@link ConcurrentModificationException},
4330	>	* and guarantees to traverse elements as they existed upon
4331	>	* construction of the iterator, and may (but is not guaranteed to)
4332	>	* reflect any modifications subsequent to construction.
4333	>	*/
4334	>	public static final class ValuesView<K,V> extends CHMView<K,V>
4335	>	implements Collection<V> {
4336	>	ValuesView(ConcurrentHashMap<K, V> map)   { super(map); }
4337	>	public final boolean contains(Object o) { return map.containsValue(o); }
4338	>	public final boolean remove(Object o) {
4339	>	if (o != null) {
4340	>	Iterator<V> it = new ValueIterator<K,V>(map);
4341	>	while (it.hasNext()) {
4342	>	if (o.equals(it.next())) {
4343	>	it.remove();
4344	>	return true;
4345	>	}
4346	>	}
4347	>	}
4348	>	return false;
4349	>	}
4350	>
4351	>	/**
4352	>	* Returns a "weakly consistent" iterator that will never
4353	>	* throw {@link ConcurrentModificationException}, and
4354	>	* guarantees to traverse elements as they existed upon
4355	>	* construction of the iterator, and may (but is not
4356	>	* guaranteed to) reflect any modifications subsequent to
4357	>	* construction.
4358	>	*
4359	>	* @return an iterator over the values of this map
4360	>	*/
4361	>	public final Iterator<V> iterator() {
4362	>	return new ValueIterator<K,V>(map);
4363	>	}
4364	>	public final boolean add(V e) {
4365	>	throw new UnsupportedOperationException();
4366	>	}
4367	>	public final boolean addAll(Collection<? extends V> c) {
4368	>	throw new UnsupportedOperationException();
4369	>	}
4370	>
4371	>	/**
4372	>	* Performs the given action for each value.
4373	>	*
4374	>	* @param action the action
4375	>	*/
4376	>	public void forEach(Action<V> action) {
4377	>	ForkJoinTasks.forEachValue
4378	>	(map, action).invoke();
4379	>	}
4380	>
4381	>	/**
4382	>	* Performs the given action for each non-null transformation
4383	>	* of each value.
4384	>	*
4385	>	* @param transformer a function returning the transformation
4386	>	* for an element, or null of there is no transformation (in
4387	>	* which case the action is not applied).
4388	>	*/
4389	>	public <U> void forEach(Fun<? super V, ? extends U> transformer,
4390	>	Action<U> action) {
4391	>	ForkJoinTasks.forEachValue
4392	>	(map, transformer, action).invoke();
4393	>	}
4394	>
4395	>	/**
4396	>	* Returns a non-null result from applying the given search
4397	>	* function on each value, or null if none.  Upon success,
4398	>	* further element processing is suppressed and the results of
4399	>	* any other parallel invocations of the search function are
4400	>	* ignored.
4401	>	*
4402	>	* @param searchFunction a function returning a non-null
4403	>	* result on success, else null
4404	>	* @return a non-null result from applying the given search
4405	>	* function on each value, or null if none
4406	>	*
4407	>	*/
4408	>	public <U> U search(Fun<? super V, ? extends U> searchFunction) {
4409	>	return ForkJoinTasks.searchValues
4410	>	(map, searchFunction).invoke();
4411		}
4412	<	public int size() {
4413	<	return ConcurrentHashMap.this.size();
4412	>
4413	>	/**
4414	>	* Returns the result of accumulating all values using the
4415	>	* given reducer to combine values, or null if none.
4416	>	*
4417	>	* @param reducer a commutative associative combining function
4418	>	* @return  the result of accumulating all values
4419	>	*/
4420	>	public V reduce(BiFun<? super V, ? super V, ? extends V> reducer) {
4421	>	return ForkJoinTasks.reduceValues
4422	>	(map, reducer).invoke();
4423		}
4424	<	public boolean isEmpty() {
4425	<	return ConcurrentHashMap.this.isEmpty();
4424	>
4425	>	/**
4426	>	* Returns the result of accumulating the given transformation
4427	>	* of all values using the given reducer to combine values, or
4428	>	* null if none.
4429	>	*
4430	>	* @param transformer a function returning the transformation
4431	>	* for an element, or null of there is no transformation (in
4432	>	* which case it is not combined).
4433	>	* @param reducer a commutative associative combining function
4434	>	* @return the result of accumulating the given transformation
4435	>	* of all values
4436	>	*/
4437	>	public <U> U reduce(Fun<? super V, ? extends U> transformer,
4438	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4439	>	return ForkJoinTasks.reduceValues
4440	>	(map, transformer, reducer).invoke();
4441	>	}
4442	>
4443	>	/**
4444	>	* Returns the result of accumulating the given transformation
4445	>	* of all values using the given reducer to combine values,
4446	>	* and the given basis as an identity value.
4447	>	*
4448	>	* @param transformer a function returning the transformation
4449	>	* for an element
4450	>	* @param basis the identity (initial default value) for the reduction
4451	>	* @param reducer a commutative associative combining function
4452	>	* @return the result of accumulating the given transformation
4453	>	* of all values
4454	>	*/
4455	>	public double reduceToDouble(ObjectToDouble<? super V> transformer,
4456	>	double basis,
4457	>	DoubleByDoubleToDouble reducer) {
4458	>	return ForkJoinTasks.reduceValuesToDouble
4459	>	(map, transformer, basis, reducer).invoke();
4460		}
4461	<	public boolean contains(Object o) {
4462	<	return ConcurrentHashMap.this.containsValue(o);
4461	>
4462	>	/**
4463	>	* Returns the result of accumulating the given transformation
4464	>	* of all values using the given reducer to combine values,
4465	>	* and the given basis as an identity value.
4466	>	*
4467	>	* @param transformer a function returning the transformation
4468	>	* for an element
4469	>	* @param basis the identity (initial default value) for the reduction
4470	>	* @param reducer a commutative associative combining function
4471	>	* @return the result of accumulating the given transformation
4472	>	* of all values
4473	>	*/
4474	>	public long reduceToLong(ObjectToLong<? super V> transformer,
4475	>	long basis,
4476	>	LongByLongToLong reducer) {
4477	>	return ForkJoinTasks.reduceValuesToLong
4478	>	(map, transformer, basis, reducer).invoke();
4479		}
4480	<	public void clear() {
4481	<	ConcurrentHashMap.this.clear();
4480	>
4481	>	/**
4482	>	* Returns the result of accumulating the given transformation
4483	>	* of all values using the given reducer to combine values,
4484	>	* and the given basis as an identity value.
4485	>	*
4486	>	* @param transformer a function returning the transformation
4487	>	* for an element
4488	>	* @param basis the identity (initial default value) for the reduction
4489	>	* @param reducer a commutative associative combining function
4490	>	* @return the result of accumulating the given transformation
4491	>	* of all values
4492	>	*/
4493	>	public int reduceToInt(ObjectToInt<? super V> transformer,
4494	>	int basis,
4495	>	IntByIntToInt reducer) {
4496	>	return ForkJoinTasks.reduceValuesToInt
4497	>	(map, transformer, basis, reducer).invoke();
4498		}
4499	+
4500		}
4501
4502	<	final class EntrySet extends AbstractSet<Map.Entry<K,V>> {
4503	<	public Iterator<Map.Entry<K,V>> iterator() {
4504	<	return new EntryIterator();
4502	>	/**
4503	>	* A view of a ConcurrentHashMap as a {@link Set} of (key, value)
4504	>	* entries.  This class cannot be directly instantiated. See
4505	>	* {@link #entrySet}.
4506	>	*/
4507	>	public static final class EntrySetView<K,V> extends CHMView<K,V>
4508	>	implements Set<Map.Entry<K,V>> {
4509	>	EntrySetView(ConcurrentHashMap<K, V> map) { super(map); }
4510	>	public final boolean contains(Object o) {
4511	>	Object k, v, r; Map.Entry<?,?> e;
4512	>	return ((o instanceof Map.Entry) &&
4513	>	(k = (e = (Map.Entry<?,?>)o).getKey()) != null &&
4514	>	(r = map.get(k)) != null &&
4515	>	(v = e.getValue()) != null &&
4516	>	(v == r || v.equals(r)));
4517	>	}
4518	>	public final boolean remove(Object o) {
4519	>	Object k, v; Map.Entry<?,?> e;
4520	>	return ((o instanceof Map.Entry) &&
4521	>	(k = (e = (Map.Entry<?,?>)o).getKey()) != null &&
4522	>	(v = e.getValue()) != null &&
4523	>	map.remove(k, v));
4524		}
4525	<	public boolean contains(Object o) {
4526	<	if (!(o instanceof Map.Entry))
4527	<	return false;
4528	<	Map.Entry<?,?> e = (Map.Entry<?,?>)o;
4529	<	V v = ConcurrentHashMap.this.get(e.getKey());
4530	<	return v != null && v.equals(e.getValue());
4525	>
4526	>	/**
4527	>	* Returns a "weakly consistent" iterator that will never
4528	>	* throw {@link ConcurrentModificationException}, and
4529	>	* guarantees to traverse elements as they existed upon
4530	>	* construction of the iterator, and may (but is not
4531	>	* guaranteed to) reflect any modifications subsequent to
4532	>	* construction.
4533	>	*
4534	>	* @return an iterator over the entries of this map
4535	>	*/
4536	>	public final Iterator<Map.Entry<K,V>> iterator() {
4537	>	return new EntryIterator<K,V>(map);
4538		}
4539	<	public boolean remove(Object o) {
4540	<	if (!(o instanceof Map.Entry))
4541	<	return false;
4542	<	Map.Entry<?,?> e = (Map.Entry<?,?>)o;
4543	<	return ConcurrentHashMap.this.remove(e.getKey(), e.getValue());
4539	>
4540	>	public final boolean add(Entry<K,V> e) {
4541	>	K key = e.getKey();
4542	>	V value = e.getValue();
4543	>	if (key == null || value == null)
4544	>	throw new NullPointerException();
4545	>	return map.internalPut(key, value) == null;
4546	>	}
4547	>	public final boolean addAll(Collection<? extends Entry<K,V>> c) {
4548	>	boolean added = false;
4549	>	for (Entry<K,V> e : c) {
4550	>	if (add(e))
4551	>	added = true;
4552	>	}
4553	>	return added;
4554	>	}
4555	>	public boolean equals(Object o) {
4556	>	Set<?> c;
4557	>	return ((o instanceof Set) &&
4558	>	((c = (Set<?>)o) == this ||
4559	>	(containsAll(c) && c.containsAll(this))));
4560	>	}
4561	>
4562	>	/**
4563	>	* Performs the given action for each entry.
4564	>	*
4565	>	* @param action the action
4566	>	*/
4567	>	public void forEach(Action<Map.Entry<K,V>> action) {
4568	>	ForkJoinTasks.forEachEntry
4569	>	(map, action).invoke();
4570	>	}
4571	>
4572	>	/**
4573	>	* Performs the given action for each non-null transformation
4574	>	* of each entry.
4575	>	*
4576	>	* @param transformer a function returning the transformation
4577	>	* for an element, or null of there is no transformation (in
4578	>	* which case the action is not applied).
4579	>	* @param action the action
4580	>	*/
4581	>	public <U> void forEach(Fun<Map.Entry<K,V>, ? extends U> transformer,
4582	>	Action<U> action) {
4583	>	ForkJoinTasks.forEachEntry
4584	>	(map, transformer, action).invoke();
4585		}
4586	<	public int size() {
4587	<	return ConcurrentHashMap.this.size();
4586	>
4587	>	/**
4588	>	* Returns a non-null result from applying the given search
4589	>	* function on each entry, or null if none.  Upon success,
4590	>	* further element processing is suppressed and the results of
4591	>	* any other parallel invocations of the search function are
4592	>	* ignored.
4593	>	*
4594	>	* @param searchFunction a function returning a non-null
4595	>	* result on success, else null
4596	>	* @return a non-null result from applying the given search
4597	>	* function on each entry, or null if none
4598	>	*/
4599	>	public <U> U search(Fun<Map.Entry<K,V>, ? extends U> searchFunction) {
4600	>	return ForkJoinTasks.searchEntries
4601	>	(map, searchFunction).invoke();
4602		}
4603	<	public boolean isEmpty() {
4604	<	return ConcurrentHashMap.this.isEmpty();
4603	>
4604	>	/**
4605	>	* Returns the result of accumulating all entries using the
4606	>	* given reducer to combine values, or null if none.
4607	>	*
4608	>	* @param reducer a commutative associative combining function
4609	>	* @return the result of accumulating all entries
4610	>	*/
4611	>	public Map.Entry<K,V> reduce(BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer) {
4612	>	return ForkJoinTasks.reduceEntries
4613	>	(map, reducer).invoke();
4614	>	}
4615	>
4616	>	/**
4617	>	* Returns the result of accumulating the given transformation
4618	>	* of all entries using the given reducer to combine values,
4619	>	* or null if none.
4620	>	*
4621	>	* @param transformer a function returning the transformation
4622	>	* for an element, or null of there is no transformation (in
4623	>	* which case it is not combined).
