ViewVC Help

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

root/jsr166/jsr166/src/jsr166e/ConcurrentHashMapV8.java

Comparing jsr166/src/jsr166e/ConcurrentHashMapV8.java (file contents):
Revision 1.13 by jsr166, Wed Aug 31 00:22:11 2011 UTC vs.
Revision 1.80 by jsr166, Sat Nov 24 03:46:28 2012 UTC

#	Line 5 | Line 5
5		*/
6
7		package jsr166e;
8	<	import jsr166e.LongAdder;
8	>
9	>	import java.util.Comparator;
10	>	import java.util.Arrays;
11		import java.util.Map;
12		import java.util.Set;
13		import java.util.Collection;
#	Line 19 | Line 21 | import java.util.Enumeration;
21		import java.util.ConcurrentModificationException;
22		import java.util.NoSuchElementException;
23		import java.util.concurrent.ConcurrentMap;
24	+	import java.util.concurrent.ThreadLocalRandom;
25	+	import java.util.concurrent.locks.LockSupport;
26	+	import java.util.concurrent.locks.AbstractQueuedSynchronizer;
27	+	import java.util.concurrent.atomic.AtomicReference;
28	+
29	+	import java.io.Serializable;
30	+
31	+	import java.util.Comparator;
32	+	import java.util.Arrays;
33	+	import java.util.Map;
34	+	import java.util.Set;
35	+	import java.util.Collection;
36	+	import java.util.AbstractMap;
37	+	import java.util.AbstractSet;
38	+	import java.util.AbstractCollection;
39	+	import java.util.Hashtable;
40	+	import java.util.HashMap;
41	+	import java.util.Iterator;
42	+	import java.util.Enumeration;
43	+	import java.util.ConcurrentModificationException;
44	+	import java.util.NoSuchElementException;
45	+	import java.util.concurrent.ConcurrentMap;
46	+	import java.util.concurrent.ThreadLocalRandom;
47	+	import java.util.concurrent.locks.LockSupport;
48	+	import java.util.concurrent.locks.AbstractQueuedSynchronizer;
49	+	import java.util.concurrent.atomic.AtomicReference;
50	+
51		import java.io.Serializable;
52
53		/**
#	Line 33 | Line 62 | import java.io.Serializable;
62		* interoperable with {@code Hashtable} in programs that rely on its
63		* thread safety but not on its synchronization details.
64		*
65	<	* <p> Retrieval operations (including {@code get}) generally do not
65	>	* <p>Retrieval operations (including {@code get}) generally do not
66		* block, so may overlap with update operations (including {@code put}
67		* and {@code remove}). Retrievals reflect the results of the most
68		* recently <em>completed</em> update operations holding upon their
69	<	* onset.  For aggregate operations such as {@code putAll} and {@code
70	<	* clear}, concurrent retrievals may reflect insertion or removal of
71	<	* only some entries.  Similarly, Iterators and Enumerations return
72	<	* elements reflecting the state of the hash table at some point at or
73	<	* since the creation of the iterator/enumeration.  They do
74	<	* <em>not</em> throw {@link ConcurrentModificationException}.
75	<	* However, iterators are designed to be used by only one thread at a
76	<	* time.  Bear in mind that the results of aggregate status methods
77	<	* including {@code size}, {@code isEmpty}, and {@code containsValue}
78	<	* are typically useful only when a map is not undergoing concurrent
79	<	* updates in other threads.  Otherwise the results of these methods
80	<	* reflect transient states that may be adequate for monitoring
81	<	* purposes, but not for program control.
82	<	*
83	<	* <p> Resizing this or any other kind of hash table is a relatively
84	<	* slow operation, so, when possible, it is a good idea to provide
85	<	* estimates of expected table sizes in constructors. Also, for
86	<	* compatibility with previous versions of this class, constructors
87	<	* may optionally specify an expected {@code concurrencyLevel} as an
88	<	* additional hint for internal sizing.
69	>	* onset. (More formally, an update operation for a given key bears a
70	>	* <em>happens-before</em> relation with any (non-null) retrieval for
71	>	* that key reporting the updated value.)  For aggregate operations
72	>	* such as {@code putAll} and {@code clear}, concurrent retrievals may
73	>	* reflect insertion or removal of only some entries.  Similarly,
74	>	* Iterators and Enumerations return elements reflecting the state of
75	>	* the hash table at some point at or since the creation of the
76	>	* iterator/enumeration.  They do <em>not</em> throw {@link
77	>	* ConcurrentModificationException}.  However, iterators are designed
78	>	* to be used by only one thread at a time.  Bear in mind that the
79	>	* results of aggregate status methods including {@code size}, {@code
80	>	* isEmpty}, and {@code containsValue} are typically useful only when
81	>	* a map is not undergoing concurrent updates in other threads.
82	>	* Otherwise the results of these methods reflect transient states
83	>	* that may be adequate for monitoring or estimation purposes, but not
84	>	* for program control.
85	>	*
86	>	* <p>The table is dynamically expanded when there are too many
87	>	* collisions (i.e., keys that have distinct hash codes but fall into
88	>	* the same slot modulo the table size), with the expected average
89	>	* effect of maintaining roughly two bins per mapping (corresponding
90	>	* to a 0.75 load factor threshold for resizing). There may be much
91	>	* variance around this average as mappings are added and removed, but
92	>	* overall, this maintains a commonly accepted time/space tradeoff for
93	>	* hash tables.  However, resizing this or any other kind of hash
94	>	* table may be a relatively slow operation. When possible, it is a
95	>	* good idea to provide a size estimate as an optional {@code
96	>	* initialCapacity} constructor argument. An additional optional
97	>	* {@code loadFactor} constructor argument provides a further means of
98	>	* customizing initial table capacity by specifying the table density
99	>	* to be used in calculating the amount of space to allocate for the
100	>	* given number of elements.  Also, for compatibility with previous
101	>	* versions of this class, constructors may optionally specify an
102	>	* expected {@code concurrencyLevel} as an additional hint for
103	>	* internal sizing.  Note that using many keys with exactly the same
104	>	* {@code hashCode()} is a sure way to slow down performance of any
105	>	* hash table.
106	>	*
107	>	* <p>A {@link Set} projection of a ConcurrentHashMapV8 may be created
108	>	* (using {@link #newKeySet()} or {@link #newKeySet(int)}), or viewed
109	>	* (using {@link #keySet(Object)} when only keys are of interest, and the
110	>	* mapped values are (perhaps transiently) not used or all take the
111	>	* same mapping value.
112	>	*
113	>	* <p>A ConcurrentHashMapV8 can be used as scalable frequency map (a
114	>	* form of histogram or multiset) by using {@link LongAdder} values
115	>	* and initializing via {@link #computeIfAbsent}. For example, to add
116	>	* a count to a {@code ConcurrentHashMapV8<String,LongAdder> freqs}, you
117	>	* can use {@code freqs.computeIfAbsent(k -> new
118	>	* LongAdder()).increment();}
119		*
120		* <p>This class and its views and iterators implement all of the
121		* <em>optional</em> methods of the {@link Map} and {@link Iterator}
122		* interfaces.
123		*
124	<	* <p> Like {@link Hashtable} but unlike {@link HashMap}, this class
124	>	* <p>Like {@link Hashtable} but unlike {@link HashMap}, this class
125		* does <em>not</em> allow {@code null} to be used as a key or value.
126		*
127	+	* <p>ConcurrentHashMapV8s support parallel operations using the {@link
128	+	* ForkJoinPool#commonPool}. (Tasks that may be used in other contexts
129	+	* are available in class {@link ForkJoinTasks}). These operations are
130	+	* designed to be safely, and often sensibly, applied even with maps
131	+	* that are being concurrently updated by other threads; for example,
132	+	* when computing a snapshot summary of the values in a shared
133	+	* registry.  There are three kinds of operation, each with four
134	+	* forms, accepting functions with Keys, Values, Entries, and (Key,
135	+	* Value) arguments and/or return values. (The first three forms are
136	+	* also available via the {@link #keySet()}, {@link #values()} and
137	+	* {@link #entrySet()} views). Because the elements of a
138	+	* ConcurrentHashMapV8 are not ordered in any particular way, and may be
139	+	* processed in different orders in different parallel executions, the
140	+	* correctness of supplied functions should not depend on any
141	+	* ordering, or on any other objects or values that may transiently
142	+	* change while computation is in progress; and except for forEach
143	+	* actions, should ideally be side-effect-free.
144	+	*
145	+	* <ul>
146	+	* <li> forEach: Perform a given action on each element.
147	+	* A variant form applies a given transformation on each element
148	+	* before performing the action.</li>
149	+	*
150	+	* <li> search: Return the first available non-null result of
151	+	* applying a given function on each element; skipping further
152	+	* search when a result is found.</li>
153	+	*
154	+	* <li> reduce: Accumulate each element.  The supplied reduction
155	+	* function cannot rely on ordering (more formally, it should be
156	+	* both associative and commutative).  There are five variants:
157	+	*
158	+	* <ul>
159	+	*
160	+	* <li> Plain reductions. (There is not a form of this method for
161	+	* (key, value) function arguments since there is no corresponding
162	+	* return type.)</li>
163	+	*
164	+	* <li> Mapped reductions that accumulate the results of a given
165	+	* function applied to each element.</li>
166	+	*
167	+	* <li> Reductions to scalar doubles, longs, and ints, using a
168	+	* given basis value.</li>
169	+	*
170	+	* </li>
171	+	* </ul>
172	+	* </ul>
173	+	*
174	+	* <p>The concurrency properties of bulk operations follow
175	+	* from those of ConcurrentHashMapV8: Any non-null result returned
176	+	* from {@code get(key)} and related access methods bears a
177	+	* happens-before relation with the associated insertion or
178	+	* update.  The result of any bulk operation reflects the
179	+	* composition of these per-element relations (but is not
180	+	* necessarily atomic with respect to the map as a whole unless it
181	+	* is somehow known to be quiescent).  Conversely, because keys
182	+	* and values in the map are never null, null serves as a reliable
183	+	* atomic indicator of the current lack of any result.  To
184	+	* maintain this property, null serves as an implicit basis for
185	+	* all non-scalar reduction operations. For the double, long, and
186	+	* int versions, the basis should be one that, when combined with
187	+	* any other value, returns that other value (more formally, it
188	+	* should be the identity element for the reduction). Most common
189	+	* reductions have these properties; for example, computing a sum
190	+	* with basis 0 or a minimum with basis MAX_VALUE.
191	+	*
192	+	* <p>Search and transformation functions provided as arguments
193	+	* should similarly return null to indicate the lack of any result
194	+	* (in which case it is not used). In the case of mapped
195	+	* reductions, this also enables transformations to serve as
196	+	* filters, returning null (or, in the case of primitive
197	+	* specializations, the identity basis) if the element should not
198	+	* be combined. You can create compound transformations and
199	+	* filterings by composing them yourself under this "null means
200	+	* there is nothing there now" rule before using them in search or
201	+	* reduce operations.
202	+	*
203	+	* <p>Methods accepting and/or returning Entry arguments maintain
204	+	* key-value associations. They may be useful for example when
205	+	* finding the key for the greatest value. Note that "plain" Entry
206	+	* arguments can be supplied using {@code new
207	+	* AbstractMap.SimpleEntry(k,v)}.
208	+	*
209	+	* <p>Bulk operations may complete abruptly, throwing an
210	+	* exception encountered in the application of a supplied
211	+	* function. Bear in mind when handling such exceptions that other
212	+	* concurrently executing functions could also have thrown
213	+	* exceptions, or would have done so if the first exception had
214	+	* not occurred.
215	+	*
216	+	* <p>Parallel speedups for bulk operations compared to sequential
217	+	* processing are common but not guaranteed.  Operations involving
218	+	* brief functions on small maps may execute more slowly than
219	+	* sequential loops if the underlying work to parallelize the
220	+	* computation is more expensive than the computation itself.
221	+	* Similarly, parallelization may not lead to much actual parallelism
222	+	* if all processors are busy performing unrelated tasks.
223	+	*
224	+	* <p>All arguments to all task methods must be non-null.
225	+	*
226	+	* <p><em>jsr166e note: During transition, this class
227	+	* uses nested functional interfaces with different names but the
228	+	* same forms as those expected for JDK8.</em>
229	+	*
230		* <p>This class is a member of the
231		* <a href="{@docRoot}/../technotes/guides/collections/index.html">
232		* Java Collections Framework</a>.