4624	>	* @param reducer a commutative associative combining function
4625	>	* @return the result of accumulating the given transformation
4626	>	* of all entries
4627	>	*/
4628	>	public <U> U reduce(Fun<Map.Entry<K,V>, ? extends U> transformer,
4629	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4630	>	return ForkJoinTasks.reduceEntries
4631	>	(map, transformer, reducer).invoke();
4632		}
4633	<	public void clear() {
4634	<	ConcurrentHashMap.this.clear();
4633	>
4634	>	/**
4635	>	* Returns the result of accumulating the given transformation
4636	>	* of all entries using the given reducer to combine values,
4637	>	* and the given basis as an identity value.
4638	>	*
4639	>	* @param transformer a function returning the transformation
4640	>	* for an element
4641	>	* @param basis the identity (initial default value) for the reduction
4642	>	* @param reducer a commutative associative combining function
4643	>	* @return the result of accumulating the given transformation
4644	>	* of all entries
4645	>	*/
4646	>	public double reduceToDouble(ObjectToDouble<Map.Entry<K,V>> transformer,
4647	>	double basis,
4648	>	DoubleByDoubleToDouble reducer) {
4649	>	return ForkJoinTasks.reduceEntriesToDouble
4650	>	(map, transformer, basis, reducer).invoke();
4651		}
4652	+
4653	+	/**
4654	+	* Returns the result of accumulating the given transformation
4655	+	* of all entries using the given reducer to combine values,
4656	+	* and the given basis as an identity value.
4657	+	*
4658	+	* @param transformer a function returning the transformation
4659	+	* for an element
4660	+	* @param basis the identity (initial default value) for the reduction
4661	+	* @param reducer a commutative associative combining function
4662	+	* @return  the result of accumulating the given transformation
4663	+	* of all entries
4664	+	*/
4665	+	public long reduceToLong(ObjectToLong<Map.Entry<K,V>> transformer,
4666	+	long basis,
4667	+	LongByLongToLong reducer) {
4668	+	return ForkJoinTasks.reduceEntriesToLong
4669	+	(map, transformer, basis, reducer).invoke();
4670	+	}
4671	+
4672	+	/**
4673	+	* Returns the result of accumulating the given transformation
4674	+	* of all entries using the given reducer to combine values,
4675	+	* and the given basis as an identity value.
4676	+	*
4677	+	* @param transformer a function returning the transformation
4678	+	* for an element
4679	+	* @param basis the identity (initial default value) for the reduction
4680	+	* @param reducer a commutative associative combining function
4681	+	* @return the result of accumulating the given transformation
4682	+	* of all entries
4683	+	*/
4684	+	public int reduceToInt(ObjectToInt<Map.Entry<K,V>> transformer,
4685	+	int basis,
4686	+	IntByIntToInt reducer) {
4687	+	return ForkJoinTasks.reduceEntriesToInt
4688	+	(map, transformer, basis, reducer).invoke();
4689	+	}
4690	+
4691		}
4692
4693	<	/* ---------------- Serialization Support -------------- */
4693	>	// ---------------------------------------------------------------------
4694
4695		/**
4696	<	* Saves the state of the <tt>ConcurrentHashMap</tt> instance to a
4697	<	* stream (i.e., serializes it).
4698	<	* @param s the stream
4699	<	* @serialData
4700	<	* the key (Object) and value (Object)
4701	<	* for each key-value mapping, followed by a null pair.
4702	<	* The key-value mappings are emitted in no particular order.
4696	>	* Predefined tasks for performing bulk parallel operations on
4697	>	* ConcurrentHashMaps. These tasks follow the forms and rules used
4698	>	* for bulk operations. Each method has the same name, but returns
4699	>	* a task rather than invoking it. These methods may be useful in
4700	>	* custom applications such as submitting a task without waiting
4701	>	* for completion, using a custom pool, or combining with other
4702	>	* tasks.
4703		*/
4704	<	private void writeObject(java.io.ObjectOutputStream s)
4705	<	throws java.io.IOException {
1402	<	// force all segments for serialization compatibility
1403	<	for (int k = 0; k < segments.length; ++k)
1404	<	ensureSegment(k);
1405	<	s.defaultWriteObject();
4704	>	public static class ForkJoinTasks {
4705	>	private ForkJoinTasks() {}
4706
4707	<	final Segment<K,V>[] segments = this.segments;
4708	<	for (int k = 0; k < segments.length; ++k) {
4709	<	Segment<K,V> seg = segmentAt(segments, k);
4710	<	seg.lock();
4711	<	try {
4712	<	HashEntry<K,V>[] tab = seg.table;
4713	<	for (int i = 0; i < tab.length; ++i) {
4714	<	HashEntry<K,V> e;
4715	<	for (e = entryAt(tab, i); e != null; e = e.next) {
4716	<	s.writeObject(e.key);
4717	<	s.writeObject(e.value);
4718	<	}
4707	>	/**
4708	>	* Returns a task that when invoked, performs the given
4709	>	* action for each (key, value)
4710	>	*
4711	>	* @param map the map
4712	>	* @param action the action
4713	>	* @return the task
4714	>	*/
4715	>	public static <K,V> ForkJoinTask<Void> forEach
4716	>	(ConcurrentHashMap<K,V> map,
4717	>	BiAction<K,V> action) {
4718	>	if (action == null) throw new NullPointerException();
4719	>	return new ForEachMappingTask<K,V>(map, null, -1, action);
4720	>	}
4721	>
4722	>	/**
4723	>	* Returns a task that when invoked, performs the given
4724	>	* action for each non-null transformation of each (key, value)
4725	>	*
4726	>	* @param map the map
4727	>	* @param transformer a function returning the transformation
4728	>	* for an element, or null if there is no transformation (in
4729	>	* which case the action is not applied)
4730	>	* @param action the action
4731	>	* @return the task
4732	>	*/
4733	>	public static <K,V,U> ForkJoinTask<Void> forEach
4734	>	(ConcurrentHashMap<K,V> map,
4735	>	BiFun<? super K, ? super V, ? extends U> transformer,
4736	>	Action<U> action) {
4737	>	if (transformer == null || action == null)
4738	>	throw new NullPointerException();
4739	>	return new ForEachTransformedMappingTask<K,V,U>
4740	>	(map, null, -1, transformer, action);
4741	>	}
4742	>
4743	>	/**
4744	>	* Returns a task that when invoked, returns a non-null result
4745	>	* from applying the given search function on each (key,
4746	>	* value), or null if none. Upon success, further element
4747	>	* processing is suppressed and the results of any other
4748	>	* parallel invocations of the search function are ignored.
4749	>	*
4750	>	* @param map the map
4751	>	* @param searchFunction a function returning a non-null
4752	>	* result on success, else null
4753	>	* @return the task
4754	>	*/
4755	>	public static <K,V,U> ForkJoinTask<U> search
4756	>	(ConcurrentHashMap<K,V> map,
4757	>	BiFun<? super K, ? super V, ? extends U> searchFunction) {
4758	>	if (searchFunction == null) throw new NullPointerException();
4759	>	return new SearchMappingsTask<K,V,U>
4760	>	(map, null, -1, searchFunction,
4761	>	new AtomicReference<U>());
4762	>	}
4763	>
4764	>	/**
4765	>	* Returns a task that when invoked, returns the result of
4766	>	* accumulating the given transformation of all (key, value) pairs
4767	>	* using the given reducer to combine values, or null if none.
4768	>	*
4769	>	* @param map the map
4770	>	* @param transformer a function returning the transformation
4771	>	* for an element, or null if there is no transformation (in
4772	>	* which case it is not combined).
4773	>	* @param reducer a commutative associative combining function
4774	>	* @return the task
4775	>	*/
4776	>	public static <K,V,U> ForkJoinTask<U> reduce
4777	>	(ConcurrentHashMap<K,V> map,
4778	>	BiFun<? super K, ? super V, ? extends U> transformer,
4779	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4780	>	if (transformer == null || reducer == null)
4781	>	throw new NullPointerException();
4782	>	return new MapReduceMappingsTask<K,V,U>
4783	>	(map, null, -1, null, transformer, reducer);
4784	>	}
4785	>
4786	>	/**
4787	>	* Returns a task that when invoked, returns the result of
4788	>	* accumulating the given transformation of all (key, value) pairs
4789	>	* using the given reducer to combine values, and the given
4790	>	* basis as an identity value.
4791	>	*
4792	>	* @param map the map
4793	>	* @param transformer a function returning the transformation
4794	>	* for an element
4795	>	* @param basis the identity (initial default value) for the reduction
4796	>	* @param reducer a commutative associative combining function
4797	>	* @return the task
4798	>	*/
4799	>	public static <K,V> ForkJoinTask<Double> reduceToDouble
4800	>	(ConcurrentHashMap<K,V> map,
4801	>	ObjectByObjectToDouble<? super K, ? super V> transformer,
4802	>	double basis,
4803	>	DoubleByDoubleToDouble reducer) {
4804	>	if (transformer == null || reducer == null)
4805	>	throw new NullPointerException();
4806	>	return new MapReduceMappingsToDoubleTask<K,V>
4807	>	(map, null, -1, null, transformer, basis, reducer);
4808	>	}
4809	>
4810	>	/**
4811	>	* Returns a task that when invoked, returns the result of
4812	>	* accumulating the given transformation of all (key, value) pairs
4813	>	* using the given reducer to combine values, and the given
4814	>	* basis as an identity value.
4815	>	*
4816	>	* @param map the map
4817	>	* @param transformer a function returning the transformation
4818	>	* for an element
4819	>	* @param basis the identity (initial default value) for the reduction
4820	>	* @param reducer a commutative associative combining function
4821	>	* @return the task
4822	>	*/
4823	>	public static <K,V> ForkJoinTask<Long> reduceToLong
4824	>	(ConcurrentHashMap<K,V> map,
4825	>	ObjectByObjectToLong<? super K, ? super V> transformer,
4826	>	long basis,
4827	>	LongByLongToLong reducer) {
4828	>	if (transformer == null || reducer == null)
4829	>	throw new NullPointerException();
4830	>	return new MapReduceMappingsToLongTask<K,V>
4831	>	(map, null, -1, null, transformer, basis, reducer);
4832	>	}
4833	>
4834	>	/**
4835	>	* Returns a task that when invoked, returns the result of
4836	>	* accumulating the given transformation of all (key, value) pairs
4837	>	* using the given reducer to combine values, and the given
4838	>	* basis as an identity value.
4839	>	*
4840	>	* @param transformer a function returning the transformation
4841	>	* for an element
4842	>	* @param basis the identity (initial default value) for the reduction
4843	>	* @param reducer a commutative associative combining function
4844	>	* @return the task
4845	>	*/
4846	>	public static <K,V> ForkJoinTask<Integer> reduceToInt
4847	>	(ConcurrentHashMap<K,V> map,
4848	>	ObjectByObjectToInt<? super K, ? super V> transformer,
4849	>	int basis,
4850	>	IntByIntToInt reducer) {
4851	>	if (transformer == null || reducer == null)
4852	>	throw new NullPointerException();
4853	>	return new MapReduceMappingsToIntTask<K,V>
4854	>	(map, null, -1, null, transformer, basis, reducer);
4855	>	}
4856	>
4857	>	/**
4858	>	* Returns a task that when invoked, performs the given action
4859	>	* for each key.
4860	>	*
4861	>	* @param map the map
4862	>	* @param action the action
4863	>	* @return the task
4864	>	*/
4865	>	public static <K,V> ForkJoinTask<Void> forEachKey
4866	>	(ConcurrentHashMap<K,V> map,
4867	>	Action<K> action) {
4868	>	if (action == null) throw new NullPointerException();
4869	>	return new ForEachKeyTask<K,V>(map, null, -1, action);
4870	>	}
4871	>
4872	>	/**
4873	>	* Returns a task that when invoked, performs the given action
4874	>	* for each non-null transformation of each key.
4875	>	*
4876	>	* @param map the map
4877	>	* @param transformer a function returning the transformation
4878	>	* for an element, or null if there is no transformation (in
4879	>	* which case the action is not applied)
4880	>	* @param action the action
4881	>	* @return the task
4882	>	*/
4883	>	public static <K,V,U> ForkJoinTask<Void> forEachKey
4884	>	(ConcurrentHashMap<K,V> map,
4885	>	Fun<? super K, ? extends U> transformer,
4886	>	Action<U> action) {
4887	>	if (transformer == null || action == null)
4888	>	throw new NullPointerException();
4889	>	return new ForEachTransformedKeyTask<K,V,U>
4890	>	(map, null, -1, transformer, action);
4891	>	}
4892	>
4893	>	/**
4894	>	* Returns a task that when invoked, returns a non-null result
4895	>	* from applying the given search function on each key, or
4896	>	* null if none.  Upon success, further element processing is
4897	>	* suppressed and the results of any other parallel
4898	>	* invocations of the search function are ignored.