233		*
72	–	* <p><em>jsr166e note: This class is a candidate replacement for
73	–	* java.util.concurrent.ConcurrentHashMap.<em>
74	–	*
234		* @since 1.5
235		* @author Doug Lea
236		* @param <K> the type of keys maintained by this map
237		* @param <V> the type of mapped values
238		*/
239		public class ConcurrentHashMapV8<K, V>
240	<	implements ConcurrentMap<K, V>, Serializable {
240	>	implements ConcurrentMap<K, V>, Serializable {
241		private static final long serialVersionUID = 7249069246763182397L;
242
243		/**
244	<	* A function computing a mapping from the given key to a value,
245	<	* or {@code null} if there is no mapping. This is a place-holder
246	<	* for an upcoming JDK8 interface.
247	<	*/
248	<	public static interface MappingFunction<K, V> {
249	<	/**
250	<	* Returns a value for the given key, or null if there is no
251	<	* mapping. If this function throws an (unchecked) exception,
252	<	* the exception is rethrown to its caller, and no mapping is
253	<	* recorded.  Because this function is invoked within
254	<	* atomicity control, the computation should be short and
255	<	* simple. The most common usage is to construct a new object
256	<	* serving as an initial mapped value.
244	>	* A partitionable iterator. A Spliterator can be traversed
245	>	* directly, but can also be partitioned (before traversal) by
246	>	* creating another Spliterator that covers a non-overlapping
247	>	* portion of the elements, and so may be amenable to parallel
248	>	* execution.
249	>	*
250	>	* <p>This interface exports a subset of expected JDK8
251	>	* functionality.
252	>	*
253	>	* <p>Sample usage: Here is one (of the several) ways to compute
254	>	* the sum of the values held in a map using the ForkJoin
255	>	* framework. As illustrated here, Spliterators are well suited to
256	>	* designs in which a task repeatedly splits off half its work
257	>	* into forked subtasks until small enough to process directly,
258	>	* and then joins these subtasks. Variants of this style can also
259	>	* be used in completion-based designs.
260	>	*
261	>	* <pre>
262	>	* {@code ConcurrentHashMapV8<String, Long> m = ...
263	>	* // split as if have 8 * parallelism, for load balance
264	>	* int n = m.size();
265	>	* int p = aForkJoinPool.getParallelism() * 8;
266	>	* int split = (n < p)? n : p;
267	>	* long sum = aForkJoinPool.invoke(new SumValues(m.valueSpliterator(), split, null));
268	>	* // ...
269	>	* static class SumValues extends RecursiveTask<Long> {
270	>	*   final Spliterator<Long> s;
271	>	*   final int split;             // split while > 1
272	>	*   final SumValues nextJoin;    // records forked subtasks to join
273	>	*   SumValues(Spliterator<Long> s, int depth, SumValues nextJoin) {
274	>	*     this.s = s; this.depth = depth; this.nextJoin = nextJoin;
275	>	*   }
276	>	*   public Long compute() {
277	>	*     long sum = 0;
278	>	*     SumValues subtasks = null; // fork subtasks
279	>	*     for (int s = split >>> 1; s > 0; s >>>= 1)
280	>	*       (subtasks = new SumValues(s.split(), s, subtasks)).fork();
281	>	*     while (s.hasNext())        // directly process remaining elements
282	>	*       sum += s.next();
283	>	*     for (SumValues t = subtasks; t != null; t = t.nextJoin)
284	>	*       sum += t.join();         // collect subtask results
285	>	*     return sum;
286	>	*   }
287	>	* }
288	>	* }</pre>
289	>	*/
290	>	public static interface Spliterator<T> extends Iterator<T> {
291	>	/**
292	>	* Returns a Spliterator covering approximately half of the
293	>	* elements, guaranteed not to overlap with those subsequently
294	>	* returned by this Spliterator.  After invoking this method,
295	>	* the current Spliterator will <em>not</em> produce any of
296	>	* the elements of the returned Spliterator, but the two
297	>	* Spliterators together will produce all of the elements that
298	>	* would have been produced by this Spliterator had this
299	>	* method not been called. The exact number of elements
300	>	* produced by the returned Spliterator is not guaranteed, and
301	>	* may be zero (i.e., with {@code hasNext()} reporting {@code
302	>	* false}) if this Spliterator cannot be further split.
303		*
304	<	* @param key the (non-null) key
305	<	* @return a value, or null if none
304	>	* @return a Spliterator covering approximately half of the
305	>	* elements
306	>	* @throws IllegalStateException if this Spliterator has
307	>	* already commenced traversing elements
308		*/
309	<	V map(K key);
309	>	Spliterator<T> split();
310		}
311
312	+
313		/*
314		* Overview:
315		*
316		* The primary design goal of this hash table is to maintain
317		* concurrent readability (typically method get(), but also
318		* iterators and related methods) while minimizing update
319	<	* contention.
319	>	* contention. Secondary goals are to keep space consumption about
320	>	* the same or better than java.util.HashMap, and to support high
321	>	* initial insertion rates on an empty table by many threads.
322		*
323		* Each key-value mapping is held in a Node.  Because Node fields
324		* can contain special values, they are defined using plain Object
325		* types. Similarly in turn, all internal methods that use them
326	<	* work off Object types. All public generic-typed methods relay
327	<	* in/out of these internal methods, supplying casts as needed.
326	>	* work off Object types. And similarly, so do the internal
327	>	* methods of auxiliary iterator and view classes.  All public
328	>	* generic typed methods relay in/out of these internal methods,
329	>	* supplying null-checks and casts as needed. This also allows
330	>	* many of the public methods to be factored into a smaller number
331	>	* of internal methods (although sadly not so for the five
332	>	* variants of put-related operations). The validation-based
333	>	* approach explained below leads to a lot of code sprawl because
334	>	* retry-control precludes factoring into smaller methods.
335		*
336		* The table is lazily initialized to a power-of-two size upon the
337	<	* first insertion.  Each bin in the table contains a (typically
338	<	* short) list of Nodes.  Table accesses require volatile/atomic
339	<	* reads, writes, and CASes.  Because there is no other way to
340	<	* arrange this without adding further indirections, we use
341	<	* intrinsics (sun.misc.Unsafe) operations.  The lists of nodes
342	<	* within bins are always accurately traversable under volatile
343	<	* reads, so long as lookups check hash code and non-nullness of
344	<	* key and value before checking key equality. (All valid hash
345	<	* codes are nonnegative. Negative values are reserved for special
346	<	* forwarding nodes; see below.)
347	<	*
348	<	* A bin may be locked during update (insert, delete, and replace)
349	<	* operations.  We do not want to waste the space required to
350	<	* associate a distinct lock object with each bin, so instead use
351	<	* the first node of a bin list itself as a lock, using builtin
352	<	* "synchronized" locks. These save space and we can live with
353	<	* only plain block-structured lock/unlock operations. Using the
354	<	* first node of a list as a lock does not by itself suffice
355	<	* though: When a node is locked, any update must first validate
356	<	* that it is still the first node, and retry if not. (Because new
357	<	* nodes are always appended to lists, once a node is first in a
358	<	* bin, it remains first until deleted or the bin becomes
359	<	* invalidated.)  However, update operations can and sometimes do
360	<	* still traverse the bin until the point of update, which helps
361	<	* reduce cache misses on retries.  This is a converse of sorts to
362	<	* the lazy locking technique described by Herlihy & Shavit. If
363	<	* there is no existing node during a put operation, then one can
364	<	* be CAS'ed in (without need for lock except in computeIfAbsent);
365	<	* the CAS serves as validation. This is on average the most
366	<	* common case for put operations -- under random hash codes, the
367	<	* distribution of nodes in bins follows a Poisson distribution
368	<	* (see http://en.wikipedia.org/wiki/Poisson_distribution) with a
369	<	* parameter of 0.5 on average under the default loadFactor of
370	<	* 0.75.  The expected number of locks covering different elements
371	<	* (i.e., bins with 2 or more nodes) is approximately 10% at
372	<	* steady state under default settings.  Lock contention
373	<	* probability for two threads accessing arbitrary distinct
374	<	* elements is, roughly, 1 / (8 * #elements).
375	<	*
376	<	* The table is resized when occupancy exceeds a threshold.  Only
377	<	* a single thread performs the resize (using field "resizing", to
378	<	* arrange exclusion), but the table otherwise remains usable for
379	<	* both reads and updates. Resizing proceeds by transferring bins,
380	<	* one by one, from the table to the next table.  Upon transfer,
381	<	* the old table bin contains only a special forwarding node (with
382	<	* negative hash code ("MOVED")) that contains the next table as
337	>	* first insertion.  Each bin in the table normally contains a
338	>	* list of Nodes (most often, the list has only zero or one Node).
339	>	* Table accesses require volatile/atomic reads, writes, and
340	>	* CASes.  Because there is no other way to arrange this without
341	>	* adding further indirections, we use intrinsics
342	>	* (sun.misc.Unsafe) operations.  The lists of nodes within bins
343	>	* are always accurately traversable under volatile reads, so long
344	>	* as lookups check hash code and non-nullness of value before
345	>	* checking key equality.
346	>	*
347	>	* We use the top two bits of Node hash fields for control
348	>	* purposes -- they are available anyway because of addressing
349	>	* constraints.  As explained further below, these top bits are
350	>	* used as follows:
351	>	*  00 - Normal
352	>	*  01 - Locked
353	>	*  11 - Locked and may have a thread waiting for lock
354	>	*  10 - Node is a forwarding node
355	>	*
356	>	* The lower 30 bits of each Node's hash field contain a
357	>	* transformation of the key's hash code, except for forwarding
358	>	* nodes, for which the lower bits are zero (and so always have
359	>	* hash field == MOVED).
360	>	*
361	>	* Insertion (via put or its variants) of the first node in an
362	>	* empty bin is performed by just CASing it to the bin.  This is
363	>	* by far the most common case for put operations under most
364	>	* key/hash distributions.  Other update operations (insert,
365	>	* delete, and replace) require locks.  We do not want to waste
366	>	* the space required to associate a distinct lock object with
367	>	* each bin, so instead use the first node of a bin list itself as
368	>	* a lock. Blocking support for these locks relies on the builtin
369	>	* "synchronized" monitors.  However, we also need a tryLock
370	>	* construction, so we overlay these by using bits of the Node
371	>	* hash field for lock control (see above), and so normally use
372	>	* builtin monitors only for blocking and signalling using
373	>	* wait/notifyAll constructions. See Node.tryAwaitLock.
374	>	*
375	>	* Using the first node of a list as a lock does not by itself
376	>	* suffice though: When a node is locked, any update must first
377	>	* validate that it is still the first node after locking it, and
378	>	* retry if not. Because new nodes are always appended to lists,
379	>	* once a node is first in a bin, it remains first until deleted
380	>	* or the bin becomes invalidated (upon resizing).  However,
381	>	* operations that only conditionally update may inspect nodes
382	>	* until the point of update. This is a converse of sorts to the
383	>	* lazy locking technique described by Herlihy & Shavit.
384	>	*
385	>	* The main disadvantage of per-bin locks is that other update
386	>	* operations on other nodes in a bin list protected by the same
387	>	* lock can stall, for example when user equals() or mapping
388	>	* functions take a long time.  However, statistically, under
389	>	* random hash codes, this is not a common problem.  Ideally, the
390	>	* frequency of nodes in bins follows a Poisson distribution
391	>	* (http://en.wikipedia.org/wiki/Poisson_distribution) with a
392	>	* parameter of about 0.5 on average, given the resizing threshold
393	>	* of 0.75, although with a large variance because of resizing
394	>	* granularity. Ignoring variance, the expected occurrences of
395	>	* list size k are (exp(-0.5) * pow(0.5, k) / factorial(k)). The
396	>	* first values are:
397	>	*
398	>	* 0:    0.60653066
399	>	* 1:    0.30326533
400	>	* 2:    0.07581633
401	>	* 3:    0.01263606
402	>	* 4:    0.00157952
403	>	* 5:    0.00015795
404	>	* 6:    0.00001316
405	>	* 7:    0.00000094
406	>	* 8:    0.00000006
407	>	* more: less than 1 in ten million
408	>	*
409	>	* Lock contention probability for two threads accessing distinct
410	>	* elements is roughly 1 / (8 * #elements) under random hashes.
411	>	*
412	>	* Actual hash code distributions encountered in practice
413	>	* sometimes deviate significantly from uniform randomness.  This
414	>	* includes the case when N > (1<<30), so some keys MUST collide.
415	>	* Similarly for dumb or hostile usages in which multiple keys are
416	>	* designed to have identical hash codes. Also, although we guard
417	>	* against the worst effects of this (see method spread), sets of
418	>	* hashes may differ only in bits that do not impact their bin
419	>	* index for a given power-of-two mask.  So we use a secondary
420	>	* strategy that applies when the number of nodes in a bin exceeds
421	>	* a threshold, and at least one of the keys implements
422	>	* Comparable.  These TreeBins use a balanced tree to hold nodes
423	>	* (a specialized form of red-black trees), bounding search time
424	>	* to O(log N).  Each search step in a TreeBin is around twice as
425	>	* slow as in a regular list, but given that N cannot exceed
426	>	* (1<<64) (before running out of addresses) this bounds search
427	>	* steps, lock hold times, etc, to reasonable constants (roughly
428	>	* 100 nodes inspected per operation worst case) so long as keys
429	>	* are Comparable (which is very common -- String, Long, etc).
430	>	* TreeBin nodes (TreeNodes) also maintain the same "next"
431	>	* traversal pointers as regular nodes, so can be traversed in
432	>	* iterators in the same way.
433	>	*
434	>	* The table is resized when occupancy exceeds a percentage
435	>	* threshold (nominally, 0.75, but see below).  Only a single
436	>	* thread performs the resize (using field "sizeCtl", to arrange
437	>	* exclusion), but the table otherwise remains usable for reads
438	>	* and updates. Resizing proceeds by transferring bins, one by
439	>	* one, from the table to the next table.  Because we are using
440	>	* power-of-two expansion, the elements from each bin must either
441	>	* stay at same index, or move with a power of two offset. We
442	>	* eliminate unnecessary node creation by catching cases where old
443	>	* nodes can be reused because their next fields won't change.  On
444	>	* average, only about one-sixth of them need cloning when a table
445	>	* doubles. The nodes they replace will be garbage collectable as
446	>	* soon as they are no longer referenced by any reader thread that
447	>	* may be in the midst of concurrently traversing table.  Upon
448	>	* transfer, the old table bin contains only a special forwarding
449	>	* node (with hash field "MOVED") that contains the next table as
450		* its key. On encountering a forwarding node, access and update
451	<	* operations restart, using the new table. To ensure concurrent
452	<	* readability of traversals, transfers must proceed from the last
453	<	* bin (table.length - 1) up towards the first.  Any traversal
454	<	* starting from the first bin can then arrange to move to the new
455	<	* table for the rest of the traversal without revisiting nodes.
456	<	* This constrains bin transfers to a particular order, and so can
457	<	* block indefinitely waiting for the next lock, and other threads
458	<	* cannot help with the transfer. However, expected stalls are
459	<	* infrequent enough to not warrant the additional overhead and
460	<	* complexity of access and iteration schemes that could admit
461	<	* out-of-order or concurrent bin transfers.
462	<	*
463	<	* A similar traversal scheme (not yet implemented) can apply to
464	<	* partial traversals during partitioned aggregate operations.
465	<	* Also, read-only operations give up if ever forwarded to a null
466	<	* table, which provides support for shutdown-style clearing,
467	<	* which is also not currently implemented.
451	>	* operations restart, using the new table.
452	>	*
453	>	* Each bin transfer requires its bin lock. However, unlike other
454	>	* cases, a transfer can skip a bin if it fails to acquire its
455	>	* lock, and revisit it later (unless it is a TreeBin). Method
456	>	* rebuild maintains a buffer of TRANSFER_BUFFER_SIZE bins that
457	>	* have been skipped because of failure to acquire a lock, and
458	>	* blocks only if none are available (i.e., only very rarely).
459	>	* The transfer operation must also ensure that all accessible
460	>	* bins in both the old and new table are usable by any traversal.
461	>	* When there are no lock acquisition failures, this is arranged
462	>	* simply by proceeding from the last bin (table.length - 1) up
463	>	* towards the first.  Upon seeing a forwarding node, traversals
464	>	* (see class Iter) arrange to move to the new table
465	>	* without revisiting nodes.  However, when any node is skipped
466	>	* during a transfer, all earlier table bins may have become
467	>	* visible, so are initialized with a reverse-forwarding node back
468	>	* to the old table until the new ones are established. (This
469	>	* sometimes requires transiently locking a forwarding node, which
470	>	* is possible under the above encoding.) These more expensive
471	>	* mechanics trigger only when necessary.
472	>	*
473	>	* The traversal scheme also applies to partial traversals of
474	>	* ranges of bins (via an alternate Traverser constructor)
475	>	* to support partitioned aggregate operations.  Also, read-only
476	>	* operations give up if ever forwarded to a null table, which
477	>	* provides support for shutdown-style clearing, which is also not
478	>	* currently implemented.
479	>	*
480	>	* Lazy table initialization minimizes footprint until first use,
481	>	* and also avoids resizings when the first operation is from a
482	>	* putAll, constructor with map argument, or deserialization.
483	>	* These cases attempt to override the initial capacity settings,
484	>	* but harmlessly fail to take effect in cases of races.
485		*
486		* The element count is maintained using a LongAdder, which avoids
487		* contention on updates but can encounter cache thrashing if read
488	<	* too frequently during concurrent updates. To avoid reading so
489	<	* often, resizing is normally attempted only upon adding to a bin
490	<	* already holding two or more nodes. Under the default threshold
491	<	* (0.75), and uniform hash distributions, the probability of this
492	<	* occurring at threshold is around 13%, meaning that only about 1
493	<	* in 8 puts check threshold (and after resizing, many fewer do
494	<	* so). But this approximation has high variance for small table
495	<	* sizes, so we check on any collision for sizes <= 64.  Further,
496	<	* to increase the probability that a resize occurs soon enough, we
497	<	* offset the threshold (see THRESHOLD_OFFSET) by the expected
498	<	* number of puts between checks. This is currently set to 8, in
499	<	* accord with the default load factor. In practice, this is
500	<	* rarely overridden, and in any case is close enough to other
501	<	* plausible values not to waste dynamic probability computation
502	<	* for more precision.
488	>	* too frequently during concurrent access. To avoid reading so
489	>	* often, resizing is attempted either when a bin lock is
490	>	* contended, or upon adding to a bin already holding two or more
491	>	* nodes (checked before adding in the xIfAbsent methods, after
492	>	* adding in others). Under uniform hash distributions, the
493	>	* probability of this occurring at threshold is around 13%,
494	>	* meaning that only about 1 in 8 puts check threshold (and after
495	>	* resizing, many fewer do so). But this approximation has high
496	>	* variance for small table sizes, so we check on any collision
497	>	* for sizes <= 64. The bulk putAll operation further reduces
498	>	* contention by only committing count updates upon these size
499	>	* checks.
500	>	*
501	>	* Maintaining API and serialization compatibility with previous
502	>	* versions of this class introduces several oddities. Mainly: We
503	>	* leave untouched but unused constructor arguments refering to
504	>	* concurrencyLevel. We accept a loadFactor constructor argument,
505	>	* but apply it only to initial table capacity (which is the only
506	>	* time that we can guarantee to honor it.) We also declare an
507	>	* unused "Segment" class that is instantiated in minimal form
508	>	* only when serializing.
509		*/
510
511		/* ---------------- Constants -------------- */
512
513		/**
514	<	* The smallest allowed table capacity.  Must be a power of 2, at
515	<	* least 2.
514	>	* The largest possible table capacity.  This value must be
515	>	* exactly 1<<30 to stay within Java array allocation and indexing
516	>	* bounds for power of two table sizes, and is further required
517	>	* because the top two bits of 32bit hash fields are used for
518	>	* control purposes.
519		*/
520	<	static final int MINIMUM_CAPACITY = 2;
520	>	private static final int MAXIMUM_CAPACITY = 1 << 30;
521
522		/**
523	<	* The largest allowed table capacity.  Must be a power of 2, at
524	<	* most 1<<30.
523	>	* The default initial table capacity.  Must be a power of 2
524	>	* (i.e., at least 1) and at most MAXIMUM_CAPACITY.
525		*/
526	<	static final int MAXIMUM_CAPACITY = 1 << 30;
526	>	private static final int DEFAULT_CAPACITY = 16;
527
528		/**
529	<	* The default initial table capacity.  Must be a power of 2, at
530	<	* least MINIMUM_CAPACITY and at most MAXIMUM_CAPACITY.
529	>	* The largest possible (non-power of two) array size.
530	>	* Needed by toArray and related methods.
531		*/
532	<	static final int DEFAULT_CAPACITY = 16;
532	>	static final int MAX_ARRAY_SIZE = Integer.MAX_VALUE - 8;
533
534		/**
535	<	* The default load factor for this table, used when not otherwise
536	<	* specified in a constructor.
535	>	* The default concurrency level for this table. Unused but
536	>	* defined for compatibility with previous versions of this class.
537		*/
538	<	static final float DEFAULT_LOAD_FACTOR = 0.75f;
538	>	private static final int DEFAULT_CONCURRENCY_LEVEL = 16;
539
540		/**
541	<	* The default concurrency level for this table. Unused, but
542	<	* defined for compatibility with previous versions of this class.
541	>	* The load factor for this table. Overrides of this value in
542	>	* constructors affect only the initial table capacity.  The
543	>	* actual floating point value isn't normally used -- it is
544	>	* simpler to use expressions such as {@code n - (n >>> 2)} for
545	>	* the associated resizing threshold.
546		*/
547	<	static final int DEFAULT_CONCURRENCY_LEVEL = 16;
547	>	private static final float LOAD_FACTOR = 0.75f;
548
549		/**
550	<	* The count value to offset thresholds to compensate for checking
551	<	* for resizing only when inserting into bins with two or more
552	<	* elements. See above for explanation.
550	>	* The buffer size for skipped bins during transfers. The
551	>	* value is arbitrary but should be large enough to avoid
552	>	* most locking stalls during resizes.
553		*/
554	<	static final int THRESHOLD_OFFSET = 8;
554	>	private static final int TRANSFER_BUFFER_SIZE = 32;
555
556		/**
557	<	* Special node hash value indicating to use table in node.key
558	<	* Must be negative.
557	>	* The bin count threshold for using a tree rather than list for a
558	>	* bin.  The value reflects the approximate break-even point for
559	>	* using tree-based operations.
560		*/
561	<	static final int MOVED = -1;
561	>	private static final int TREE_THRESHOLD = 8;
562	>
563	>	/*
564	>	* Encodings for special uses of Node hash fields. See above for
565	>	* explanation.
566	>	*/
567	>	static final int MOVED     = 0x80000000; // hash field for forwarding nodes
568	>	static final int LOCKED    = 0x40000000; // set/tested only as a bit
569	>	static final int WAITING   = 0xc0000000; // both bits set/tested together
570	>	static final int HASH_BITS = 0x3fffffff; // usable bits of normal node hash
571
572		/* ---------------- Fields -------------- */
573
574		/**
575		* The array of bins. Lazily initialized upon first insertion.
576	<	* Size is always a power of two. Accessed directly by inner
254	<	* classes.
576	>	* Size is always a power of two. Accessed directly by iterators.
577		*/
578		transient volatile Node[] table;
579
580	<	/** The counter maintaining number of elements. */
580	>	/**
581	>	* The counter maintaining number of elements.
582	>	*/
583		private transient final LongAdder counter;
584	<	/** Nonzero when table is being initialized or resized. Updated via CAS. */
585	<	private transient volatile int resizing;
586	<	/** The target load factor for the table. */
587	<	private transient float loadFactor;
588	<	/** The next element count value upon which to resize the table. */
589	<	private transient int threshold;
590	<	/** The initial capacity of the table. */
591	<	private transient int initCap;
584	>
585	>	/**
586	>	* Table initialization and resizing control.  When negative, the
587	>	* table is being initialized or resized. Otherwise, when table is
588	>	* null, holds the initial table size to use upon creation, or 0
589	>	* for default. After initialization, holds the next element count
590	>	* value upon which to resize the table.
591	>	*/
592	>	private transient volatile int sizeCtl;
593
594		// views
595	<	transient Set<K> keySet;
596	<	transient Set<Map.Entry<K,V>> entrySet;
597	<	transient Collection<V> values;
595	>	private transient KeySetView<K,V> keySet;
596	>	private transient ValuesView<K,V> values;
597	>	private transient EntrySetView<K,V> entrySet;
598
599		/** For serialization compatibility. Null unless serialized; see below */
600	<	Segment<K,V>[] segments;
600	>	private Segment<K,V>[] segments;
601
602	<	/**
603	<	* Applies a supplemental hash function to a given hashCode, which
604	<	* defends against poor quality hash functions.  The result must
605	<	* be non-negative, and for reasonable performance must have good
606	<	* avalanche properties; i.e., that each bit of the argument
607	<	* affects each bit (except sign bit) of the result.
602	>	/* ---------------- Table element access -------------- */
603	>
604	>	/*
605	>	* Volatile access methods are used for table elements as well as
606	>	* elements of in-progress next table while resizing.  Uses are
607	>	* null checked by callers, and implicitly bounds-checked, relying
608	>	* on the invariants that tab arrays have non-zero size, and all
609	>	* indices are masked with (tab.length - 1) which is never
610	>	* negative and always less than length. Note that, to be correct
611	>	* wrt arbitrary concurrency errors by users, bounds checks must
612	>	* operate on local variables, which accounts for some odd-looking
613	>	* inline assignments below.