4899	>	*
4900	>	* @param map the map
4901	>	* @param searchFunction a function returning a non-null
4902	>	* result on success, else null
4903	>	* @return the task
4904	>	*/
4905	>	public static <K,V,U> ForkJoinTask<U> searchKeys
4906	>	(ConcurrentHashMap<K,V> map,
4907	>	Fun<? super K, ? extends U> searchFunction) {
4908	>	if (searchFunction == null) throw new NullPointerException();
4909	>	return new SearchKeysTask<K,V,U>
4910	>	(map, null, -1, searchFunction,
4911	>	new AtomicReference<U>());
4912	>	}
4913	>
4914	>	/**
4915	>	* Returns a task that when invoked, returns the result of
4916	>	* accumulating all keys using the given reducer to combine
4917	>	* values, or null if none.
4918	>	*
4919	>	* @param map the map
4920	>	* @param reducer a commutative associative combining function
4921	>	* @return the task
4922	>	*/
4923	>	public static <K,V> ForkJoinTask<K> reduceKeys
4924	>	(ConcurrentHashMap<K,V> map,
4925	>	BiFun<? super K, ? super K, ? extends K> reducer) {
4926	>	if (reducer == null) throw new NullPointerException();
4927	>	return new ReduceKeysTask<K,V>
4928	>	(map, null, -1, null, reducer);
4929	>	}
4930	>
4931	>	/**
4932	>	* Returns a task that when invoked, returns the result of
4933	>	* accumulating the given transformation of all keys using the given
4934	>	* reducer to combine values, or null if none.
4935	>	*
4936	>	* @param map the map
4937	>	* @param transformer a function returning the transformation
4938	>	* for an element, or null if there is no transformation (in
4939	>	* which case it is not combined).
4940	>	* @param reducer a commutative associative combining function
4941	>	* @return the task
4942	>	*/
4943	>	public static <K,V,U> ForkJoinTask<U> reduceKeys
4944	>	(ConcurrentHashMap<K,V> map,
4945	>	Fun<? super K, ? extends U> transformer,
4946	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4947	>	if (transformer == null || reducer == null)
4948	>	throw new NullPointerException();
4949	>	return new MapReduceKeysTask<K,V,U>
4950	>	(map, null, -1, null, transformer, reducer);
4951	>	}
4952	>
4953	>	/**
4954	>	* Returns a task that when invoked, returns the result of
4955	>	* accumulating the given transformation of all keys using the given
4956	>	* reducer to combine values, and the given basis as an
4957	>	* identity value.
4958	>	*
4959	>	* @param map the map
4960	>	* @param transformer a function returning the transformation
4961	>	* for an element
4962	>	* @param basis the identity (initial default value) for the reduction
4963	>	* @param reducer a commutative associative combining function
4964	>	* @return the task
4965	>	*/
4966	>	public static <K,V> ForkJoinTask<Double> reduceKeysToDouble
4967	>	(ConcurrentHashMap<K,V> map,
4968	>	ObjectToDouble<? super K> transformer,
4969	>	double basis,
4970	>	DoubleByDoubleToDouble reducer) {
4971	>	if (transformer == null || reducer == null)
4972	>	throw new NullPointerException();
4973	>	return new MapReduceKeysToDoubleTask<K,V>
4974	>	(map, null, -1, null, transformer, basis, reducer);
4975	>	}
4976	>
4977	>	/**
4978	>	* Returns a task that when invoked, returns the result of
4979	>	* accumulating the given transformation of all keys using the given
4980	>	* reducer to combine values, and the given basis as an
4981	>	* identity value.
4982	>	*
4983	>	* @param map the map
4984	>	* @param transformer a function returning the transformation
4985	>	* for an element
4986	>	* @param basis the identity (initial default value) for the reduction
4987	>	* @param reducer a commutative associative combining function
4988	>	* @return the task
4989	>	*/
4990	>	public static <K,V> ForkJoinTask<Long> reduceKeysToLong
4991	>	(ConcurrentHashMap<K,V> map,
4992	>	ObjectToLong<? super K> transformer,
4993	>	long basis,
4994	>	LongByLongToLong reducer) {
4995	>	if (transformer == null || reducer == null)
4996	>	throw new NullPointerException();
4997	>	return new MapReduceKeysToLongTask<K,V>
4998	>	(map, null, -1, null, transformer, basis, reducer);
4999	>	}
5000	>
5001	>	/**
5002	>	* Returns a task that when invoked, returns the result of
5003	>	* accumulating the given transformation of all keys using the given
5004	>	* reducer to combine values, and the given basis as an
5005	>	* identity value.
5006	>	*
5007	>	* @param map the map
5008	>	* @param transformer a function returning the transformation
5009	>	* for an element
5010	>	* @param basis the identity (initial default value) for the reduction
5011	>	* @param reducer a commutative associative combining function
5012	>	* @return the task
5013	>	*/
5014	>	public static <K,V> ForkJoinTask<Integer> reduceKeysToInt
5015	>	(ConcurrentHashMap<K,V> map,
5016	>	ObjectToInt<? super K> transformer,
5017	>	int basis,
5018	>	IntByIntToInt reducer) {
5019	>	if (transformer == null || reducer == null)
5020	>	throw new NullPointerException();
5021	>	return new MapReduceKeysToIntTask<K,V>
5022	>	(map, null, -1, null, transformer, basis, reducer);
5023	>	}
5024	>
5025	>	/**
5026	>	* Returns a task that when invoked, performs the given action
5027	>	* for each value.
5028	>	*
5029	>	* @param map the map
5030	>	* @param action the action
5031	>	*/
5032	>	public static <K,V> ForkJoinTask<Void> forEachValue
5033	>	(ConcurrentHashMap<K,V> map,
5034	>	Action<V> action) {
5035	>	if (action == null) throw new NullPointerException();
5036	>	return new ForEachValueTask<K,V>(map, null, -1, action);
5037	>	}
5038	>
5039	>	/**
5040	>	* Returns a task that when invoked, performs the given action
5041	>	* for each non-null transformation of each value.
5042	>	*
5043	>	* @param map the map
5044	>	* @param transformer a function returning the transformation
5045	>	* for an element, or null if there is no transformation (in
5046	>	* which case the action is not applied)
5047	>	* @param action the action
5048	>	*/
5049	>	public static <K,V,U> ForkJoinTask<Void> forEachValue
5050	>	(ConcurrentHashMap<K,V> map,
5051	>	Fun<? super V, ? extends U> transformer,
5052	>	Action<U> action) {
5053	>	if (transformer == null || action == null)
5054	>	throw new NullPointerException();
5055	>	return new ForEachTransformedValueTask<K,V,U>
5056	>	(map, null, -1, transformer, action);
5057	>	}
5058	>
5059	>	/**
5060	>	* Returns a task that when invoked, returns a non-null result
5061	>	* from applying the given search function on each value, or
5062	>	* null if none.  Upon success, further element processing is
5063	>	* suppressed and the results of any other parallel
5064	>	* invocations of the search function are ignored.
5065	>	*
5066	>	* @param map the map
5067	>	* @param searchFunction a function returning a non-null
5068	>	* result on success, else null
5069	>	* @return the task
5070	>	*/
5071	>	public static <K,V,U> ForkJoinTask<U> searchValues
5072	>	(ConcurrentHashMap<K,V> map,
5073	>	Fun<? super V, ? extends U> searchFunction) {
5074	>	if (searchFunction == null) throw new NullPointerException();
5075	>	return new SearchValuesTask<K,V,U>
5076	>	(map, null, -1, searchFunction,
5077	>	new AtomicReference<U>());
5078	>	}
5079	>
5080	>	/**
5081	>	* Returns a task that when invoked, returns the result of
5082	>	* accumulating all values using the given reducer to combine
5083	>	* values, or null if none.
5084	>	*
5085	>	* @param map the map
5086	>	* @param reducer a commutative associative combining function
5087	>	* @return the task
5088	>	*/
5089	>	public static <K,V> ForkJoinTask<V> reduceValues
5090	>	(ConcurrentHashMap<K,V> map,
5091	>	BiFun<? super V, ? super V, ? extends V> reducer) {
5092	>	if (reducer == null) throw new NullPointerException();
5093	>	return new ReduceValuesTask<K,V>
5094	>	(map, null, -1, null, reducer);
5095	>	}
5096	>
5097	>	/**
5098	>	* Returns a task that when invoked, returns the result of
5099	>	* accumulating the given transformation of all values using the
5100	>	* given reducer to combine values, or null if none.
5101	>	*
5102	>	* @param map the map
5103	>	* @param transformer a function returning the transformation
5104	>	* for an element, or null if there is no transformation (in
5105	>	* which case it is not combined).
5106	>	* @param reducer a commutative associative combining function
5107	>	* @return the task
5108	>	*/
5109	>	public static <K,V,U> ForkJoinTask<U> reduceValues
5110	>	(ConcurrentHashMap<K,V> map,
5111	>	Fun<? super V, ? extends U> transformer,
5112	>	BiFun<? super U, ? super U, ? extends U> reducer) {
5113	>	if (transformer == null || reducer == null)
5114	>	throw new NullPointerException();
5115	>	return new MapReduceValuesTask<K,V,U>
5116	>	(map, null, -1, null, transformer, reducer);
5117	>	}
5118	>
5119	>	/**
5120	>	* Returns a task that when invoked, returns the result of
5121	>	* accumulating the given transformation of all values using the
5122	>	* given reducer to combine values, and the given basis as an
5123	>	* identity value.
5124	>	*
5125	>	* @param map the map
5126	>	* @param transformer a function returning the transformation
5127	>	* for an element
5128	>	* @param basis the identity (initial default value) for the reduction
5129	>	* @param reducer a commutative associative combining function
5130	>	* @return the task
5131	>	*/
5132	>	public static <K,V> ForkJoinTask<Double> reduceValuesToDouble
5133	>	(ConcurrentHashMap<K,V> map,
5134	>	ObjectToDouble<? super V> transformer,
5135	>	double basis,
5136	>	DoubleByDoubleToDouble reducer) {
5137	>	if (transformer == null || reducer == null)
5138	>	throw new NullPointerException();
5139	>	return new MapReduceValuesToDoubleTask<K,V>
5140	>	(map, null, -1, null, transformer, basis, reducer);
5141	>	}
5142	>
5143	>	/**
5144	>	* Returns a task that when invoked, returns the result of
5145	>	* accumulating the given transformation of all values using the
5146	>	* given reducer to combine values, and the given basis as an
5147	>	* identity value.
5148	>	*
5149	>	* @param map the map
5150	>	* @param transformer a function returning the transformation
5151	>	* for an element
5152	>	* @param basis the identity (initial default value) for the reduction
5153	>	* @param reducer a commutative associative combining function
5154	>	* @return the task
5155	>	*/
5156	>	public static <K,V> ForkJoinTask<Long> reduceValuesToLong
5157	>	(ConcurrentHashMap<K,V> map,
5158	>	ObjectToLong<? super V> transformer,
5159	>	long basis,
5160	>	LongByLongToLong reducer) {
5161	>	if (transformer == null || reducer == null)
5162	>	throw new NullPointerException();
5163	>	return new MapReduceValuesToLongTask<K,V>
5164	>	(map, null, -1, null, transformer, basis, reducer);
5165	>	}
5166	>
5167	>	/**
5168	>	* Returns a task that when invoked, returns the result of
5169	>	* accumulating the given transformation of all values using the
5170	>	* given reducer to combine values, and the given basis as an
5171	>	* identity value.
5172	>	*
5173	>	* @param map the map
5174	>	* @param transformer a function returning the transformation
5175	>	* for an element
5176	>	* @param basis the identity (initial default value) for the reduction
5177	>	* @param reducer a commutative associative combining function
5178	>	* @return the task
5179	>	*/
5180	>	public static <K,V> ForkJoinTask<Integer> reduceValuesToInt
5181	>	(ConcurrentHashMap<K,V> map,
5182	>	ObjectToInt<? super V> transformer,
5183	>	int basis,
5184	>	IntByIntToInt reducer) {
5185	>	if (transformer == null || reducer == null)
5186	>	throw new NullPointerException();
5187	>	return new MapReduceValuesToIntTask<K,V>
5188	>	(map, null, -1, null, transformer, basis, reducer);
5189	>	}
5190	>
5191	>	/**
5192	>	* Returns a task that when invoked, perform the given action
5193	>	* for each entry.
5194	>	*
5195	>	* @param map the map
5196	>	* @param action the action
5197	>	*/
5198	>	public static <K,V> ForkJoinTask<Void> forEachEntry
5199	>	(ConcurrentHashMap<K,V> map,
5200	>	Action<Map.Entry<K,V>> action) {
5201	>	if (action == null) throw new NullPointerException();
5202	>	return new ForEachEntryTask<K,V>(map, null, -1, action);
5203	>	}
5204	>
5205	>	/**
5206	>	* Returns a task that when invoked, perform the given action
5207	>	* for each non-null transformation of each entry.
5208	>	*
5209	>	* @param map the map
5210	>	* @param transformer a function returning the transformation
5211	>	* for an element, or null if there is no transformation (in
5212	>	* which case the action is not applied)
5213	>	* @param action the action
5214	>	*/
5215	>	public static <K,V,U> ForkJoinTask<Void> forEachEntry
5216	>	(ConcurrentHashMap<K,V> map,
5217	>	Fun<Map.Entry<K,V>, ? extends U> transformer,
5218	>	Action<U> action) {
5219	>	if (transformer == null || action == null)
5220	>	throw new NullPointerException();
5221	>	return new ForEachTransformedEntryTask<K,V,U>
5222	>	(map, null, -1, transformer, action);
5223	>	}
5224	>
5225	>	/**
5226	>	* Returns a task that when invoked, returns a non-null result
5227	>	* from applying the given search function on each entry, or
5228	>	* null if none.  Upon success, further element processing is
5229	>	* suppressed and the results of any other parallel
5230	>	* invocations of the search function are ignored.
5231	>	*
5232	>	* @param map the map
5233	>	* @param searchFunction a function returning a non-null
5234	>	* result on success, else null
5235	>	* @return the task
5236	>	*/
5237	>	public static <K,V,U> ForkJoinTask<U> searchEntries
5238	>	(ConcurrentHashMap<K,V> map,
5239	>	Fun<Map.Entry<K,V>, ? extends U> searchFunction) {
5240	>	if (searchFunction == null) throw new NullPointerException();
5241	>	return new SearchEntriesTask<K,V,U>
5242	>	(map, null, -1, searchFunction,
5243	>	new AtomicReference<U>());
5244	>	}
5245	>
5246	>	/**
5247	>	* Returns a task that when invoked, returns the result of
5248	>	* accumulating all entries using the given reducer to combine
5249	>	* values, or null if none.
5250	>	*
5251	>	* @param map the map
5252	>	* @param reducer a commutative associative combining function
5253	>	* @return the task
5254	>	*/
5255	>	public static <K,V> ForkJoinTask<Map.Entry<K,V>> reduceEntries
5256	>	(ConcurrentHashMap<K,V> map,
5257	>	BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer) {
5258	>	if (reducer == null) throw new NullPointerException();
5259	>	return new ReduceEntriesTask<K,V>
5260	>	(map, null, -1, null, reducer);
5261	>	}
5262	>
5263	>	/**
5264	>	* Returns a task that when invoked, returns the result of
5265	>	* accumulating the given transformation of all entries using the
5266	>	* given reducer to combine values, or null if none.
5267	>	*
5268	>	* @param map the map
5269	>	* @param transformer a function returning the transformation
5270	>	* for an element, or null if there is no transformation (in
5271	>	* which case it is not combined).
5272	>	* @param reducer a commutative associative combining function
5273	>	* @return the task
5274	>	*/
5275	>	public static <K,V,U> ForkJoinTask<U> reduceEntries
5276	>	(ConcurrentHashMap<K,V> map,
5277	>	Fun<Map.Entry<K,V>, ? extends U> transformer,
5278	>	BiFun<? super U, ? super U, ? extends U> reducer) {
5279	>	if (transformer == null || reducer == null)
5280	>	throw new NullPointerException();
5281	>	return new MapReduceEntriesTask<K,V,U>
5282	>	(map, null, -1, null, transformer, reducer);
5283	>	}
5284	>
5285	>	/**
5286	>	* Returns a task that when invoked, returns the result of
5287	>	* accumulating the given transformation of all entries using the
5288	>	* given reducer to combine values, and the given basis as an
5289	>	* identity value.
5290	>	*
5291	>	* @param map the map
5292	>	* @param transformer a function returning the transformation
5293	>	* for an element
5294	>	* @param basis the identity (initial default value) for the reduction
5295	>	* @param reducer a commutative associative combining function
5296	>	* @return the task
5297	>	*/
5298	>	public static <K,V> ForkJoinTask<Double> reduceEntriesToDouble
5299	>	(ConcurrentHashMap<K,V> map,
5300	>	ObjectToDouble<Map.Entry<K,V>> transformer,
5301	>	double basis,
5302	>	DoubleByDoubleToDouble reducer) {
5303	>	if (transformer == null || reducer == null)
5304	>	throw new NullPointerException();
5305	>	return new MapReduceEntriesToDoubleTask<K,V>
5306	>	(map, null, -1, null, transformer, basis, reducer);
5307	>	}
5308	>
5309	>	/**
5310	>	* Returns a task that when invoked, returns the result of
5311	>	* accumulating the given transformation of all entries using the
5312	>	* given reducer to combine values, and the given basis as an
5313	>	* identity value.
5314	>	*
5315	>	* @param map the map
5316	>	* @param transformer a function returning the transformation
5317	>	* for an element
5318	>	* @param basis the identity (initial default value) for the reduction
5319	>	* @param reducer a commutative associative combining function
5320	>	* @return the task
5321	>	*/
5322	>	public static <K,V> ForkJoinTask<Long> reduceEntriesToLong
5323	>	(ConcurrentHashMap<K,V> map,
5324	>	ObjectToLong<Map.Entry<K,V>> transformer,
5325	>	long basis,
5326	>	LongByLongToLong reducer) {
5327	>	if (transformer == null || reducer == null)
5328	>	throw new NullPointerException();
5329	>	return new MapReduceEntriesToLongTask<K,V>
5330	>	(map, null, -1, null, transformer, basis, reducer);
5331	>	}
5332	>
5333	>	/**
5334	>	* Returns a task that when invoked, returns the result of
5335	>	* accumulating the given transformation of all entries using the
5336	>	* given reducer to combine values, and the given basis as an
5337	>	* identity value.
5338	>	*
5339	>	* @param map the map
5340	>	* @param transformer a function returning the transformation
5341	>	* for an element
5342	>	* @param basis the identity (initial default value) for the reduction
5343	>	* @param reducer a commutative associative combining function
5344	>	* @return the task
5345	>	*/
5346	>	public static <K,V> ForkJoinTask<Integer> reduceEntriesToInt
5347	>	(ConcurrentHashMap<K,V> map,
5348	>	ObjectToInt<Map.Entry<K,V>> transformer,
5349	>	int basis,
5350	>	IntByIntToInt reducer) {
5351	>	if (transformer == null || reducer == null)
5352	>	throw new NullPointerException();
5353	>	return new MapReduceEntriesToIntTask<K,V>
5354	>	(map, null, -1, null, transformer, basis, reducer);
5355	>	}
5356	>	}
5357	>
5358	>	// -------------------------------------------------------
5359	>
5360	>	/*
5361	>	* Task classes. Coded in a regular but ugly format/style to
5362	>	* simplify checks that each variant differs in the right way from
5363	>	* others.
5364	>	*/
5365	>
5366	>	@SuppressWarnings("serial") static final class ForEachKeyTask<K,V>
5367	>	extends Traverser<K,V,Void> {
5368	>	final Action<K> action;
5369	>	ForEachKeyTask
5370	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5371	>	Action<K> action) {
5372	>	super(m, p, b);
5373	>	this.action = action;
5374	>	}
5375	>	@SuppressWarnings("unchecked") public final void compute() {
5376	>	final Action<K> action;
5377	>	if ((action = this.action) == null)
5378	>	throw new NullPointerException();
5379	>	for (int b; (b = preSplit()) > 0;)
5380	>	new ForEachKeyTask<K,V>(map, this, b, action).fork();
5381	>	while (advance() != null)
5382	>	action.apply((K)nextKey);
5383	>	propagateCompletion();
5384	>	}
5385	>	}
5386	>
5387	>	@SuppressWarnings("serial") static final class ForEachValueTask<K,V>
5388	>	extends Traverser<K,V,Void> {
5389	>	final Action<V> action;
5390	>	ForEachValueTask
5391	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5392	>	Action<V> action) {
5393	>	super(m, p, b);
5394	>	this.action = action;
5395	>	}
5396	>	@SuppressWarnings("unchecked") public final void compute() {
5397	>	final Action<V> action;
5398	>	if ((action = this.action) == null)
5399	>	throw new NullPointerException();
5400	>	for (int b; (b = preSplit()) > 0;)
5401	>	new ForEachValueTask<K,V>(map, this, b, action).fork();
5402	>	Object v;
5403	>	while ((v = advance()) != null)
5404	>	action.apply((V)v);
5405	>	propagateCompletion();
5406	>	}
5407	>	}
5408	>
5409	>	@SuppressWarnings("serial") static final class ForEachEntryTask<K,V>
5410	>	extends Traverser<K,V,Void> {
5411	>	final Action<Entry<K,V>> action;
5412	>	ForEachEntryTask
5413	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5414	>	Action<Entry<K,V>> action) {
5415	>	super(m, p, b);
5416	>	this.action = action;
5417	>	}
5418	>	@SuppressWarnings("unchecked") public final void compute() {
5419	>	final Action<Entry<K,V>> action;
5420	>	if ((action = this.action) == null)
5421	>	throw new NullPointerException();
5422	>	for (int b; (b = preSplit()) > 0;)
5423	>	new ForEachEntryTask<K,V>(map, this, b, action).fork();
5424	>	Object v;
5425	>	while ((v = advance()) != null)
5426	>	action.apply(entryFor((K)nextKey, (V)v));
5427	>	propagateCompletion();
5428	>	}
5429	>	}
5430	>
5431	>	@SuppressWarnings("serial") static final class ForEachMappingTask<K,V>
5432	>	extends Traverser<K,V,Void> {
5433	>	final BiAction<K,V> action;
5434	>	ForEachMappingTask
5435	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5436	>	BiAction<K,V> action) {
5437	>	super(m, p, b);
5438	>	this.action = action;
5439	>	}
5440	>	@SuppressWarnings("unchecked") public final void compute() {
5441	>	final BiAction<K,V> action;
5442	>	if ((action = this.action) == null)
5443	>	throw new NullPointerException();
5444	>	for (int b; (b = preSplit()) > 0;)
5445	>	new ForEachMappingTask<K,V>(map, this, b, action).fork();
5446	>	Object v;
5447	>	while ((v = advance()) != null)
5448	>	action.apply((K)nextKey, (V)v);
5449	>	propagateCompletion();
5450	>	}
5451	>	}
5452	>
5453	>	@SuppressWarnings("serial") static final class ForEachTransformedKeyTask<K,V,U>
5454	>	extends Traverser<K,V,Void> {
5455	>	final Fun<? super K, ? extends U> transformer;
5456	>	final Action<U> action;
5457	>	ForEachTransformedKeyTask
5458	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5459	>	Fun<? super K, ? extends U> transformer, Action<U> action) {
5460	>	super(m, p, b);
5461	>	this.transformer = transformer; this.action = action;
5462	>	}
5463	>	@SuppressWarnings("unchecked") public final void compute() {
5464	>	final Fun<? super K, ? extends U> transformer;
5465	>	final Action<U> action;
5466	>	if ((transformer = this.transformer) == null ||
5467	>	(action = this.action) == null)
5468	>	throw new NullPointerException();
5469	>	for (int b; (b = preSplit()) > 0;)
5470	>	new ForEachTransformedKeyTask<K,V,U>
5471	>	(map, this, b, transformer, action).fork();
5472	>	U u;
5473	>	while (advance() != null) {
5474	>	if ((u = transformer.apply((K)nextKey)) != null)
5475	>	action.apply(u);
5476	>	}
5477	>	propagateCompletion();
5478	>	}
5479	>	}
5480	>
5481	>	@SuppressWarnings("serial") static final class ForEachTransformedValueTask<K,V,U>
5482	>	extends Traverser<K,V,Void> {
5483	>	final Fun<? super V, ? extends U> transformer;
5484	>	final Action<U> action;
5485	>	ForEachTransformedValueTask
5486	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5487	>	Fun<? super V, ? extends U> transformer, Action<U> action) {
5488	>	super(m, p, b);
5489	>	this.transformer = transformer; this.action = action;
5490	>	}
5491	>	@SuppressWarnings("unchecked") public final void compute() {
5492	>	final Fun<? super V, ? extends U> transformer;
5493	>	final Action<U> action;
5494	>	if ((transformer = this.transformer) == null ||