614		*/
615	<	private static final int spread(int h) {
616	<	// Apply base step of MurmurHash; see http://code.google.com/p/smhasher/
617	<	h ^= h >>> 16;
618	<	h *= 0x85ebca6b;
619	<	h ^= h >>> 13;
620	<	h *= 0xc2b2ae35;
621	<	return (h >>> 16) ^ (h & 0x7fffffff); // mask out sign bit
615	>
616	>	static final Node tabAt(Node[] tab, int i) { // used by Iter
617	>	return (Node)UNSAFE.getObjectVolatile(tab, ((long)i<<ASHIFT)+ABASE);
618	>	}
619	>
620	>	private static final boolean casTabAt(Node[] tab, int i, Node c, Node v) {
621	>	return UNSAFE.compareAndSwapObject(tab, ((long)i<<ASHIFT)+ABASE, c, v);
622		}
623
624	+	private static final void setTabAt(Node[] tab, int i, Node v) {
625	+	UNSAFE.putObjectVolatile(tab, ((long)i<<ASHIFT)+ABASE, v);
626	+	}
627	+
628	+	/* ---------------- Nodes -------------- */
629	+
630		/**
631		* Key-value entry. Note that this is never exported out as a
632	<	* user-visible Map.Entry.
632	>	* user-visible Map.Entry (see MapEntry below). Nodes with a hash
633	>	* field of MOVED are special, and do not contain user keys or
634	>	* values.  Otherwise, keys are never null, and null val fields
635	>	* indicate that a node is in the process of being deleted or
636	>	* created. For purposes of read-only access, a key may be read
637	>	* before a val, but can only be used after checking val to be
638	>	* non-null.
639		*/
640	<	static final class Node {
641	<	final int hash;
640	>	static class Node {
641	>	volatile int hash;
642		final Object key;
643		volatile Object val;
644		volatile Node next;
#	Line 306 | Line 649 | public class ConcurrentHashMapV8<K, V>
649		this.val = val;
650		this.next = next;
651		}
309	–	}
652
653	<	/*
654	<	* Volatile access methods are used for table elements as well as
655	<	* elements of in-progress next table while resizing.  Uses in
656	<	* access and update methods are null checked by callers, and
315	<	* implicitly bounds-checked, relying on the invariants that tab
316	<	* arrays have non-zero size, and all indices are masked with
317	<	* (tab.length - 1) which is never negative and always less than
318	<	* length. The "relaxed" non-volatile forms are used only during
319	<	* table initialization. The only other usage is in
320	<	* HashIterator.advance, which performs explicit checks.
321	<	*/
653	>	/** CompareAndSet the hash field */
654	>	final boolean casHash(int cmp, int val) {
655	>	return UNSAFE.compareAndSwapInt(this, hashOffset, cmp, val);
656	>	}
657
658	<	static final Node tabAt(Node[] tab, int i) { // used in HashIterator
659	<	return (Node)UNSAFE.getObjectVolatile(tab, ((long)i<<ASHIFT)+ABASE);
660	<	}
658	>	/** The number of spins before blocking for a lock */
659	>	static final int MAX_SPINS =
660	>	Runtime.getRuntime().availableProcessors() > 1 ? 64 : 1;
661
662	<	private static final boolean casTabAt(Node[] tab, int i, Node c, Node v) {
663	<	return UNSAFE.compareAndSwapObject(tab, ((long)i<<ASHIFT)+ABASE, c, v);
664	<	}
662	>	/**
663	>	* Spins a while if LOCKED bit set and this node is the first
664	>	* of its bin, and then sets WAITING bits on hash field and
665	>	* blocks (once) if they are still set.  It is OK for this
666	>	* method to return even if lock is not available upon exit,
667	>	* which enables these simple single-wait mechanics.
668	>	*
669	>	* The corresponding signalling operation is performed within
670	>	* callers: Upon detecting that WAITING has been set when
671	>	* unlocking lock (via a failed CAS from non-waiting LOCKED
672	>	* state), unlockers acquire the sync lock and perform a
673	>	* notifyAll.
674	>	*
675	>	* The initial sanity check on tab and bounds is not currently
676	>	* necessary in the only usages of this method, but enables
677	>	* use in other future contexts.
678	>	*/
679	>	final void tryAwaitLock(Node[] tab, int i) {
680	>	if (tab != null && i >= 0 && i < tab.length) { // sanity check
681	>	int r = ThreadLocalRandom.current().nextInt(); // randomize spins
682	>	int spins = MAX_SPINS, h;
683	>	while (tabAt(tab, i) == this && ((h = hash) & LOCKED) != 0) {
684	>	if (spins >= 0) {
685	>	r ^= r << 1; r ^= r >>> 3; r ^= r << 10; // xorshift
686	>	if (r >= 0 && --spins == 0)
687	>	Thread.yield();  // yield before block
688	>	}
689	>	else if (casHash(h, h | WAITING)) {
690	>	synchronized (this) {
691	>	if (tabAt(tab, i) == this &&
692	>	(hash & WAITING) == WAITING) {
693	>	try {
694	>	wait();
695	>	} catch (InterruptedException ie) {
696	>	try {
697	>	Thread.currentThread().interrupt();
698	>	} catch (SecurityException ignore) {
699	>	}
700	>	}
701	>	}
702	>	else
703	>	notifyAll(); // possibly won race vs signaller
704	>	}
705	>	break;
706	>	}
707	>	}
708	>	}
709	>	}
710
711	<	private static final void setTabAt(Node[] tab, int i, Node v) {
712	<	UNSAFE.putObjectVolatile(tab, ((long)i<<ASHIFT)+ABASE, v);
713	<	}
711	>	// Unsafe mechanics for casHash
712	>	private static final sun.misc.Unsafe UNSAFE;
713	>	private static final long hashOffset;
714
715	<	private static final Node relaxedTabAt(Node[] tab, int i) {
716	<	return (Node)UNSAFE.getObject(tab, ((long)i<<ASHIFT)+ABASE);
715	>	static {
716	>	try {
717	>	UNSAFE = getUnsafe();
718	>	Class<?> k = Node.class;
719	>	hashOffset = UNSAFE.objectFieldOffset
720	>	(k.getDeclaredField("hash"));
721	>	} catch (Exception e) {
722	>	throw new Error(e);
723	>	}
724	>	}
725		}
726
727	<	private static final void relaxedSetTabAt(Node[] tab, int i, Node v) {
728	<	UNSAFE.putObject(tab, ((long)i<<ASHIFT)+ABASE, v);
727	>	/* ---------------- TreeBins -------------- */
728	>
729	>	/**
730	>	* Nodes for use in TreeBins
731	>	*/
732	>	static final class TreeNode extends Node {
733	>	TreeNode parent;  // red-black tree links
734	>	TreeNode left;
735	>	TreeNode right;
736	>	TreeNode prev;    // needed to unlink next upon deletion
737	>	boolean red;
738	>
739	>	TreeNode(int hash, Object key, Object val, Node next, TreeNode parent) {
740	>	super(hash, key, val, next);
741	>	this.parent = parent;
742	>	}
743		}
744
745	<	/* ---------------- Access and update operations -------------- */
745	>	/**
746	>	* A specialized form of red-black tree for use in bins
747	>	* whose size exceeds a threshold.
748	>	*
749	>	* TreeBins use a special form of comparison for search and
750	>	* related operations (which is the main reason we cannot use
751	>	* existing collections such as TreeMaps). TreeBins contain
752	>	* Comparable elements, but may contain others, as well as
753	>	* elements that are Comparable but not necessarily Comparable<T>
754	>	* for the same T, so we cannot invoke compareTo among them. To
755	>	* handle this, the tree is ordered primarily by hash value, then
756	>	* by getClass().getName() order, and then by Comparator order
757	>	* among elements of the same class.  On lookup at a node, if
758	>	* elements are not comparable or compare as 0, both left and
759	>	* right children may need to be searched in the case of tied hash
760	>	* values. (This corresponds to the full list search that would be
761	>	* necessary if all elements were non-Comparable and had tied
762	>	* hashes.)  The red-black balancing code is updated from
763	>	* pre-jdk-collections
764	>	* (http://gee.cs.oswego.edu/dl/classes/collections/RBCell.java)
765	>	* based in turn on Cormen, Leiserson, and Rivest "Introduction to
766	>	* Algorithms" (CLR).
767	>	*
768	>	* TreeBins also maintain a separate locking discipline than
769	>	* regular bins. Because they are forwarded via special MOVED
770	>	* nodes at bin heads (which can never change once established),
771	>	* we cannot use those nodes as locks. Instead, TreeBin
772	>	* extends AbstractQueuedSynchronizer to support a simple form of
773	>	* read-write lock. For update operations and table validation,
774	>	* the exclusive form of lock behaves in the same way as bin-head
775	>	* locks. However, lookups use shared read-lock mechanics to allow
776	>	* multiple readers in the absence of writers.  Additionally,
777	>	* these lookups do not ever block: While the lock is not
778	>	* available, they proceed along the slow traversal path (via
779	>	* next-pointers) until the lock becomes available or the list is
780	>	* exhausted, whichever comes first. (These cases are not fast,
781	>	* but maximize aggregate expected throughput.)  The AQS mechanics
782	>	* for doing this are straightforward.  The lock state is held as
783	>	* AQS getState().  Read counts are negative; the write count (1)
784	>	* is positive.  There are no signalling preferences among readers
785	>	* and writers. Since we don't need to export full Lock API, we
786	>	* just override the minimal AQS methods and use them directly.
787	>	*/
788	>	static final class TreeBin extends AbstractQueuedSynchronizer {
789	>	private static final long serialVersionUID = 2249069246763182397L;
790	>	transient TreeNode root;  // root of tree
791	>	transient TreeNode first; // head of next-pointer list
792
793	<	/** Implementation for get and containsKey */
794	<	private final Object internalGet(Object k) {
795	<	int h = spread(k.hashCode());
796	<	Node[] tab = table;
797	<	retry: while (tab != null) {
798	<	Node e = tabAt(tab, (tab.length - 1) & h);
799	<	while (e != null) {
800	<	int eh = e.hash;
801	<	if (eh == h) {
802	<	Object ek = e.key, ev = e.val;
803	<	if (ev != null && ek != null && (k == ek || k.equals(ek)))
804	<	return ev;
805	<	}
806	<	else if (eh < 0) { // bin was moved during resize
807	<	tab = (Node[])e.key;
808	<	continue retry;
809	<	}
810	<	e = e.next;
793	>	/* AQS overrides */
794	>	public final boolean isHeldExclusively() { return getState() > 0; }
795	>	public final boolean tryAcquire(int ignore) {
796	>	if (compareAndSetState(0, 1)) {
797	>	setExclusiveOwnerThread(Thread.currentThread());
798	>	return true;
799	>	}
800	>	return false;
801	>	}
802	>	public final boolean tryRelease(int ignore) {
803	>	setExclusiveOwnerThread(null);
804	>	setState(0);
805	>	return true;
806	>	}
807	>	public final int tryAcquireShared(int ignore) {
808	>	for (int c;;) {
809	>	if ((c = getState()) > 0)
810	>	return -1;
811	>	if (compareAndSetState(c, c -1))
812	>	return 1;
813	>	}
814	>	}
815	>	public final boolean tryReleaseShared(int ignore) {
816	>	int c;
817	>	do {} while (!compareAndSetState(c = getState(), c + 1));
818	>	return c == -1;
819	>	}
820	>
821	>	/** From CLR */
822	>	private void rotateLeft(TreeNode p) {
823	>	if (p != null) {
824	>	TreeNode r = p.right, pp, rl;
825	>	if ((rl = p.right = r.left) != null)
826	>	rl.parent = p;
827	>	if ((pp = r.parent = p.parent) == null)
828	>	root = r;
829	>	else if (pp.left == p)
830	>	pp.left = r;
831	>	else
832	>	pp.right = r;
833	>	r.left = p;
834	>	p.parent = r;
835		}
364	–	break;
836		}
366	–	return null;
367	–	}
837
838	+	/** From CLR */
839	+	private void rotateRight(TreeNode p) {
840	+	if (p != null) {
841	+	TreeNode l = p.left, pp, lr;
842	+	if ((lr = p.left = l.right) != null)
843	+	lr.parent = p;
844	+	if ((pp = l.parent = p.parent) == null)
845	+	root = l;
846	+	else if (pp.right == p)
847	+	pp.right = l;
848	+	else
849	+	pp.left = l;
850	+	l.right = p;
851	+	p.parent = l;
852	+	}
853	+	}
854
855	<	/** Implementation for put and putIfAbsent */
856	<	private final Object internalPut(Object k, Object v, boolean replace) {
857	<	int h = spread(k.hashCode());
858	<	Object oldVal = null;  // the previous value or null if none
859	<	Node[] tab = table;
860	<	for (;;) {
861	<	Node e; int i;
862	<	if (tab == null)
863	<	tab = grow(0);
864	<	else if ((e = tabAt(tab, i = (tab.length - 1) & h)) == null) {
865	<	if (casTabAt(tab, i, null, new Node(h, k, v, null)))
855	>	/**
856	>	* Returns the TreeNode (or null if not found) for the given key
857	>	* starting at given root.
858	>	*/
859	>	@SuppressWarnings("unchecked") final TreeNode getTreeNode
860	>	(int h, Object k, TreeNode p) {
861	>	Class<?> c = k.getClass();
862	>	while (p != null) {
863	>	int dir, ph;  Object pk; Class<?> pc;
864	>	if ((ph = p.hash) == h) {
865	>	if ((pk = p.key) == k || k.equals(pk))
866	>	return p;
867	>	if (c != (pc = pk.getClass()) ||
868	>	!(k instanceof Comparable) ||
869	>	(dir = ((Comparable)k).compareTo((Comparable)pk)) == 0) {
870	>	dir = (c == pc) ? 0 : c.getName().compareTo(pc.getName());
871	>	TreeNode r = null, s = null, pl, pr;
872	>	if (dir >= 0) {
873	>	if ((pl = p.left) != null && h <= pl.hash)
874	>	s = pl;
875	>	}
876	>	else if ((pr = p.right) != null && h >= pr.hash)
877	>	s = pr;
878	>	if (s != null && (r = getTreeNode(h, k, s)) != null)
879	>	return r;
880	>	}
881	>	}
882	>	else
883	>	dir = (h < ph) ? -1 : 1;
884	>	p = (dir > 0) ? p.right : p.left;
885	>	}
886	>	return null;
887	>	}
888	>
889	>	/**
890	>	* Wrapper for getTreeNode used by CHM.get. Tries to obtain
891	>	* read-lock to call getTreeNode, but during failure to get
892	>	* lock, searches along next links.
893	>	*/
894	>	final Object getValue(int h, Object k) {
895	>	Node r = null;
896	>	int c = getState(); // Must read lock state first
897	>	for (Node e = first; e != null; e = e.next) {
898	>	if (c <= 0 && compareAndSetState(c, c - 1)) {
899	>	try {
900	>	r = getTreeNode(h, k, root);
901	>	} finally {
902	>	releaseShared(0);
903	>	}
904		break;
905	+	}
906	+	else if ((e.hash & HASH_BITS) == h && k.equals(e.key)) {
907	+	r = e;
908	+	break;
909	+	}
910	+	else
911	+	c = getState();
912	+	}
913	+	return r == null ? null : r.val;
914	+	}
915	+
916	+	/**
917	+	* Finds or adds a node.
918	+	* @return null if added
919	+	*/
920	+	@SuppressWarnings("unchecked") final TreeNode putTreeNode
921	+	(int h, Object k, Object v) {
922	+	Class<?> c = k.getClass();
923	+	TreeNode pp = root, p = null;
924	+	int dir = 0;
925	+	while (pp != null) { // find existing node or leaf to insert at
926	+	int ph;  Object pk; Class<?> pc;
927	+	p = pp;
928	+	if ((ph = p.hash) == h) {
929	+	if ((pk = p.key) == k || k.equals(pk))
930	+	return p;
931	+	if (c != (pc = pk.getClass()) ||
932	+	!(k instanceof Comparable) ||
933	+	(dir = ((Comparable)k).compareTo((Comparable)pk)) == 0) {
934	+	dir = (c == pc) ? 0 : c.getName().compareTo(pc.getName());
935	+	TreeNode r = null, s = null, pl, pr;
936	+	if (dir >= 0) {
937	+	if ((pl = p.left) != null && h <= pl.hash)
938	+	s = pl;
939	+	}
940	+	else if ((pr = p.right) != null && h >= pr.hash)
941	+	s = pr;
942	+	if (s != null && (r = getTreeNode(h, k, s)) != null)
943	+	return r;
944	+	}
945	+	}
946	+	else
947	+	dir = (h < ph) ? -1 : 1;
948	+	pp = (dir > 0) ? p.right : p.left;
949	+	}
950	+
951	+	TreeNode f = first;
952	+	TreeNode x = first = new TreeNode(h, k, v, f, p);
953	+	if (p == null)
954	+	root = x;
955	+	else { // attach and rebalance; adapted from CLR
956	+	TreeNode xp, xpp;
957	+	if (f != null)
958	+	f.prev = x;
959	+	if (dir <= 0)
960	+	p.left = x;
961	+	else
962	+	p.right = x;
963	+	x.red = true;
964	+	while (x != null && (xp = x.parent) != null && xp.red &&
965	+	(xpp = xp.parent) != null) {
966	+	TreeNode xppl = xpp.left;
967	+	if (xp == xppl) {
968	+	TreeNode y = xpp.right;
969	+	if (y != null && y.red) {
970	+	y.red = false;
971	+	xp.red = false;
972	+	xpp.red = true;
973	+	x = xpp;
974	+	}
975	+	else {
976	+	if (x == xp.right) {
977	+	rotateLeft(x = xp);
978	+	xpp = (xp = x.parent) == null ? null : xp.parent;
979	+	}
980	+	if (xp != null) {
981	+	xp.red = false;
982	+	if (xpp != null) {
983	+	xpp.red = true;
984	+	rotateRight(xpp);
985	+	}
986	+	}
987	+	}
988	+	}
989	+	else {
990	+	TreeNode y = xppl;
991	+	if (y != null && y.red) {
992	+	y.red = false;
993	+	xp.red = false;
994	+	xpp.red = true;
995	+	x = xpp;
996	+	}
997	+	else {
998	+	if (x == xp.left) {
999	+	rotateRight(x = xp);
1000	+	xpp = (xp = x.parent) == null ? null : xp.parent;
1001	+	}
1002	+	if (xp != null) {
1003	+	xp.red = false;
1004	+	if (xpp != null) {
1005	+	xpp.red = true;
1006	+	rotateLeft(xpp);
1007	+	}
1008	+	}
1009	+	}
1010	+	}
1011	+	}
1012	+	TreeNode r = root;
1013	+	if (r != null && r.red)
1014	+	r.red = false;
1015	+	}
1016	+	return null;
1017	+	}
1018	+
1019	+	/**
1020	+	* Removes the given node, that must be present before this
1021	+	* call.  This is messier than typical red-black deletion code
1022	+	* because we cannot swap the contents of an interior node
1023	+	* with a leaf successor that is pinned by "next" pointers
1024	+	* that are accessible independently of lock. So instead we
1025	+	* swap the tree linkages.
1026	+	*/
1027	+	final void deleteTreeNode(TreeNode p) {
1028	+	TreeNode next = (TreeNode)p.next; // unlink traversal pointers
1029	+	TreeNode pred = p.prev;
1030	+	if (pred == null)
1031	+	first = next;
1032	+	else
1033	+	pred.next = next;
1034	+	if (next != null)
1035	+	next.prev = pred;
1036	+	TreeNode replacement;
1037	+	TreeNode pl = p.left;
1038	+	TreeNode pr = p.right;
1039	+	if (pl != null && pr != null) {
1040	+	TreeNode s = pr, sl;
1041	+	while ((sl = s.left) != null) // find successor
1042	+	s = sl;
1043	+	boolean c = s.red; s.red = p.red; p.red = c; // swap colors
1044	+	TreeNode sr = s.right;
1045	+	TreeNode pp = p.parent;
1046	+	if (s == pr) { // p was s's direct parent
1047	+	p.parent = s;
1048	+	s.right = p;
1049	+	}
1050	+	else {
1051	+	TreeNode sp = s.parent;
1052	+	if ((p.parent = sp) != null) {
1053	+	if (s == sp.left)
1054	+	sp.left = p;
1055	+	else
1056	+	sp.right = p;
1057	+	}
1058	+	if ((s.right = pr) != null)
1059	+	pr.parent = s;
1060	+	}
1061	+	p.left = null;
1062	+	if ((p.right = sr) != null)
1063	+	sr.parent = p;
1064	+	if ((s.left = pl) != null)
1065	+	pl.parent = s;
1066	+	if ((s.parent = pp) == null)
1067	+	root = s;
1068	+	else if (p == pp.left)
1069	+	pp.left = s;
1070	+	else
1071	+	pp.right = s;
1072	+	replacement = sr;
1073	+	}
1074	+	else
1075	+	replacement = (pl != null) ? pl : pr;
1076	+	TreeNode pp = p.parent;
1077	+	if (replacement == null) {
1078	+	if (pp == null) {
1079	+	root = null;
1080	+	return;
1081	+	}
1082	+	replacement = p;
1083		}
383	–	else if (e.hash < 0)
384	–	tab = (Node[])e.key;
1084		else {
1085	<	boolean validated = false;
1086	<	boolean checkSize = false;
1087	<	synchronized (e) {
1088	<	if (tabAt(tab, i) == e) {
1089	<	validated = true;
1090	<	for (Node first = e;;) {
1091	<	Object ek, ev;
1092	<	if (e.hash == h &&
1093	<	(ek = e.key) != null &&
1094	<	(ev = e.val) != null &&
1095	<	(k == ek || k.equals(ek))) {
1096	<	oldVal = ev;
1097	<	if (replace)
1098	<	e.val = v;
1099	<	break;
1085	>	replacement.parent = pp;
1086	>	if (pp == null)
1087	>	root = replacement;
1088	>	else if (p == pp.left)
1089	>	pp.left = replacement;
1090	>	else
1091	>	pp.right = replacement;
1092	>	p.left = p.right = p.parent = null;
1093	>	}
1094	>	if (!p.red) { // rebalance, from CLR
1095	>	TreeNode x = replacement;
1096	>	while (x != null) {
1097	>	TreeNode xp, xpl;
1098	>	if (x.red || (xp = x.parent) == null) {
1099	>	x.red = false;
1100	>	break;
1101	>	}
1102	>	if (x == (xpl = xp.left)) {
1103	>	TreeNode sib = xp.right;
1104	>	if (sib != null && sib.red) {
1105	>	sib.red = false;
1106	>	xp.red = true;
1107	>	rotateLeft(xp);
1108	>	sib = (xp = x.parent) == null ? null : xp.right;
1109	>	}
1110	>	if (sib == null)
1111	>	x = xp;
1112	>	else {
1113	>	TreeNode sl = sib.left, sr = sib.right;
1114	>	if ((sr == null || !sr.red) &&
1115	>	(sl == null || !sl.red)) {
1116	>	sib.red = true;
1117	>	x = xp;
1118		}
1119	<	Node last = e;
1120	<	if ((e = e.next) == null) {
1121	<	last.next = new Node(h, k, v, null);
1122	<	if (last != first || tab.length <= 64)
1123	<	checkSize = true;
1124	<	break;
1119	>	else {
1120	>	if (sr == null || !sr.red) {
1121	>	if (sl != null)
1122	>	sl.red = false;
1123	>	sib.red = true;
1124	>	rotateRight(sib);
1125	>	sib = (xp = x.parent) == null ? null : xp.right;
1126	>	}
1127	>	if (sib != null) {
1128	>	sib.red = (xp == null) ? false : xp.red;
1129	>	if ((sr = sib.right) != null)
1130	>	sr.red = false;
1131	>	}
1132	>	if (xp != null) {
1133	>	xp.red = false;
1134	>	rotateLeft(xp);
1135	>	}
1136	>	x = root;
1137	>	}
1138	>	}
1139	>	}
1140	>	else { // symmetric
1141	>	TreeNode sib = xpl;
1142	>	if (sib != null && sib.red) {
1143	>	sib.red = false;
1144	>	xp.red = true;
1145	>	rotateRight(xp);
1146	>	sib = (xp = x.parent) == null ? null : xp.left;
1147	>	}
1148	>	if (sib == null)
1149	>	x = xp;
1150	>	else {
1151	>	TreeNode sl = sib.left, sr = sib.right;
1152	>	if ((sl == null || !sl.red) &&
1153	>	(sr == null || !sr.red)) {
1154	>	sib.red = true;
1155	>	x = xp;
1156	>	}
1157	>	else {
1158	>	if (sl == null || !sl.red) {
1159	>	if (sr != null)
1160	>	sr.red = false;
1161	>	sib.red = true;
1162	>	rotateLeft(sib);
1163	>	sib = (xp = x.parent) == null ? null : xp.left;
1164	>	}
1165	>	if (sib != null) {
1166	>	sib.red = (xp == null) ? false : xp.red;
1167	>	if ((sl = sib.left) != null)
1168	>	sl.red = false;
1169	>	}
1170	>	if (xp != null) {
1171	>	xp.red = false;
1172	>	rotateRight(xp);
1173	>	}
1174	>	x = root;
1175		}
1176		}
1177		}
1178		}
1179	<	if (validated) {
1180	<	if (checkSize && tab.length < MAXIMUM_CAPACITY &&
1181	<	resizing == 0 && counter.sum() >= threshold)
1182	<	grow(0);
1183	<	break;
1179	>	}
1180	>	if (p == replacement && (pp = p.parent) != null) {
1181	>	if (p == pp.left) // detach pointers
1182	>	pp.left = null;
1183	>	else if (p == pp.right)
1184	>	pp.right = null;
1185	>	p.parent = null;
1186	>	}
1187	>	}
1188	>	}
1189	>
1190	>	/* ---------------- Collision reduction methods -------------- */
1191	>
1192	>	/**
1193	>	* Spreads higher bits to lower, and also forces top 2 bits to 0.