5495	>	(action = this.action) == null)
5496	>	throw new NullPointerException();
5497	>	for (int b; (b = preSplit()) > 0;)
5498	>	new ForEachTransformedValueTask<K,V,U>
5499	>	(map, this, b, transformer, action).fork();
5500	>	Object v; U u;
5501	>	while ((v = advance()) != null) {
5502	>	if ((u = transformer.apply((V)v)) != null)
5503	>	action.apply(u);
5504	>	}
5505	>	propagateCompletion();
5506	>	}
5507	>	}
5508	>
5509	>	@SuppressWarnings("serial") static final class ForEachTransformedEntryTask<K,V,U>
5510	>	extends Traverser<K,V,Void> {
5511	>	final Fun<Map.Entry<K,V>, ? extends U> transformer;
5512	>	final Action<U> action;
5513	>	ForEachTransformedEntryTask
5514	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5515	>	Fun<Map.Entry<K,V>, ? extends U> transformer, Action<U> action) {
5516	>	super(m, p, b);
5517	>	this.transformer = transformer; this.action = action;
5518	>	}
5519	>	@SuppressWarnings("unchecked") public final void compute() {
5520	>	final Fun<Map.Entry<K,V>, ? extends U> transformer;
5521	>	final Action<U> action;
5522	>	if ((transformer = this.transformer) == null ||
5523	>	(action = this.action) == null)
5524	>	throw new NullPointerException();
5525	>	for (int b; (b = preSplit()) > 0;)
5526	>	new ForEachTransformedEntryTask<K,V,U>
5527	>	(map, this, b, transformer, action).fork();
5528	>	Object v; U u;
5529	>	while ((v = advance()) != null) {
5530	>	if ((u = transformer.apply(entryFor((K)nextKey, (V)v))) != null)
5531	>	action.apply(u);
5532	>	}
5533	>	propagateCompletion();
5534	>	}
5535	>	}
5536	>
5537	>	@SuppressWarnings("serial") static final class ForEachTransformedMappingTask<K,V,U>
5538	>	extends Traverser<K,V,Void> {
5539	>	final BiFun<? super K, ? super V, ? extends U> transformer;
5540	>	final Action<U> action;
5541	>	ForEachTransformedMappingTask
5542	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5543	>	BiFun<? super K, ? super V, ? extends U> transformer,
5544	>	Action<U> action) {
5545	>	super(m, p, b);
5546	>	this.transformer = transformer; this.action = action;
5547	>	}
5548	>	@SuppressWarnings("unchecked") public final void compute() {
5549	>	final BiFun<? super K, ? super V, ? extends U> transformer;
5550	>	final Action<U> action;
5551	>	if ((transformer = this.transformer) == null ||
5552	>	(action = this.action) == null)
5553	>	throw new NullPointerException();
5554	>	for (int b; (b = preSplit()) > 0;)
5555	>	new ForEachTransformedMappingTask<K,V,U>
5556	>	(map, this, b, transformer, action).fork();
5557	>	Object v; U u;
5558	>	while ((v = advance()) != null) {
5559	>	if ((u = transformer.apply((K)nextKey, (V)v)) != null)
5560	>	action.apply(u);
5561	>	}
5562	>	propagateCompletion();
5563	>	}
5564	>	}
5565	>
5566	>	@SuppressWarnings("serial") static final class SearchKeysTask<K,V,U>
5567	>	extends Traverser<K,V,U> {
5568	>	final Fun<? super K, ? extends U> searchFunction;
5569	>	final AtomicReference<U> result;
5570	>	SearchKeysTask
5571	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5572	>	Fun<? super K, ? extends U> searchFunction,
5573	>	AtomicReference<U> result) {
5574	>	super(m, p, b);
5575	>	this.searchFunction = searchFunction; this.result = result;
5576	>	}
5577	>	public final U getRawResult() { return result.get(); }
5578	>	@SuppressWarnings("unchecked") public final void compute() {
5579	>	final Fun<? super K, ? extends U> searchFunction;
5580	>	final AtomicReference<U> result;
5581	>	if ((searchFunction = this.searchFunction) == null ||
5582	>	(result = this.result) == null)
5583	>	throw new NullPointerException();
5584	>	for (int b;;) {
5585	>	if (result.get() != null)
5586	>	return;
5587	>	if ((b = preSplit()) <= 0)
5588	>	break;
5589	>	new SearchKeysTask<K,V,U>
5590	>	(map, this, b, searchFunction, result).fork();
5591	>	}
5592	>	while (result.get() == null) {
5593	>	U u;
5594	>	if (advance() == null) {
5595	>	propagateCompletion();
5596	>	break;
5597	>	}
5598	>	if ((u = searchFunction.apply((K)nextKey)) != null) {
5599	>	if (result.compareAndSet(null, u))
5600	>	quietlyCompleteRoot();
5601	>	break;
5602		}
1420	–	} finally {
1421	–	seg.unlock();
5603		}
5604		}
1424	–	s.writeObject(null);
1425	–	s.writeObject(null);
5605		}
5606
5607	<	/**
5608	<	* Reconstitutes the <tt>ConcurrentHashMap</tt> instance from a
5609	<	* stream (i.e., deserializes it).
5610	<	* @param s the stream
5611	<	*/
5612	<	@SuppressWarnings("unchecked")
5613	<	private void readObject(java.io.ObjectInputStream s)
5614	<	throws java.io.IOException, ClassNotFoundException {
5615	<	s.defaultReadObject();
5607	>	@SuppressWarnings("serial") static final class SearchValuesTask<K,V,U>
5608	>	extends Traverser<K,V,U> {
5609	>	final Fun<? super V, ? extends U> searchFunction;
5610	>	final AtomicReference<U> result;
5611	>	SearchValuesTask
5612	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5613	>	Fun<? super V, ? extends U> searchFunction,
5614	>	AtomicReference<U> result) {
5615	>	super(m, p, b);
5616	>	this.searchFunction = searchFunction; this.result = result;
5617	>	}
5618	>	public final U getRawResult() { return result.get(); }
5619	>	@SuppressWarnings("unchecked") public final void compute() {
5620	>	final Fun<? super V, ? extends U> searchFunction;
5621	>	final AtomicReference<U> result;
5622	>	if ((searchFunction = this.searchFunction) == null ||
5623	>	(result = this.result) == null)
5624	>	throw new NullPointerException();
5625	>	for (int b;;) {
5626	>	if (result.get() != null)
5627	>	return;
5628	>	if ((b = preSplit()) <= 0)
5629	>	break;
5630	>	new SearchValuesTask<K,V,U>
5631	>	(map, this, b, searchFunction, result).fork();
5632	>	}
5633	>	while (result.get() == null) {
5634	>	Object v; U u;
5635	>	if ((v = advance()) == null) {
5636	>	propagateCompletion();
5637	>	break;
5638	>	}
5639	>	if ((u = searchFunction.apply((V)v)) != null) {
5640	>	if (result.compareAndSet(null, u))
5641	>	quietlyCompleteRoot();
5642	>	break;
5643	>	}
5644	>	}
5645	>	}
5646	>	}
5647
5648	<	// Re-initialize segments to be minimally sized, and let grow.
5649	<	int cap = MIN_SEGMENT_TABLE_CAPACITY;
5650	<	final Segment<K,V>[] segments = this.segments;
5651	<	for (int k = 0; k < segments.length; ++k) {
5652	<	Segment<K,V> seg = segments[k];
5653	<	if (seg != null) {
5654	<	seg.threshold = (int)(cap * seg.loadFactor);
5655	<	seg.table = (HashEntry<K,V>[]) new HashEntry<?,?>[cap];
5648	>	@SuppressWarnings("serial") static final class SearchEntriesTask<K,V,U>
5649	>	extends Traverser<K,V,U> {
5650	>	final Fun<Entry<K,V>, ? extends U> searchFunction;
5651	>	final AtomicReference<U> result;
5652	>	SearchEntriesTask
5653	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5654	>	Fun<Entry<K,V>, ? extends U> searchFunction,
5655	>	AtomicReference<U> result) {
5656	>	super(m, p, b);
5657	>	this.searchFunction = searchFunction; this.result = result;
5658	>	}
5659	>	public final U getRawResult() { return result.get(); }
5660	>	@SuppressWarnings("unchecked") public final void compute() {
5661	>	final Fun<Entry<K,V>, ? extends U> searchFunction;
5662	>	final AtomicReference<U> result;
5663	>	if ((searchFunction = this.searchFunction) == null ||
5664	>	(result = this.result) == null)
5665	>	throw new NullPointerException();
5666	>	for (int b;;) {
5667	>	if (result.get() != null)
5668	>	return;
5669	>	if ((b = preSplit()) <= 0)
5670	>	break;
5671	>	new SearchEntriesTask<K,V,U>
5672	>	(map, this, b, searchFunction, result).fork();
5673	>	}
5674	>	while (result.get() == null) {
5675	>	Object v; U u;
5676	>	if ((v = advance()) == null) {
5677	>	propagateCompletion();
5678	>	break;
5679	>	}
5680	>	if ((u = searchFunction.apply(entryFor((K)nextKey, (V)v))) != null) {
5681	>	if (result.compareAndSet(null, u))
5682	>	quietlyCompleteRoot();
5683	>	return;
5684	>	}
5685		}
5686		}
5687	+	}
5688
5689	<	// Read the keys and values, and put the mappings in the table
5690	<	for (;;) {
5691	<	K key = (K) s.readObject();
5692	<	V value = (V) s.readObject();
5693	<	if (key == null)
5694	<	break;
5695	<	put(key, value);
5689	>	@SuppressWarnings("serial") static final class SearchMappingsTask<K,V,U>
5690	>	extends Traverser<K,V,U> {
5691	>	final BiFun<? super K, ? super V, ? extends U> searchFunction;
5692	>	final AtomicReference<U> result;
5693	>	SearchMappingsTask
5694	>	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5695	>	BiFun<? super K, ? super V, ? extends U> searchFunction,
5696	>	AtomicReference<U> result) {
5697	>	super(m, p, b);
5698	>	this.searchFunction = searchFunction; this.result = result;
5699	>	}
5700	>	public final U getRawResult() { return result.get(); }
5701	>	@SuppressWarnings("unchecked") public final void compute() {
5702	>	final BiFun<? super K, ? super V, ? extends U> searchFunction;
5703	>	final AtomicReference<U> result;
5704	>	if ((searchFunction = this.searchFunction) == null ||
5705	>	(result = this.result) == null)
5706	>	throw new NullPointerException();
5707	>	for (int b;;) {
5708	>	if (result.get() != null)
5709	>	return;
5710	>	if ((b = preSplit()) <= 0)
5711	>	break;
5712	>	new SearchMappingsTask<K,V,U>
5713	>	(map, this, b, searchFunction, result).fork();
5714	>	}
5715	>	while (result.get() == null) {
5716	>	Object v; U u;
5717	>	if ((v = advance()) == null) {
5718	>	propagateCompletion();
5719	>	break;
5720	>	}
5721	>	if ((u = searchFunction.apply((K)nextKey, (V)v)) != null) {
5722	>	if (result.compareAndSet(null, u))
5723	>	quietlyCompleteRoot();
5724	>	break;
5725	>	}
5726	>	}
5727		}
5728		}
5729
5730	+	@SuppressWarnings("serial") static final class ReduceKeysTask<K,V>
5731	+	extends Traverser<K,V,K> {
5732	+	final BiFun<? super K, ? super K, ? extends K> reducer;
5733	+	K result;
5734	+	ReduceKeysTask<K,V> rights, nextRight;
5735	+	ReduceKeysTask
5736	+	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5737	+	ReduceKeysTask<K,V> nextRight,
5738	+	BiFun<? super K, ? super K, ? extends K> reducer) {
5739	+	super(m, p, b); this.nextRight = nextRight;
5740	+	this.reducer = reducer;
5741	+	}
5742	+	public final K getRawResult() { return result; }
5743	+	@SuppressWarnings("unchecked") public final void compute() {
5744	+	final BiFun<? super K, ? super K, ? extends K> reducer =
5745	+	this.reducer;
5746	+	if (reducer == null)
5747	+	throw new NullPointerException();
5748	+	for (int b; (b = preSplit()) > 0;)
5749	+	(rights = new ReduceKeysTask<K,V>
5750	+	(map, this, b, rights, reducer)).fork();
5751	+	K r = null;
5752	+	while (advance() != null) {
5753	+	K u = (K)nextKey;
5754	+	r = (r == null) ? u : reducer.apply(r, u);
5755	+	}
5756	+	result = r;
5757	+	CountedCompleter<?> c;
5758	+	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5759	+	ReduceKeysTask<K,V>
5760	+	t = (ReduceKeysTask<K,V>)c,
5761	+	s = t.rights;
5762	+	while (s != null) {
5763	+	K tr, sr;
5764	+	if ((sr = s.result) != null)
5765	+	t.result = (((tr = t.result) == null) ? sr :
5766	+	reducer.apply(tr, sr));
5767	+	s = t.rights = s.nextRight;
5768	+	}
5769	+	}
5770	+	}
5771	+	}
5772	+
5773	+	@SuppressWarnings("serial") static final class ReduceValuesTask<K,V>
5774	+	extends Traverser<K,V,V> {