1194	>	* Because the table uses power-of-two masking, sets of hashes
1195	>	* that vary only in bits above the current mask will always
1196	>	* collide. (Among known examples are sets of Float keys holding
1197	>	* consecutive whole numbers in small tables.)  To counter this,
1198	>	* we apply a transform that spreads the impact of higher bits
1199	>	* downward. There is a tradeoff between speed, utility, and
1200	>	* quality of bit-spreading. Because many common sets of hashes
1201	>	* are already reasonably distributed across bits (so don't benefit
1202	>	* from spreading), and because we use trees to handle large sets
1203	>	* of collisions in bins, we don't need excessively high quality.
1204	>	*/
1205	>	private static final int spread(int h) {
1206	>	h ^= (h >>> 18) ^ (h >>> 12);
1207	>	return (h ^ (h >>> 10)) & HASH_BITS;
1208	>	}
1209	>
1210	>	/**
1211	>	* Replaces a list bin with a tree bin. Call only when locked.
1212	>	* Fails to replace if the given key is non-comparable or table
1213	>	* is, or needs, resizing.
1214	>	*/
1215	>	private final void replaceWithTreeBin(Node[] tab, int index, Object key) {
1216	>	if ((key instanceof Comparable) &&
1217	>	(tab.length >= MAXIMUM_CAPACITY || counter.sum() < (long)sizeCtl)) {
1218	>	TreeBin t = new TreeBin();
1219	>	for (Node e = tabAt(tab, index); e != null; e = e.next)
1220	>	t.putTreeNode(e.hash & HASH_BITS, e.key, e.val);
1221	>	setTabAt(tab, index, new Node(MOVED, t, null, null));
1222	>	}
1223	>	}
1224	>
1225	>	/* ---------------- Internal access and update methods -------------- */
1226	>
1227	>	/** Implementation for get and containsKey */
1228	>	private final Object internalGet(Object k) {
1229	>	int h = spread(k.hashCode());
1230	>	retry: for (Node[] tab = table; tab != null;) {
1231	>	Node e, p; Object ek, ev; int eh;      // locals to read fields once
1232	>	for (e = tabAt(tab, (tab.length - 1) & h); e != null; e = e.next) {
1233	>	if ((eh = e.hash) == MOVED) {
1234	>	if ((ek = e.key) instanceof TreeBin)  // search TreeBin
1235	>	return ((TreeBin)ek).getValue(h, k);
1236	>	else {                        // restart with new table
1237	>	tab = (Node[])ek;
1238	>	continue retry;
1239	>	}
1240		}
1241	+	else if ((eh & HASH_BITS) == h && (ev = e.val) != null &&
1242	+	((ek = e.key) == k || k.equals(ek)))
1243	+	return ev;
1244		}
1245	+	break;
1246		}
1247	<	if (oldVal == null)
421	<	counter.increment();
422	<	return oldVal;
1247	>	return null;
1248		}
1249
1250		/**
#	Line 430 | Line 1255 | public class ConcurrentHashMapV8<K, V>
1255		private final Object internalReplace(Object k, Object v, Object cv) {
1256		int h = spread(k.hashCode());
1257		Object oldVal = null;
1258	<	Node e; int i;
1259	<	Node[] tab = table;
1260	<	while (tab != null &&
1261	<	(e = tabAt(tab, i = (tab.length - 1) & h)) != null) {
1262	<	if (e.hash < 0)
1263	<	tab = (Node[])e.key;
1264	<	else {
1258	>	for (Node[] tab = table;;) {
1259	>	Node f; int i, fh; Object fk;
1260	>	if (tab == null ||
1261	>	(f = tabAt(tab, i = (tab.length - 1) & h)) == null)
1262	>	break;
1263	>	else if ((fh = f.hash) == MOVED) {
1264	>	if ((fk = f.key) instanceof TreeBin) {
1265	>	TreeBin t = (TreeBin)fk;
1266	>	boolean validated = false;
1267	>	boolean deleted = false;
1268	>	t.acquire(0);
1269	>	try {
1270	>	if (tabAt(tab, i) == f) {
1271	>	validated = true;
1272	>	TreeNode p = t.getTreeNode(h, k, t.root);
1273	>	if (p != null) {
1274	>	Object pv = p.val;
1275	>	if (cv == null || cv == pv || cv.equals(pv)) {
1276	>	oldVal = pv;
1277	>	if ((p.val = v) == null) {
1278	>	deleted = true;
1279	>	t.deleteTreeNode(p);
1280	>	}
1281	>	}
1282	>	}
1283	>	}
1284	>	} finally {
1285	>	t.release(0);
1286	>	}
1287	>	if (validated) {
1288	>	if (deleted)
1289	>	counter.add(-1L);
1290	>	break;
1291	>	}
1292	>	}
1293	>	else
1294	>	tab = (Node[])fk;
1295	>	}
1296	>	else if ((fh & HASH_BITS) != h && f.next == null) // precheck
1297	>	break;                          // rules out possible existence
1298	>	else if ((fh & LOCKED) != 0) {
1299	>	checkForResize();               // try resizing if can't get lock
1300	>	f.tryAwaitLock(tab, i);
1301	>	}
1302	>	else if (f.casHash(fh, fh | LOCKED)) {
1303		boolean validated = false;
1304		boolean deleted = false;
1305	<	synchronized (e) {
1306	<	if (tabAt(tab, i) == e) {
1305	>	try {
1306	>	if (tabAt(tab, i) == f) {
1307		validated = true;
1308	<	Node pred = null;
446	<	do {
1308	>	for (Node e = f, pred = null;;) {
1309		Object ek, ev;
1310	<	if (e.hash == h &&
1311	<	(ek = e.key) != null &&
1312	<	(ev = e.val) != null &&
451	<	(k == ek || k.equals(ek))) {
1310	>	if ((e.hash & HASH_BITS) == h &&
1311	>	((ev = e.val) != null) &&
1312	>	((ek = e.key) == k || k.equals(ek))) {
1313		if (cv == null || cv == ev || cv.equals(ev)) {
1314		oldVal = ev;
1315		if ((e.val = v) == null) {
#	Line 463 | Line 1324 | public class ConcurrentHashMapV8<K, V>
1324		break;
1325		}
1326		pred = e;
1327	<	} while ((e = e.next) != null);
1327	>	if ((e = e.next) == null)
1328	>	break;
1329	>	}
1330	>	}
1331	>	} finally {
1332	>	if (!f.casHash(fh | LOCKED, fh)) {
1333	>	f.hash = fh;
1334	>	synchronized (f) { f.notifyAll(); };
1335		}
1336		}
1337		if (validated) {
1338		if (deleted)
1339	<	counter.decrement();
1339	>	counter.add(-1L);
1340		break;
1341		}
1342		}
#	Line 476 | Line 1344 | public class ConcurrentHashMapV8<K, V>
1344		return oldVal;
1345		}
1346
1347	<	/** Implementation for computeIfAbsent and compute */
1348	<	@SuppressWarnings("unchecked")
1349	<	private final V internalCompute(K k,
1350	<	MappingFunction<? super K, ? extends V> f,
1351	<	boolean replace) {
1347	>	/*
1348	>	* Internal versions of the six insertion methods, each a
1349	>	* little more complicated than the last. All have
1350	>	* the same basic structure as the first (internalPut):
1351	>	*  1. If table uninitialized, create
1352	>	*  2. If bin empty, try to CAS new node
1353	>	*  3. If bin stale, use new table
1354	>	*  4. if bin converted to TreeBin, validate and relay to TreeBin methods
1355	>	*  5. Lock and validate; if valid, scan and add or update
1356	>	*
1357	>	* The others interweave other checks and/or alternative actions:
1358	>	*  * Plain put checks for and performs resize after insertion.
1359	>	*  * putIfAbsent prescans for mapping without lock (and fails to add
1360	>	*    if present), which also makes pre-emptive resize checks worthwhile.
1361	>	*  * computeIfAbsent extends form used in putIfAbsent with additional
1362	>	*    mechanics to deal with, calls, potential exceptions and null
1363	>	*    returns from function call.
1364	>	*  * compute uses the same function-call mechanics, but without
1365	>	*    the prescans
1366	>	*  * merge acts as putIfAbsent in the absent case, but invokes the
1367	>	*    update function if present
1368	>	*  * putAll attempts to pre-allocate enough table space
1369	>	*    and more lazily performs count updates and checks.
1370	>	*
1371	>	* Someday when details settle down a bit more, it might be worth
1372	>	* some factoring to reduce sprawl.