5775	+	final BiFun<? super V, ? super V, ? extends V> reducer;
5776	+	V result;
5777	+	ReduceValuesTask<K,V> rights, nextRight;
5778	+	ReduceValuesTask
5779	+	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5780	+	ReduceValuesTask<K,V> nextRight,
5781	+	BiFun<? super V, ? super V, ? extends V> reducer) {
5782	+	super(m, p, b); this.nextRight = nextRight;
5783	+	this.reducer = reducer;
5784	+	}
5785	+	public final V getRawResult() { return result; }
5786	+	@SuppressWarnings("unchecked") public final void compute() {
5787	+	final BiFun<? super V, ? super V, ? extends V> reducer =
5788	+	this.reducer;
5789	+	if (reducer == null)
5790	+	throw new NullPointerException();
5791	+	for (int b; (b = preSplit()) > 0;)
5792	+	(rights = new ReduceValuesTask<K,V>
5793	+	(map, this, b, rights, reducer)).fork();
5794	+	V r = null;
5795	+	Object v;
5796	+	while ((v = advance()) != null) {
5797	+	V u = (V)v;
5798	+	r = (r == null) ? u : reducer.apply(r, u);
5799	+	}
5800	+	result = r;
5801	+	CountedCompleter<?> c;
5802	+	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5803	+	ReduceValuesTask<K,V>
5804	+	t = (ReduceValuesTask<K,V>)c,
5805	+	s = t.rights;
5806	+	while (s != null) {
5807	+	V tr, sr;
5808	+	if ((sr = s.result) != null)
5809	+	t.result = (((tr = t.result) == null) ? sr :
5810	+	reducer.apply(tr, sr));
5811	+	s = t.rights = s.nextRight;
5812	+	}
5813	+	}
5814	+	}
5815	+	}
5816	+
5817	+	@SuppressWarnings("serial") static final class ReduceEntriesTask<K,V>
5818	+	extends Traverser<K,V,Map.Entry<K,V>> {
5819	+	final BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer;
5820	+	Map.Entry<K,V> result;
5821	+	ReduceEntriesTask<K,V> rights, nextRight;
5822	+	ReduceEntriesTask
5823	+	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5824	+	ReduceEntriesTask<K,V> nextRight,
5825	+	BiFun<Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer) {
5826	+	super(m, p, b); this.nextRight = nextRight;
5827	+	this.reducer = reducer;
5828	+	}
5829	+	public final Map.Entry<K,V> getRawResult() { return result; }
5830	+	@SuppressWarnings("unchecked") public final void compute() {
5831	+	final BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer =
5832	+	this.reducer;
5833	+	if (reducer == null)
5834	+	throw new NullPointerException();
5835	+	for (int b; (b = preSplit()) > 0;)
5836	+	(rights = new ReduceEntriesTask<K,V>
5837	+	(map, this, b, rights, reducer)).fork();
5838	+	Map.Entry<K,V> r = null;
5839	+	Object v;
5840	+	while ((v = advance()) != null) {
5841	+	Map.Entry<K,V> u = entryFor((K)nextKey, (V)v);
5842	+	r = (r == null) ? u : reducer.apply(r, u);
5843	+	}
5844	+	result = r;
5845	+	CountedCompleter<?> c;
5846	+	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5847	+	ReduceEntriesTask<K,V>
5848	+	t = (ReduceEntriesTask<K,V>)c,
5849	+	s = t.rights;
5850	+	while (s != null) {
5851	+	Map.Entry<K,V> tr, sr;
5852	+	if ((sr = s.result) != null)
5853	+	t.result = (((tr = t.result) == null) ? sr :
5854	+	reducer.apply(tr, sr));
5855	+	s = t.rights = s.nextRight;
5856	+	}
5857	+	}
5858	+	}
5859	+	}
5860	+
5861	+	@SuppressWarnings("serial") static final class MapReduceKeysTask<K,V,U>
5862	+	extends Traverser<K,V,U> {
5863	+	final Fun<? super K, ? extends U> transformer;
5864	+	final BiFun<? super U, ? super U, ? extends U> reducer;
5865	+	U result;
5866	+	MapReduceKeysTask<K,V,U> rights, nextRight;
5867	+	MapReduceKeysTask
5868	+	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5869	+	MapReduceKeysTask<K,V,U> nextRight,
5870	+	Fun<? super K, ? extends U> transformer,
5871	+	BiFun<? super U, ? super U, ? extends U> reducer) {
5872	+	super(m, p, b); this.nextRight = nextRight;
5873	+	this.transformer = transformer;
5874	+	this.reducer = reducer;
5875	+	}
5876	+	public final U getRawResult() { return result; }
5877	+	@SuppressWarnings("unchecked") public final void compute() {
5878	+	final Fun<? super K, ? extends U> transformer =
5879	+	this.transformer;
5880	+	final BiFun<? super U, ? super U, ? extends U> reducer =
5881	+	this.reducer;
5882	+	if (transformer == null || reducer == null)
5883	+	throw new NullPointerException();
5884	+	for (int b; (b = preSplit()) > 0;)
5885	+	(rights = new MapReduceKeysTask<K,V,U>
5886	+	(map, this, b, rights, transformer, reducer)).fork();
5887	+	U r = null, u;
5888	+	while (advance() != null) {
5889	+	if ((u = transformer.apply((K)nextKey)) != null)
5890	+	r = (r == null) ? u : reducer.apply(r, u);
5891	+	}
5892	+	result = r;
5893	+	CountedCompleter<?> c;
5894	+	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5895	+	MapReduceKeysTask<K,V,U>
5896	+	t = (MapReduceKeysTask<K,V,U>)c,
5897	+	s = t.rights;
5898	+	while (s != null) {
5899	+	U tr, sr;
5900	+	if ((sr = s.result) != null)
5901	+	t.result = (((tr = t.result) == null) ? sr :
5902	+	reducer.apply(tr, sr));
5903	+	s = t.rights = s.nextRight;
5904	+	}
5905	+	}
5906	+	}
5907	+	}
5908	+
5909	+	@SuppressWarnings("serial") static final class MapReduceValuesTask<K,V,U>
5910	+	extends Traverser<K,V,U> {
5911	+	final Fun<? super V, ? extends U> transformer;
5912	+	final BiFun<? super U, ? super U, ? extends U> reducer;
5913	+	U result;
5914	+	MapReduceValuesTask<K,V,U> rights, nextRight;
5915	+	MapReduceValuesTask
5916	+	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5917	+	MapReduceValuesTask<K,V,U> nextRight,
5918	+	Fun<? super V, ? extends U> transformer,
5919	+	BiFun<? super U, ? super U, ? extends U> reducer) {
5920	+	super(m, p, b); this.nextRight = nextRight;
5921	+	this.transformer = transformer;
5922	+	this.reducer = reducer;
5923	+	}
5924	+	public final U getRawResult() { return result; }
5925	+	@SuppressWarnings("unchecked") public final void compute() {
5926	+	final Fun<? super V, ? extends U> transformer =
5927	+	this.transformer;
5928	+	final BiFun<? super U, ? super U, ? extends U> reducer =
5929	+	this.reducer;
5930	+	if (transformer == null || reducer == null)
5931	+	throw new NullPointerException();
5932	+	for (int b; (b = preSplit()) > 0;)
5933	+	(rights = new MapReduceValuesTask<K,V,U>
5934	+	(map, this, b, rights, transformer, reducer)).fork();
5935	+	U r = null, u;
5936	+	Object v;
5937	+	while ((v = advance()) != null) {
5938	+	if ((u = transformer.apply((V)v)) != null)
5939	+	r = (r == null) ? u : reducer.apply(r, u);
5940	+	}
5941	+	result = r;
5942	+	CountedCompleter<?> c;
5943	+	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5944	+	MapReduceValuesTask<K,V,U>
5945	+	t = (MapReduceValuesTask<K,V,U>)c,
5946	+	s = t.rights;
5947	+	while (s != null) {
5948	+	U tr, sr;
5949	+	if ((sr = s.result) != null)
5950	+	t.result = (((tr = t.result) == null) ? sr :
5951	+	reducer.apply(tr, sr));
5952	+	s = t.rights = s.nextRight;
5953	+	}
5954	+	}
5955	+	}
5956	+	}
5957	+
5958	+	@SuppressWarnings("serial") static final class MapReduceEntriesTask<K,V,U>
5959	+	extends Traverser<K,V,U> {
5960	+	final Fun<Map.Entry<K,V>, ? extends U> transformer;
5961	+	final BiFun<? super U, ? super U, ? extends U> reducer;
5962	+	U result;
5963	+	MapReduceEntriesTask<K,V,U> rights, nextRight;
5964	+	MapReduceEntriesTask
5965	+	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
5966	+	MapReduceEntriesTask<K,V,U> nextRight,
5967	+	Fun<Map.Entry<K,V>, ? extends U> transformer,
5968	+	BiFun<? super U, ? super U, ? extends U> reducer) {
5969	+	super(m, p, b); this.nextRight = nextRight;
5970	+	this.transformer = transformer;
5971	+	this.reducer = reducer;
5972	+	}
5973	+	public final U getRawResult() { return result; }
5974	+	@SuppressWarnings("unchecked") public final void compute() {
5975	+	final Fun<Map.Entry<K,V>, ? extends U> transformer =
5976	+	this.transformer;
5977	+	final BiFun<? super U, ? super U, ? extends U> reducer =
5978	+	this.reducer;
5979	+	if (transformer == null || reducer == null)
5980	+	throw new NullPointerException();
5981	+	for (int b; (b = preSplit()) > 0;)
5982	+	(rights = new MapReduceEntriesTask<K,V,U>
5983	+	(map, this, b, rights, transformer, reducer)).fork();
5984	+	U r = null, u;
5985	+	Object v;
5986	+	while ((v = advance()) != null) {
5987	+	if ((u = transformer.apply(entryFor((K)nextKey, (V)v))) != null)
5988	+	r = (r == null) ? u : reducer.apply(r, u);
5989	+	}
5990	+	result = r;
5991	+	CountedCompleter<?> c;
5992	+	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5993	+	MapReduceEntriesTask<K,V,U>
5994	+	t = (MapReduceEntriesTask<K,V,U>)c,
5995	+	s = t.rights;
5996	+	while (s != null) {
5997	+	U tr, sr;
5998	+	if ((sr = s.result) != null)
5999	+	t.result = (((tr = t.result) == null) ? sr :
6000	+	reducer.apply(tr, sr));
6001	+	s = t.rights = s.nextRight;
6002	+	}
6003	+	}
6004	+	}
6005	+	}
6006	+
6007	+	@SuppressWarnings("serial") static final class MapReduceMappingsTask<K,V,U>
6008	+	extends Traverser<K,V,U> {
6009	+	final BiFun<? super K, ? super V, ? extends U> transformer;
6010	+	final BiFun<? super U, ? super U, ? extends U> reducer;
6011	+	U result;
6012	+	MapReduceMappingsTask<K,V,U> rights, nextRight;
6013	+	MapReduceMappingsTask
6014	+	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6015	+	MapReduceMappingsTask<K,V,U> nextRight,
6016	+	BiFun<? super K, ? super V, ? extends U> transformer,
6017	+	BiFun<? super U, ? super U, ? extends U> reducer) {
6018	+	super(m, p, b); this.nextRight = nextRight;
6019	+	this.transformer = transformer;
6020	+	this.reducer = reducer;
6021	+	}
6022	+	public final U getRawResult() { return result; }
6023	+	@SuppressWarnings("unchecked") public final void compute() {
6024	+	final BiFun<? super K, ? super V, ? extends U> transformer =
6025	+	this.transformer;
6026	+	final BiFun<? super U, ? super U, ? extends U> reducer =
6027	+	this.reducer;
6028	+	if (transformer == null || reducer == null)
6029	+	throw new NullPointerException();
6030	+	for (int b; (b = preSplit()) > 0;)
6031	+	(rights = new MapReduceMappingsTask<K,V,U>
6032	+	(map, this, b, rights, transformer, reducer)).fork();
6033	+	U r = null, u;
6034	+	Object v;
6035	+	while ((v = advance()) != null) {
6036	+	if ((u = transformer.apply((K)nextKey, (V)v)) != null)
6037	+	r = (r == null) ? u : reducer.apply(r, u);
6038	+	}
6039	+	result = r;
6040	+	CountedCompleter<?> c;
6041	+	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6042	+	MapReduceMappingsTask<K,V,U>
6043	+	t = (MapReduceMappingsTask<K,V,U>)c,
6044	+	s = t.rights;
6045	+	while (s != null) {
6046	+	U tr, sr;
6047	+	if ((sr = s.result) != null)
6048	+	t.result = (((tr = t.result) == null) ? sr :
6049	+	reducer.apply(tr, sr));
6050	+	s = t.rights = s.nextRight;
6051	+	}
6052	+	}
6053	+	}
6054	+	}
6055	+
6056	+	@SuppressWarnings("serial") static final class MapReduceKeysToDoubleTask<K,V>
6057	+	extends Traverser<K,V,Double> {
6058	+	final ObjectToDouble<? super K> transformer;
6059	+	final DoubleByDoubleToDouble reducer;
6060	+	final double basis;
6061	+	double result;
6062	+	MapReduceKeysToDoubleTask<K,V> rights, nextRight;
6063	+	MapReduceKeysToDoubleTask
6064	+	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6065	+	MapReduceKeysToDoubleTask<K,V> nextRight,
6066	+	ObjectToDouble<? super K> transformer,
6067	+	double basis,
6068	+	DoubleByDoubleToDouble reducer) {
6069	+	super(m, p, b); this.nextRight = nextRight;
6070	+	this.transformer = transformer;
6071	+	this.basis = basis; this.reducer = reducer;
6072	+	}
6073	+	public final Double getRawResult() { return result; }
6074	+	@SuppressWarnings("unchecked") public final void compute() {
6075	+	final ObjectToDouble<? super K> transformer =
6076	+	this.transformer;
6077	+	final DoubleByDoubleToDouble reducer = this.reducer;
6078	+	if (transformer == null || reducer == null)
6079	+	throw new NullPointerException();
6080	+	double r = this.basis;
6081	+	for (int b; (b = preSplit()) > 0;)
6082	+	(rights = new MapReduceKeysToDoubleTask<K,V>
6083	+	(map, this, b, rights, transformer, r, reducer)).fork();
6084	+	while (advance() != null)
6085	+	r = reducer.apply(r, transformer.apply((K)nextKey));
6086	+	result = r;
6087	+	CountedCompleter<?> c;
6088	+	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6089	+	MapReduceKeysToDoubleTask<K,V>
6090	+	t = (MapReduceKeysToDoubleTask<K,V>)c,
6091	+	s = t.rights;
6092	+	while (s != null) {
6093	+	t.result = reducer.apply(t.result, s.result);
6094	+	s = t.rights = s.nextRight;
6095	+	}
6096	+	}
6097	+	}
6098	+	}
6099	+
6100	+	@SuppressWarnings("serial") static final class MapReduceValuesToDoubleTask<K,V>
6101	+	extends Traverser<K,V,Double> {
6102	+	final ObjectToDouble<? super V> transformer;
6103	+	final DoubleByDoubleToDouble reducer;
6104	+	final double basis;
6105	+	double result;
6106	+	MapReduceValuesToDoubleTask<K,V> rights, nextRight;
6107	+	MapReduceValuesToDoubleTask
6108	+	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6109	+	MapReduceValuesToDoubleTask<K,V> nextRight,
6110	+	ObjectToDouble<? super V> transformer,
6111	+	double basis,
6112	+	DoubleByDoubleToDouble reducer) {
6113	+	super(m, p, b); this.nextRight = nextRight;
6114	+	this.transformer = transformer;
6115	+	this.basis = basis; this.reducer = reducer;
6116	+	}
6117	+	public final Double getRawResult() { return result; }
6118	+	@SuppressWarnings("unchecked") public final void compute() {
6119	+	final ObjectToDouble<? super V> transformer =
6120	+	this.transformer;
6121	+	final DoubleByDoubleToDouble reducer = this.reducer;
6122	+	if (transformer == null || reducer == null)
6123	+	throw new NullPointerException();
6124	+	double r = this.basis;
6125	+	for (int b; (b = preSplit()) > 0;)
6126	+	(rights = new MapReduceValuesToDoubleTask<K,V>
6127	+	(map, this, b, rights, transformer, r, reducer)).fork();
6128	+	Object v;
6129	+	while ((v = advance()) != null)
6130	+	r = reducer.apply(r, transformer.apply((V)v));
6131	+	result = r;
6132	+	CountedCompleter<?> c;
6133	+	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6134	+	MapReduceValuesToDoubleTask<K,V>
6135	+	t = (MapReduceValuesToDoubleTask<K,V>)c,
6136	+	s = t.rights;
6137	+	while (s != null) {
6138	+	t.result = reducer.apply(t.result, s.result);
6139	+	s = t.rights = s.nextRight;
6140	+	}
6141	+	}
6142	+	}
6143	+	}
6144	+
6145	+	@SuppressWarnings("serial") static final class MapReduceEntriesToDoubleTask<K,V>
6146	+	extends Traverser<K,V,Double> {
6147	+	final ObjectToDouble<Map.Entry<K,V>> transformer;
6148	+	final DoubleByDoubleToDouble reducer;
6149	+	final double basis;
6150	+	double result;
6151	+	MapReduceEntriesToDoubleTask<K,V> rights, nextRight;
6152	+	MapReduceEntriesToDoubleTask
6153	+	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6154	+	MapReduceEntriesToDoubleTask<K,V> nextRight,
6155	+	ObjectToDouble<Map.Entry<K,V>> transformer,
6156	+	double basis,
6157	+	DoubleByDoubleToDouble reducer) {
6158	+	super(m, p, b); this.nextRight = nextRight;
6159	+	this.transformer = transformer;
6160	+	this.basis = basis; this.reducer = reducer;
6161	+	}
6162	+	public final Double getRawResult() { return result; }
6163	+	@SuppressWarnings("unchecked") public final void compute() {
6164	+	final ObjectToDouble<Map.Entry<K,V>> transformer =
6165	+	this.transformer;
6166	+	final DoubleByDoubleToDouble reducer = this.reducer;
6167	+	if (transformer == null || reducer == null)
6168	+	throw new NullPointerException();
6169	+	double r = this.basis;
6170	+	for (int b; (b = preSplit()) > 0;)
6171	+	(rights = new MapReduceEntriesToDoubleTask<K,V>
6172	+	(map, this, b, rights, transformer, r, reducer)).fork();
6173	+	Object v;
6174	+	while ((v = advance()) != null)
6175	+	r = reducer.apply(r, transformer.apply(entryFor((K)nextKey, (V)v)));
6176	+	result = r;
6177	+	CountedCompleter<?> c;
6178	+	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6179	+	MapReduceEntriesToDoubleTask<K,V>
6180	+	t = (MapReduceEntriesToDoubleTask<K,V>)c,
6181	+	s = t.rights;
6182	+	while (s != null) {
6183	+	t.result = reducer.apply(t.result, s.result);
6184	+	s = t.rights = s.nextRight;
6185	+	}
6186	+	}
6187	+	}
6188	+	}
6189	+
6190	+	@SuppressWarnings("serial") static final class MapReduceMappingsToDoubleTask<K,V>
6191	+	extends Traverser<K,V,Double> {
6192	+	final ObjectByObjectToDouble<? super K, ? super V> transformer;
6193	+	final DoubleByDoubleToDouble reducer;
6194	+	final double basis;
6195	+	double result;
6196	+	MapReduceMappingsToDoubleTask<K,V> rights, nextRight;
6197	+	MapReduceMappingsToDoubleTask
6198	+	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6199	+	MapReduceMappingsToDoubleTask<K,V> nextRight,
6200	+	ObjectByObjectToDouble<? super K, ? super V> transformer,
6201	+	double basis,
6202	+	DoubleByDoubleToDouble reducer) {
6203	+	super(m, p, b); this.nextRight = nextRight;
6204	+	this.transformer = transformer;
6205	+	this.basis = basis; this.reducer = reducer;
6206	+	}
6207	+	public final Double getRawResult() { return result; }
6208	+	@SuppressWarnings("unchecked") public final void compute() {
6209	+	final ObjectByObjectToDouble<? super K, ? super V> transformer =
6210	+	this.transformer;
6211	+	final DoubleByDoubleToDouble reducer = this.reducer;
6212	+	if (transformer == null || reducer == null)
6213	+	throw new NullPointerException();
6214	+	double r = this.basis;
6215	+	for (int b; (b = preSplit()) > 0;)
6216	+	(rights = new MapReduceMappingsToDoubleTask<K,V>
6217	+	(map, this, b, rights, transformer, r, reducer)).fork();
6218	+	Object v;
6219	+	while ((v = advance()) != null)
6220	+	r = reducer.apply(r, transformer.apply((K)nextKey, (V)v));
6221	+	result = r;
6222	+	CountedCompleter<?> c;
6223	+	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6224	+	MapReduceMappingsToDoubleTask<K,V>
6225	+	t = (MapReduceMappingsToDoubleTask<K,V>)c,
6226	+	s = t.rights;
6227	+	while (s != null) {
6228	+	t.result = reducer.apply(t.result, s.result);
6229	+	s = t.rights = s.nextRight;
6230	+	}
6231	+	}
6232	+	}
6233	+	}
6234	+
6235	+	@SuppressWarnings("serial") static final class MapReduceKeysToLongTask<K,V>
6236	+	extends Traverser<K,V,Long> {
6237	+	final ObjectToLong<? super K> transformer;
6238	+	final LongByLongToLong reducer;
6239	+	final long basis;
6240	+	long result;
6241	+	MapReduceKeysToLongTask<K,V> rights, nextRight;
6242	+	MapReduceKeysToLongTask
6243	+	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6244	+	MapReduceKeysToLongTask<K,V> nextRight,
6245	+	ObjectToLong<? super K> transformer,
6246	+	long basis,
6247	+	LongByLongToLong reducer) {
6248	+	super(m, p, b); this.nextRight = nextRight;
6249	+	this.transformer = transformer;
6250	+	this.basis = basis; this.reducer = reducer;
6251	+	}
6252	+	public final Long getRawResult() { return result; }
6253	+	@SuppressWarnings("unchecked") public final void compute() {
6254	+	final ObjectToLong<? super K> transformer =
6255	+	this.transformer;
6256	+	final LongByLongToLong reducer = this.reducer;
6257	+	if (transformer == null || reducer == null)
6258	+	throw new NullPointerException();
6259	+	long r = this.basis;
6260	+	for (int b; (b = preSplit()) > 0;)
6261	+	(rights = new MapReduceKeysToLongTask<K,V>
6262	+	(map, this, b, rights, transformer, r, reducer)).fork();
6263	+	while (advance() != null)
6264	+	r = reducer.apply(r, transformer.apply((K)nextKey));
6265	+	result = r;
6266	+	CountedCompleter<?> c;
6267	+	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6268	+	MapReduceKeysToLongTask<K,V>
6269	+	t = (MapReduceKeysToLongTask<K,V>)c,
6270	+	s = t.rights;
6271	+	while (s != null) {
6272	+	t.result = reducer.apply(t.result, s.result);
6273	+	s = t.rights = s.nextRight;
6274	+	}
6275	+	}
6276	+	}
6277	+	}
6278	+
6279	+	@SuppressWarnings("serial") static final class MapReduceValuesToLongTask<K,V>
6280	+	extends Traverser<K,V,Long> {
6281	+	final ObjectToLong<? super V> transformer;
6282	+	final LongByLongToLong reducer;
6283	+	final long basis;
6284	+	long result;
6285	+	MapReduceValuesToLongTask<K,V> rights, nextRight;
6286	+	MapReduceValuesToLongTask
6287	+	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6288	+	MapReduceValuesToLongTask<K,V> nextRight,
6289	+	ObjectToLong<? super V> transformer,
6290	+	long basis,
6291	+	LongByLongToLong reducer) {
6292	+	super(m, p, b); this.nextRight = nextRight;
6293	+	this.transformer = transformer;
6294	+	this.basis = basis; this.reducer = reducer;
6295	+	}
6296	+	public final Long getRawResult() { return result; }
6297	+	@SuppressWarnings("unchecked") public final void compute() {
6298	+	final ObjectToLong<? super V> transformer =
6299	+	this.transformer;
6300	+	final LongByLongToLong reducer = this.reducer;
6301	+	if (transformer == null || reducer == null)
6302	+	throw new NullPointerException();
6303	+	long r = this.basis;
6304	+	for (int b; (b = preSplit()) > 0;)
6305	+	(rights = new MapReduceValuesToLongTask<K,V>
6306	+	(map, this, b, rights, transformer, r, reducer)).fork();
6307	+	Object v;
6308	+	while ((v = advance()) != null)
6309	+	r = reducer.apply(r, transformer.apply((V)v));
6310	+	result = r;
6311	+	CountedCompleter<?> c;
6312	+	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6313	+	MapReduceValuesToLongTask<K,V>
6314	+	t = (MapReduceValuesToLongTask<K,V>)c,
6315	+	s = t.rights;
6316	+	while (s != null) {
6317	+	t.result = reducer.apply(t.result, s.result);
6318	+	s = t.rights = s.nextRight;
6319	+	}
6320	+	}
6321	+	}
6322	+	}
6323	+
6324	+	@SuppressWarnings("serial") static final class MapReduceEntriesToLongTask<K,V>
6325	+	extends Traverser<K,V,Long> {
6326	+	final ObjectToLong<Map.Entry<K,V>> transformer;
6327	+	final LongByLongToLong reducer;
6328	+	final long basis;
6329	+	long result;
6330	+	MapReduceEntriesToLongTask<K,V> rights, nextRight;
6331	+	MapReduceEntriesToLongTask
6332	+	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6333	+	MapReduceEntriesToLongTask<K,V> nextRight,
6334	+	ObjectToLong<Map.Entry<K,V>> transformer,
6335	+	long basis,
6336	+	LongByLongToLong reducer) {
6337	+	super(m, p, b); this.nextRight = nextRight;
6338	+	this.transformer = transformer;
6339	+	this.basis = basis; this.reducer = reducer;
6340	+	}
6341	+	public final Long getRawResult() { return result; }
6342	+	@SuppressWarnings("unchecked") public final void compute() {
6343	+	final ObjectToLong<Map.Entry<K,V>> transformer =
6344	+	this.transformer;
6345	+	final LongByLongToLong reducer = this.reducer;