1373	>	*/
1374	>
1375	>	/** Implementation for put */
1376	>	private final Object internalPut(Object k, Object v) {
1377		int h = spread(k.hashCode());
1378	<	V val = null;
1379	<	boolean added = false;
1380	<	Node[] tab = table;
488	<	for (;;) {
489	<	Node e; int i;
1378	>	int count = 0;
1379	>	for (Node[] tab = table;;) {
1380	>	int i; Node f; int fh; Object fk;
1381		if (tab == null)
1382	<	tab = grow(0);
1383	<	else if ((e = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1384	<	Node node = new Node(h, k, null, null);
1385	<	boolean validated = false;
1386	<	synchronized (node) {
1387	<	if (casTabAt(tab, i, null, node)) {
1388	<	validated = true;
1389	<	try {
1390	<	val = f.map(k);
1391	<	if (val != null) {
1392	<	node.val = val;
1393	<	added = true;
1382	>	tab = initTable();
1383	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1384	>	if (casTabAt(tab, i, null, new Node(h, k, v, null)))
1385	>	break;                   // no lock when adding to empty bin
1386	>	}
1387	>	else if ((fh = f.hash) == MOVED) {
1388	>	if ((fk = f.key) instanceof TreeBin) {
1389	>	TreeBin t = (TreeBin)fk;
1390	>	Object oldVal = null;
1391	>	t.acquire(0);
1392	>	try {
1393	>	if (tabAt(tab, i) == f) {
1394	>	count = 2;
1395	>	TreeNode p = t.putTreeNode(h, k, v);
1396	>	if (p != null) {
1397	>	oldVal = p.val;
1398	>	p.val = v;
1399		}
1400	<	} finally {
1401	<	if (!added)
1402	<	setTabAt(tab, i, null);
1400	>	}
1401	>	} finally {
1402	>	t.release(0);
1403	>	}
1404	>	if (count != 0) {
1405	>	if (oldVal != null)
1406	>	return oldVal;
1407	>	break;
1408	>	}
1409	>	}
1410	>	else
1411	>	tab = (Node[])fk;
1412	>	}
1413	>	else if ((fh & LOCKED) != 0) {
1414	>	checkForResize();
1415	>	f.tryAwaitLock(tab, i);
1416	>	}
1417	>	else if (f.casHash(fh, fh | LOCKED)) {
1418	>	Object oldVal = null;
1419	>	try {                        // needed in case equals() throws
1420	>	if (tabAt(tab, i) == f) {
1421	>	count = 1;
1422	>	for (Node e = f;; ++count) {
1423	>	Object ek, ev;
1424	>	if ((e.hash & HASH_BITS) == h &&
1425	>	(ev = e.val) != null &&
1426	>	((ek = e.key) == k || k.equals(ek))) {
1427	>	oldVal = ev;
1428	>	e.val = v;
1429	>	break;
1430	>	}
1431	>	Node last = e;
1432	>	if ((e = e.next) == null) {
1433	>	last.next = new Node(h, k, v, null);
1434	>	if (count >= TREE_THRESHOLD)
1435	>	replaceWithTreeBin(tab, i, k);
1436	>	break;
1437	>	}
1438	>	}
1439	>	}
1440	>	} finally {                  // unlock and signal if needed
1441	>	if (!f.casHash(fh | LOCKED, fh)) {
1442	>	f.hash = fh;
1443	>	synchronized (f) { f.notifyAll(); };
1444	>	}
1445	>	}
1446	>	if (count != 0) {
1447	>	if (oldVal != null)
1448	>	return oldVal;
1449	>	if (tab.length <= 64)
1450	>	count = 2;
1451	>	break;
1452	>	}
1453	>	}
1454	>	}
1455	>	counter.add(1L);
1456	>	if (count > 1)
1457	>	checkForResize();
1458	>	return null;
1459	>	}
1460	>
1461	>	/** Implementation for putIfAbsent */
1462	>	private final Object internalPutIfAbsent(Object k, Object v) {
1463	>	int h = spread(k.hashCode());
1464	>	int count = 0;
1465	>	for (Node[] tab = table;;) {
1466	>	int i; Node f; int fh; Object fk, fv;
1467	>	if (tab == null)
1468	>	tab = initTable();
1469	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1470	>	if (casTabAt(tab, i, null, new Node(h, k, v, null)))
1471	>	break;
1472	>	}
1473	>	else if ((fh = f.hash) == MOVED) {
1474	>	if ((fk = f.key) instanceof TreeBin) {
1475	>	TreeBin t = (TreeBin)fk;
1476	>	Object oldVal = null;
1477	>	t.acquire(0);
1478	>	try {
1479	>	if (tabAt(tab, i) == f) {
1480	>	count = 2;
1481	>	TreeNode p = t.putTreeNode(h, k, v);
1482	>	if (p != null)
1483	>	oldVal = p.val;
1484	>	}
1485	>	} finally {
1486	>	t.release(0);
1487	>	}
1488	>	if (count != 0) {
1489	>	if (oldVal != null)
1490	>	return oldVal;
1491	>	break;
1492	>	}
1493	>	}
1494	>	else
1495	>	tab = (Node[])fk;
1496	>	}
1497	>	else if ((fh & HASH_BITS) == h && (fv = f.val) != null &&
1498	>	((fk = f.key) == k || k.equals(fk)))
1499	>	return fv;
1500	>	else {
1501	>	Node g = f.next;
1502	>	if (g != null) { // at least 2 nodes -- search and maybe resize
1503	>	for (Node e = g;;) {
1504	>	Object ek, ev;
1505	>	if ((e.hash & HASH_BITS) == h && (ev = e.val) != null &&
1506	>	((ek = e.key) == k || k.equals(ek)))
1507	>	return ev;
1508	>	if ((e = e.next) == null) {
1509	>	checkForResize();
1510	>	break;
1511	>	}
1512	>	}
1513	>	}
1514	>	if (((fh = f.hash) & LOCKED) != 0) {
1515	>	checkForResize();
1516	>	f.tryAwaitLock(tab, i);
1517	>	}
1518	>	else if (tabAt(tab, i) == f && f.casHash(fh, fh | LOCKED)) {
1519	>	Object oldVal = null;
1520	>	try {
1521	>	if (tabAt(tab, i) == f) {
1522	>	count = 1;
1523	>	for (Node e = f;; ++count) {
1524	>	Object ek, ev;
1525	>	if ((e.hash & HASH_BITS) == h &&
1526	>	(ev = e.val) != null &&
1527	>	((ek = e.key) == k || k.equals(ek))) {
1528	>	oldVal = ev;
1529	>	break;
1530	>	}
1531	>	Node last = e;
1532	>	if ((e = e.next) == null) {
1533	>	last.next = new Node(h, k, v, null);
1534	>	if (count >= TREE_THRESHOLD)
1535	>	replaceWithTreeBin(tab, i, k);
1536	>	break;
1537	>	}
1538	>	}
1539	>	}
1540	>	} finally {
1541	>	if (!f.casHash(fh | LOCKED, fh)) {
1542	>	f.hash = fh;
1543	>	synchronized (f) { f.notifyAll(); };
1544		}
1545		}
1546	+	if (count != 0) {
1547	+	if (oldVal != null)
1548	+	return oldVal;
1549	+	if (tab.length <= 64)
1550	+	count = 2;
1551	+	break;
1552	+	}
1553		}
1554	<	if (validated)
1554	>	}
1555	>	}
1556	>	counter.add(1L);
1557	>	if (count > 1)
1558	>	checkForResize();
1559	>	return null;
1560	>	}
1561	>
1562	>	/** Implementation for computeIfAbsent */
1563	>	private final Object internalComputeIfAbsent(K k,
1564	>	Fun<? super K, ?> mf) {
1565	>	int h = spread(k.hashCode());
1566	>	Object val = null;
1567	>	int count = 0;
1568	>	for (Node[] tab = table;;) {
1569	>	Node f; int i, fh; Object fk, fv;
1570	>	if (tab == null)
1571	>	tab = initTable();
1572	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1573	>	Node node = new Node(fh = h | LOCKED, k, null, null);
1574	>	if (casTabAt(tab, i, null, node)) {
1575	>	count = 1;
1576	>	try {
1577	>	if ((val = mf.apply(k)) != null)
1578	>	node.val = val;
1579	>	} finally {
1580	>	if (val == null)
1581	>	setTabAt(tab, i, null);
1582	>	if (!node.casHash(fh, h)) {
1583	>	node.hash = h;
1584	>	synchronized (node) { node.notifyAll(); };
1585	>	}
1586	>	}
1587	>	}
1588	>	if (count != 0)
1589		break;
1590		}
1591	<	else if (e.hash < 0)
1592	<	tab = (Node[])e.key;
1593	<	else if (Thread.holdsLock(e))
1594	<	throw new IllegalStateException("Recursive map computation");
1591	>	else if ((fh = f.hash) == MOVED) {
1592	>	if ((fk = f.key) instanceof TreeBin) {
1593	>	TreeBin t = (TreeBin)fk;
1594	>	boolean added = false;
1595	>	t.acquire(0);
1596	>	try {
1597	>	if (tabAt(tab, i) == f) {
1598	>	count = 1;
1599	>	TreeNode p = t.getTreeNode(h, k, t.root);
1600	>	if (p != null)
1601	>	val = p.val;
1602	>	else if ((val = mf.apply(k)) != null) {
1603	>	added = true;
1604	>	count = 2;
1605	>	t.putTreeNode(h, k, val);
1606	>	}
1607	>	}
1608	>	} finally {
1609	>	t.release(0);
1610	>	}
1611	>	if (count != 0) {
1612	>	if (!added)
1613	>	return val;
1614	>	break;
1615	>	}
1616	>	}
1617	>	else
1618	>	tab = (Node[])fk;
1619	>	}
1620	>	else if ((fh & HASH_BITS) == h && (fv = f.val) != null &&
1621	>	((fk = f.key) == k || k.equals(fk)))
1622	>	return fv;
1623		else {
1624	<	boolean validated = false;
1625	<	boolean checkSize = false;
1626	<	synchronized (e) {
1627	<	if (tabAt(tab, i) == e) {
1628	<	validated = true;
1629	<	for (Node first = e;;) {
1630	<	Object ek, ev, fv;
1631	<	if (e.hash == h &&
1632	<	(ek = e.key) != null &&
1624	>	Node g = f.next;
1625	>	if (g != null) {
1626	>	for (Node e = g;;) {
1627	>	Object ek, ev;
1628	>	if ((e.hash & HASH_BITS) == h && (ev = e.val) != null &&
1629	>	((ek = e.key) == k || k.equals(ek)))
1630	>	return ev;
1631	>	if ((e = e.next) == null) {
1632	>	checkForResize();
1633	>	break;
1634	>	}
1635	>	}
1636	>	}
1637	>	if (((fh = f.hash) & LOCKED) != 0) {
1638	>	checkForResize();
1639	>	f.tryAwaitLock(tab, i);
1640	>	}
1641	>	else if (tabAt(tab, i) == f && f.casHash(fh, fh | LOCKED)) {
1642	>	boolean added = false;
1643	>	try {
1644	>	if (tabAt(tab, i) == f) {
1645	>	count = 1;
1646	>	for (Node e = f;; ++count) {
1647	>	Object ek, ev;
1648	>	if ((e.hash & HASH_BITS) == h &&
1649	>	(ev = e.val) != null &&
1650	>	((ek = e.key) == k || k.equals(ek))) {
1651	>	val = ev;
1652	>	break;
1653	>	}
1654	>	Node last = e;
1655	>	if ((e = e.next) == null) {
1656	>	if ((val = mf.apply(k)) != null) {
1657	>	added = true;
1658	>	last.next = new Node(h, k, val, null);
1659	>	if (count >= TREE_THRESHOLD)
1660	>	replaceWithTreeBin(tab, i, k);
1661	>	}
1662	>	break;
1663	>	}
1664	>	}
1665	>	}
1666	>	} finally {
1667	>	if (!f.casHash(fh | LOCKED, fh)) {
1668	>	f.hash = fh;
1669	>	synchronized (f) { f.notifyAll(); };
1670	>	}
1671	>	}
1672	>	if (count != 0) {
1673	>	if (!added)
1674	>	return val;
1675	>	if (tab.length <= 64)
1676	>	count = 2;
1677	>	break;
1678	>	}
1679	>	}
1680	>	}
1681	>	}
1682	>	if (val != null) {
1683	>	counter.add(1L);
1684	>	if (count > 1)
1685	>	checkForResize();
1686	>	}
1687	>	return val;
1688	>	}
1689	>
1690	>	/** Implementation for compute */
1691	>	@SuppressWarnings("unchecked") private final Object internalCompute
1692	>	(K k, boolean onlyIfPresent, BiFun<? super K, ? super V, ? extends V> mf) {
1693	>	int h = spread(k.hashCode());
1694	>	Object val = null;
1695	>	int delta = 0;
1696	>	int count = 0;
1697	>	for (Node[] tab = table;;) {
1698	>	Node f; int i, fh; Object fk;
1699	>	if (tab == null)
1700	>	tab = initTable();
1701	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1702	>	if (onlyIfPresent)
1703	>	break;
1704	>	Node node = new Node(fh = h | LOCKED, k, null, null);
1705	>	if (casTabAt(tab, i, null, node)) {
1706	>	try {
1707	>	count = 1;
1708	>	if ((val = mf.apply(k, null)) != null) {
1709	>	node.val = val;
1710	>	delta = 1;
1711	>	}
1712	>	} finally {
1713	>	if (delta == 0)
1714	>	setTabAt(tab, i, null);
1715	>	if (!node.casHash(fh, h)) {
1716	>	node.hash = h;
1717	>	synchronized (node) { node.notifyAll(); };
1718	>	}
1719	>	}
1720	>	}
1721	>	if (count != 0)
1722	>	break;
1723	>	}
1724	>	else if ((fh = f.hash) == MOVED) {
1725	>	if ((fk = f.key) instanceof TreeBin) {
1726	>	TreeBin t = (TreeBin)fk;
1727	>	t.acquire(0);
1728	>	try {
1729	>	if (tabAt(tab, i) == f) {
1730	>	count = 1;
1731	>	TreeNode p = t.getTreeNode(h, k, t.root);
1732	>	Object pv = (p == null) ? null : p.val;
1733	>	if ((val = mf.apply(k, (V)pv)) != null) {
1734	>	if (p != null)
1735	>	p.val = val;
1736	>	else {
1737	>	count = 2;
1738	>	delta = 1;
1739	>	t.putTreeNode(h, k, val);
1740	>	}
1741	>	}
1742	>	else if (p != null) {
1743	>	delta = -1;
1744	>	t.deleteTreeNode(p);
1745	>	}
1746	>	}
1747	>	} finally {
1748	>	t.release(0);
1749	>	}
1750	>	if (count != 0)
1751	>	break;
1752	>	}
1753	>	else
1754	>	tab = (Node[])fk;
1755	>	}
1756	>	else if ((fh & LOCKED) != 0) {
1757	>	checkForResize();
1758	>	f.tryAwaitLock(tab, i);
1759	>	}
1760	>	else if (f.casHash(fh, fh | LOCKED)) {
1761	>	try {
1762	>	if (tabAt(tab, i) == f) {
1763	>	count = 1;
1764	>	for (Node e = f, pred = null;; ++count) {
1765	>	Object ek, ev;
1766	>	if ((e.hash & HASH_BITS) == h &&
1767		(ev = e.val) != null &&
1768	<	(k == ek || k.equals(ek))) {
1769	<	if (replace && (fv = f.map(k)) != null)
1770	<	ev = e.val = fv;
1771	<	val = (V)ev;
1768	>	((ek = e.key) == k || k.equals(ek))) {
1769	>	val = mf.apply(k, (V)ev);
1770	>	if (val != null)
1771	>	e.val = val;
1772	>	else {
1773	>	delta = -1;
1774	>	Node en = e.next;
1775	>	if (pred != null)
1776	>	pred.next = en;
1777	>	else
1778	>	setTabAt(tab, i, en);
1779	>	}
1780		break;
1781		}
1782	<	Node last = e;
1782	>	pred = e;
1783		if ((e = e.next) == null) {
1784	<	if ((val = f.map(k)) != null) {
1785	<	last.next = new Node(h, k, val, null);
1786	<	added = true;
1787	<	if (last != first || tab.length <= 64)
1788	<	checkSize = true;
1784	>	if (!onlyIfPresent && (val = mf.apply(k, null)) != null) {
1785	>	pred.next = new Node(h, k, val, null);
1786	>	delta = 1;
1787	>	if (count >= TREE_THRESHOLD)
1788	>	replaceWithTreeBin(tab, i, k);
1789		}
1790		break;
1791		}
1792		}
1793		}
1794	+	} finally {
1795	+	if (!f.casHash(fh | LOCKED, fh)) {
1796	+	f.hash = fh;
1797	+	synchronized (f) { f.notifyAll(); };
1798	+	}
1799		}
1800	<	if (validated) {
1801	<	if (checkSize && tab.length < MAXIMUM_CAPACITY &&
1802	<	resizing == 0 && counter.sum() >= threshold)
550	<	grow(0);
1800	>	if (count != 0) {
1801	>	if (tab.length <= 64)
1802	>	count = 2;
1803		break;
1804		}
1805		}
1806		}
1807	<	if (added)
1808	<	counter.increment();
1807	>	if (delta != 0) {
1808	>	counter.add((long)delta);
1809	>	if (count > 1)
1810	>	checkForResize();
1811	>	}
1812		return val;
1813		}
1814
1815	<	/*
1816	<	* Reclassifies nodes in each bin to new table.  Because we are
1817	<	* using power-of-two expansion, the elements from each bin must
1818	<	* either stay at same index, or move with a power of two
1819	<	* offset. We eliminate unnecessary node creation by catching
1820	<	* cases where old nodes can be reused because their next fields
1821	<	* won't change.  Statistically, at the default threshold, only
1822	<	* about one-sixth of them need cloning when a table doubles. The
1823	<	* nodes they replace will be garbage collectable as soon as they
1824	<	* are no longer referenced by any reader thread that may be in
1825	<	* the midst of concurrently traversing table.