6346	+	if (transformer == null || reducer == null)
6347	+	throw new NullPointerException();
6348	+	long r = this.basis;
6349	+	for (int b; (b = preSplit()) > 0;)
6350	+	(rights = new MapReduceEntriesToLongTask<K,V>
6351	+	(map, this, b, rights, transformer, r, reducer)).fork();
6352	+	Object v;
6353	+	while ((v = advance()) != null)
6354	+	r = reducer.apply(r, transformer.apply(entryFor((K)nextKey, (V)v)));
6355	+	result = r;
6356	+	CountedCompleter<?> c;
6357	+	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6358	+	MapReduceEntriesToLongTask<K,V>
6359	+	t = (MapReduceEntriesToLongTask<K,V>)c,
6360	+	s = t.rights;
6361	+	while (s != null) {
6362	+	t.result = reducer.apply(t.result, s.result);
6363	+	s = t.rights = s.nextRight;
6364	+	}
6365	+	}
6366	+	}
6367	+	}
6368	+
6369	+	@SuppressWarnings("serial") static final class MapReduceMappingsToLongTask<K,V>
6370	+	extends Traverser<K,V,Long> {
6371	+	final ObjectByObjectToLong<? super K, ? super V> transformer;
6372	+	final LongByLongToLong reducer;
6373	+	final long basis;
6374	+	long result;
6375	+	MapReduceMappingsToLongTask<K,V> rights, nextRight;
6376	+	MapReduceMappingsToLongTask
6377	+	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6378	+	MapReduceMappingsToLongTask<K,V> nextRight,
6379	+	ObjectByObjectToLong<? super K, ? super V> transformer,
6380	+	long basis,
6381	+	LongByLongToLong reducer) {
6382	+	super(m, p, b); this.nextRight = nextRight;
6383	+	this.transformer = transformer;
6384	+	this.basis = basis; this.reducer = reducer;
6385	+	}
6386	+	public final Long getRawResult() { return result; }
6387	+	@SuppressWarnings("unchecked") public final void compute() {
6388	+	final ObjectByObjectToLong<? super K, ? super V> transformer =
6389	+	this.transformer;
6390	+	final LongByLongToLong reducer = this.reducer;
6391	+	if (transformer == null || reducer == null)
6392	+	throw new NullPointerException();
6393	+	long r = this.basis;
6394	+	for (int b; (b = preSplit()) > 0;)
6395	+	(rights = new MapReduceMappingsToLongTask<K,V>
6396	+	(map, this, b, rights, transformer, r, reducer)).fork();
6397	+	Object v;
6398	+	while ((v = advance()) != null)
6399	+	r = reducer.apply(r, transformer.apply((K)nextKey, (V)v));
6400	+	result = r;
6401	+	CountedCompleter<?> c;
6402	+	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6403	+	MapReduceMappingsToLongTask<K,V>
6404	+	t = (MapReduceMappingsToLongTask<K,V>)c,
6405	+	s = t.rights;
6406	+	while (s != null) {
6407	+	t.result = reducer.apply(t.result, s.result);
6408	+	s = t.rights = s.nextRight;
6409	+	}
6410	+	}
6411	+	}
6412	+	}
6413	+
6414	+	@SuppressWarnings("serial") static final class MapReduceKeysToIntTask<K,V>
6415	+	extends Traverser<K,V,Integer> {
6416	+	final ObjectToInt<? super K> transformer;
6417	+	final IntByIntToInt reducer;
6418	+	final int basis;
6419	+	int result;
6420	+	MapReduceKeysToIntTask<K,V> rights, nextRight;
6421	+	MapReduceKeysToIntTask
6422	+	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6423	+	MapReduceKeysToIntTask<K,V> nextRight,
6424	+	ObjectToInt<? super K> transformer,
6425	+	int basis,
6426	+	IntByIntToInt reducer) {
6427	+	super(m, p, b); this.nextRight = nextRight;
6428	+	this.transformer = transformer;
6429	+	this.basis = basis; this.reducer = reducer;
6430	+	}
6431	+	public final Integer getRawResult() { return result; }
6432	+	@SuppressWarnings("unchecked") public final void compute() {
6433	+	final ObjectToInt<? super K> transformer =
6434	+	this.transformer;
6435	+	final IntByIntToInt reducer = this.reducer;
6436	+	if (transformer == null || reducer == null)
6437	+	throw new NullPointerException();
6438	+	int r = this.basis;
6439	+	for (int b; (b = preSplit()) > 0;)
6440	+	(rights = new MapReduceKeysToIntTask<K,V>
6441	+	(map, this, b, rights, transformer, r, reducer)).fork();
6442	+	while (advance() != null)
6443	+	r = reducer.apply(r, transformer.apply((K)nextKey));
6444	+	result = r;
6445	+	CountedCompleter<?> c;
6446	+	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6447	+	MapReduceKeysToIntTask<K,V>
6448	+	t = (MapReduceKeysToIntTask<K,V>)c,
6449	+	s = t.rights;
6450	+	while (s != null) {
6451	+	t.result = reducer.apply(t.result, s.result);
6452	+	s = t.rights = s.nextRight;
6453	+	}
6454	+	}
6455	+	}
6456	+	}
6457	+
6458	+	@SuppressWarnings("serial") static final class MapReduceValuesToIntTask<K,V>
6459	+	extends Traverser<K,V,Integer> {
6460	+	final ObjectToInt<? super V> transformer;
6461	+	final IntByIntToInt reducer;
6462	+	final int basis;
6463	+	int result;
6464	+	MapReduceValuesToIntTask<K,V> rights, nextRight;
6465	+	MapReduceValuesToIntTask
6466	+	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6467	+	MapReduceValuesToIntTask<K,V> nextRight,
6468	+	ObjectToInt<? super V> transformer,
6469	+	int basis,
6470	+	IntByIntToInt reducer) {
6471	+	super(m, p, b); this.nextRight = nextRight;
6472	+	this.transformer = transformer;
6473	+	this.basis = basis; this.reducer = reducer;
6474	+	}
6475	+	public final Integer getRawResult() { return result; }
6476	+	@SuppressWarnings("unchecked") public final void compute() {
6477	+	final ObjectToInt<? super V> transformer =
6478	+	this.transformer;
6479	+	final IntByIntToInt reducer = this.reducer;
6480	+	if (transformer == null || reducer == null)
6481	+	throw new NullPointerException();
6482	+	int r = this.basis;
6483	+	for (int b; (b = preSplit()) > 0;)
6484	+	(rights = new MapReduceValuesToIntTask<K,V>
6485	+	(map, this, b, rights, transformer, r, reducer)).fork();
6486	+	Object v;
6487	+	while ((v = advance()) != null)
6488	+	r = reducer.apply(r, transformer.apply((V)v));
6489	+	result = r;
6490	+	CountedCompleter<?> c;
6491	+	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6492	+	MapReduceValuesToIntTask<K,V>
6493	+	t = (MapReduceValuesToIntTask<K,V>)c,
6494	+	s = t.rights;
6495	+	while (s != null) {
6496	+	t.result = reducer.apply(t.result, s.result);
6497	+	s = t.rights = s.nextRight;
6498	+	}
6499	+	}
6500	+	}
6501	+	}
6502	+
6503	+	@SuppressWarnings("serial") static final class MapReduceEntriesToIntTask<K,V>
6504	+	extends Traverser<K,V,Integer> {
6505	+	final ObjectToInt<Map.Entry<K,V>> transformer;
6506	+	final IntByIntToInt reducer;
6507	+	final int basis;
6508	+	int result;
6509	+	MapReduceEntriesToIntTask<K,V> rights, nextRight;
6510	+	MapReduceEntriesToIntTask
6511	+	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6512	+	MapReduceEntriesToIntTask<K,V> nextRight,
6513	+	ObjectToInt<Map.Entry<K,V>> transformer,
6514	+	int basis,
6515	+	IntByIntToInt reducer) {
6516	+	super(m, p, b); this.nextRight = nextRight;
6517	+	this.transformer = transformer;
6518	+	this.basis = basis; this.reducer = reducer;
6519	+	}
6520	+	public final Integer getRawResult() { return result; }
6521	+	@SuppressWarnings("unchecked") public final void compute() {
6522	+	final ObjectToInt<Map.Entry<K,V>> transformer =
6523	+	this.transformer;
6524	+	final IntByIntToInt reducer = this.reducer;
6525	+	if (transformer == null || reducer == null)
6526	+	throw new NullPointerException();
6527	+	int r = this.basis;
6528	+	for (int b; (b = preSplit()) > 0;)
6529	+	(rights = new MapReduceEntriesToIntTask<K,V>
6530	+	(map, this, b, rights, transformer, r, reducer)).fork();
6531	+	Object v;
6532	+	while ((v = advance()) != null)
6533	+	r = reducer.apply(r, transformer.apply(entryFor((K)nextKey, (V)v)));
6534	+	result = r;
6535	+	CountedCompleter<?> c;
6536	+	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6537	+	MapReduceEntriesToIntTask<K,V>
6538	+	t = (MapReduceEntriesToIntTask<K,V>)c,
6539	+	s = t.rights;
6540	+	while (s != null) {
6541	+	t.result = reducer.apply(t.result, s.result);
6542	+	s = t.rights = s.nextRight;
6543	+	}
6544	+	}
6545	+	}
6546	+	}
6547	+
6548	+	@SuppressWarnings("serial") static final class MapReduceMappingsToIntTask<K,V>
6549	+	extends Traverser<K,V,Integer> {
6550	+	final ObjectByObjectToInt<? super K, ? super V> transformer;
6551	+	final IntByIntToInt reducer;
6552	+	final int basis;
6553	+	int result;
6554	+	MapReduceMappingsToIntTask<K,V> rights, nextRight;
6555	+	MapReduceMappingsToIntTask
6556	+	(ConcurrentHashMap<K,V> m, Traverser<K,V,?> p, int b,
6557	+	MapReduceMappingsToIntTask<K,V> nextRight,
6558	+	ObjectByObjectToInt<? super K, ? super V> transformer,
6559	+	int basis,
6560	+	IntByIntToInt reducer) {
6561	+	super(m, p, b); this.nextRight = nextRight;
6562	+	this.transformer = transformer;
6563	+	this.basis = basis; this.reducer = reducer;
6564	+	}
6565	+	public final Integer getRawResult() { return result; }
6566	+	@SuppressWarnings("unchecked") public final void compute() {
6567	+	final ObjectByObjectToInt<? super K, ? super V> transformer =
6568	+	this.transformer;
6569	+	final IntByIntToInt reducer = this.reducer;
6570	+	if (transformer == null || reducer == null)
6571	+	throw new NullPointerException();
6572	+	int r = this.basis;
6573	+	for (int b; (b = preSplit()) > 0;)
6574	+	(rights = new MapReduceMappingsToIntTask<K,V>
6575	+	(map, this, b, rights, transformer, r, reducer)).fork();
6576	+	Object v;
6577	+	while ((v = advance()) != null)
6578	+	r = reducer.apply(r, transformer.apply((K)nextKey, (V)v));
6579	+	result = r;
6580	+	CountedCompleter<?> c;
6581	+	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6582	+	MapReduceMappingsToIntTask<K,V>
6583	+	t = (MapReduceMappingsToIntTask<K,V>)c,
6584	+	s = t.rights;
6585	+	while (s != null) {
6586	+	t.result = reducer.apply(t.result, s.result);
6587	+	s = t.rights = s.nextRight;
6588	+	}
6589	+	}
6590	+	}
6591	+	}
6592	+
6593	+
6594		// Unsafe mechanics
6595		private static final sun.misc.Unsafe UNSAFE;
6596	<	private static final long SBASE;
6597	<	private static final int SSHIFT;
6598	<	private static final long TBASE;
6599	<	private static final int TSHIFT;
6596	>	private static final long counterOffset;
6597	>	private static final long sizeCtlOffset;
6598	>	private static final long ABASE;
6599	>	private static final int ASHIFT;
6600
6601		static {
6602	<	int ss, ts;
6602	>	int ss;
6603		try {
6604		UNSAFE = sun.misc.Unsafe.getUnsafe();
6605	<	Class<?> tc = HashEntry[].class;
6606	<	Class<?> sc = Segment[].class;
6607	<	TBASE = UNSAFE.arrayBaseOffset(tc);
6608	<	SBASE = UNSAFE.arrayBaseOffset(sc);
6609	<	ts = UNSAFE.arrayIndexScale(tc);
6605	>	Class<?> k = ConcurrentHashMap.class;
6606	>	counterOffset = UNSAFE.objectFieldOffset
6607	>	(k.getDeclaredField("counter"));
6608	>	sizeCtlOffset = UNSAFE.objectFieldOffset
6609	>	(k.getDeclaredField("sizeCtl"));
6610	>	Class<?> sc = Node[].class;
6611	>	ABASE = UNSAFE.arrayBaseOffset(sc);
6612		ss = UNSAFE.arrayIndexScale(sc);
6613		} catch (Exception e) {
6614		throw new Error(e);
6615		}
6616	<	if ((ss & (ss-1)) != 0 || (ts & (ts-1)) != 0)
6616	>	if ((ss & (ss-1)) != 0)
6617		throw new Error("data type scale not a power of two");
6618	<	SSHIFT = 31 - Integer.numberOfLeadingZeros(ss);
1482	<	TSHIFT = 31 - Integer.numberOfLeadingZeros(ts);
6618	>	ASHIFT = 31 - Integer.numberOfLeadingZeros(ss);
6619		}
1484	–
6620		}

Diff Legend

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