1826	<	*
1827	<	* Transfers are done from the bottom up to preserve iterator
1828	<	* traversability. On each step, the old bin is locked,
1829	<	* moved/copied, and then replaced with a forwarding node.
1830	<	*/
576	<	private static final void transfer(Node[] tab, Node[] nextTab) {
577	<	int n = tab.length;
578	<	int mask = nextTab.length - 1;
579	<	Node fwd = new Node(MOVED, nextTab, null, null);
580	<	for (int i = n - 1; i >= 0; --i) {
581	<	for (Node e;;) {
582	<	if ((e = tabAt(tab, i)) == null) {
583	<	if (casTabAt(tab, i, e, fwd))
584	<	break;
1815	>	/** Implementation for merge */
1816	>	@SuppressWarnings("unchecked") private final Object internalMerge
1817	>	(K k, V v, BiFun<? super V, ? super V, ? extends V> mf) {
1818	>	int h = spread(k.hashCode());
1819	>	Object val = null;
1820	>	int delta = 0;
1821	>	int count = 0;
1822	>	for (Node[] tab = table;;) {
1823	>	int i; Node f; int fh; Object fk, fv;
1824	>	if (tab == null)
1825	>	tab = initTable();
1826	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1827	>	if (casTabAt(tab, i, null, new Node(h, k, v, null))) {
1828	>	delta = 1;
1829	>	val = v;
1830	>	break;
1831		}
1832	<	else {
1833	<	int idx = e.hash & mask;
1834	<	boolean validated = false;
1835	<	synchronized (e) {
1836	<	if (tabAt(tab, i) == e) {
1837	<	validated = true;
1838	<	Node lastRun = e;
1839	<	for (Node p = e.next; p != null; p = p.next) {
1840	<	int j = p.hash & mask;
1841	<	if (j != idx) {
1842	<	idx = j;
1843	<	lastRun = p;
1832	>	}
1833	>	else if ((fh = f.hash) == MOVED) {
1834	>	if ((fk = f.key) instanceof TreeBin) {
1835	>	TreeBin t = (TreeBin)fk;
1836	>	t.acquire(0);
1837	>	try {
1838	>	if (tabAt(tab, i) == f) {
1839	>	count = 1;
1840	>	TreeNode p = t.getTreeNode(h, k, t.root);
1841	>	val = (p == null) ? v : mf.apply((V)p.val, v);
1842	>	if (val != null) {
1843	>	if (p != null)
1844	>	p.val = val;
1845	>	else {
1846	>	count = 2;
1847	>	delta = 1;
1848	>	t.putTreeNode(h, k, val);
1849		}
1850		}
1851	<	relaxedSetTabAt(nextTab, idx, lastRun);
1852	<	for (Node p = e; p != lastRun; p = p.next) {
1853	<	int h = p.hash;
603	<	int j = h & mask;
604	<	Node r = relaxedTabAt(nextTab, j);
605	<	relaxedSetTabAt(nextTab, j,
606	<	new Node(h, p.key, p.val, r));
1851	>	else if (p != null) {
1852	>	delta = -1;
1853	>	t.deleteTreeNode(p);
1854		}
608	–	setTabAt(tab, i, fwd);
1855		}
1856	+	} finally {
1857	+	t.release(0);
1858		}
1859	<	if (validated)
1859	>	if (count != 0)
1860		break;
1861		}
1862	+	else
1863	+	tab = (Node[])fk;
1864	+	}
1865	+	else if ((fh & LOCKED) != 0) {
1866	+	checkForResize();
1867	+	f.tryAwaitLock(tab, i);
1868	+	}
1869	+	else if (f.casHash(fh, fh | LOCKED)) {
1870	+	try {
1871	+	if (tabAt(tab, i) == f) {
1872	+	count = 1;
1873	+	for (Node e = f, pred = null;; ++count) {
1874	+	Object ek, ev;
1875	+	if ((e.hash & HASH_BITS) == h &&
1876	+	(ev = e.val) != null &&
1877	+	((ek = e.key) == k || k.equals(ek))) {
1878	+	val = mf.apply(v, (V)ev);
1879	+	if (val != null)
1880	+	e.val = val;
1881	+	else {
1882	+	delta = -1;
1883	+	Node en = e.next;
1884	+	if (pred != null)
1885	+	pred.next = en;
1886	+	else
1887	+	setTabAt(tab, i, en);
1888	+	}
1889	+	break;
1890	+	}
1891	+	pred = e;
1892	+	if ((e = e.next) == null) {
1893	+	val = v;
1894	+	pred.next = new Node(h, k, val, null);
1895	+	delta = 1;
1896	+	if (count >= TREE_THRESHOLD)
1897	+	replaceWithTreeBin(tab, i, k);
1898	+	break;
1899	+	}
1900	+	}
1901	+	}
1902	+	} finally {
1903	+	if (!f.casHash(fh | LOCKED, fh)) {
1904	+	f.hash = fh;
1905	+	synchronized (f) { f.notifyAll(); };
1906	+	}
1907	+	}
1908	+	if (count != 0) {
1909	+	if (tab.length <= 64)
1910	+	count = 2;
1911	+	break;
1912	+	}
1913		}
1914		}
1915	+	if (delta != 0) {
1916	+	counter.add((long)delta);
1917	+	if (count > 1)
1918	+	checkForResize();
1919	+	}
1920	+	return val;
1921		}
1922
1923	+	/** Implementation for putAll */
1924	+	private final void internalPutAll(Map<?, ?> m) {
1925	+	tryPresize(m.size());
1926	+	long delta = 0L;     // number of uncommitted additions
1927	+	boolean npe = false; // to throw exception on exit for nulls
1928	+	try {                // to clean up counts on other exceptions
1929	+	for (Map.Entry<?, ?> entry : m.entrySet()) {
1930	+	Object k, v;
1931	+	if (entry == null || (k = entry.getKey()) == null ||
1932	+	(v = entry.getValue()) == null) {
1933	+	npe = true;
1934	+	break;
1935	+	}
1936	+	int h = spread(k.hashCode());
1937	+	for (Node[] tab = table;;) {
1938	+	int i; Node f; int fh; Object fk;
1939	+	if (tab == null)
1940	+	tab = initTable();
1941	+	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null){
1942	+	if (casTabAt(tab, i, null, new Node(h, k, v, null))) {
1943	+	++delta;
1944	+	break;
1945	+	}
1946	+	}
1947	+	else if ((fh = f.hash) == MOVED) {
1948	+	if ((fk = f.key) instanceof TreeBin) {
1949	+	TreeBin t = (TreeBin)fk;
1950	+	boolean validated = false;
1951	+	t.acquire(0);
1952	+	try {
1953	+	if (tabAt(tab, i) == f) {
1954	+	validated = true;
1955	+	TreeNode p = t.getTreeNode(h, k, t.root);
1956	+	if (p != null)
1957	+	p.val = v;
1958	+	else {
1959	+	t.putTreeNode(h, k, v);
1960	+	++delta;
1961	+	}
1962	+	}
1963	+	} finally {
1964	+	t.release(0);
1965	+	}
1966	+	if (validated)
1967	+	break;
1968	+	}
1969	+	else
1970	+	tab = (Node[])fk;
1971	+	}
1972	+	else if ((fh & LOCKED) != 0) {
1973	+	counter.add(delta);
1974	+	delta = 0L;
1975	+	checkForResize();
1976	+	f.tryAwaitLock(tab, i);
1977	+	}
1978	+	else if (f.casHash(fh, fh | LOCKED)) {
1979	+	int count = 0;
1980	+	try {
1981	+	if (tabAt(tab, i) == f) {
1982	+	count = 1;
1983	+	for (Node e = f;; ++count) {
1984	+	Object ek, ev;
1985	+	if ((e.hash & HASH_BITS) == h &&
1986	+	(ev = e.val) != null &&
1987	+	((ek = e.key) == k || k.equals(ek))) {
1988	+	e.val = v;
1989	+	break;
1990	+	}
1991	+	Node last = e;
1992	+	if ((e = e.next) == null) {
1993	+	++delta;
1994	+	last.next = new Node(h, k, v, null);
1995	+	if (count >= TREE_THRESHOLD)
1996	+	replaceWithTreeBin(tab, i, k);
1997	+	break;
1998	+	}
1999	+	}
2000	+	}
2001	+	} finally {
2002	+	if (!f.casHash(fh | LOCKED, fh)) {
2003	+	f.hash = fh;
2004	+	synchronized (f) { f.notifyAll(); };
2005	+	}
2006	+	}
2007	+	if (count != 0) {
2008	+	if (count > 1) {
2009	+	counter.add(delta);
2010	+	delta = 0L;
2011	+	checkForResize();
2012	+	}
2013	+	break;
2014	+	}
2015	+	}
2016	+	}
2017	+	}
2018	+	} finally {
2019	+	if (delta != 0)
2020	+	counter.add(delta);
2021	+	}
2022	+	if (npe)
2023	+	throw new NullPointerException();
2024	+	}
2025	+
2026	+	/* ---------------- Table Initialization and Resizing -------------- */
2027	+
2028		/**
2029	<	* If not already resizing, initializes or creates next table and
2030	<	* transfers bins. Rechecks occupancy after a transfer to see if
621	<	* another resize is already needed because resizings are lagging
622	<	* additions.
623	<	*
624	<	* @param sizeHint overridden capacity target (nonzero only from putAll)
625	<	* @return current table
2029	>	* Returns a power of two table size for the given desired capacity.
2030	>	* See Hackers Delight, sec 3.2
2031		*/
2032	<	private final Node[] grow(int sizeHint) {
2033	<	if (resizing == 0 &&
2034	<	UNSAFE.compareAndSwapInt(this, resizingOffset, 0, 1)) {
2035	<	try {
2036	<	for (;;) {
2037	<	int cap, n;
2038	<	Node[] tab = table;
2039	<	if (tab == null) {
2040	<	int c = initCap;
2041	<	if (c < sizeHint)
2042	<	c = sizeHint;
2043	<	if (c == DEFAULT_CAPACITY)
2044	<	cap = c;
2045	<	else if (c >= MAXIMUM_CAPACITY)
2046	<	cap = MAXIMUM_CAPACITY;
2047	<	else {
2048	<	cap = MINIMUM_CAPACITY;
2049	<	while (cap < c)
2050	<	cap <<= 1;
2051	<	}
2032	>	private static final int tableSizeFor(int c) {
2033	>	int n = c - 1;
2034	>	n |= n >>> 1;
2035	>	n |= n >>> 2;
2036	>	n |= n >>> 4;
2037	>	n |= n >>> 8;
2038	>	n |= n >>> 16;
2039	>	return (n < 0) ? 1 : (n >= MAXIMUM_CAPACITY) ? MAXIMUM_CAPACITY : n + 1;
2040	>	}
2041	>
2042	>	/**
2043	>	* Initializes table, using the size recorded in sizeCtl.
2044	>	*/
2045	>	private final Node[] initTable() {
2046	>	Node[] tab; int sc;
2047	>	while ((tab = table) == null) {
2048	>	if ((sc = sizeCtl) < 0)
2049	>	Thread.yield(); // lost initialization race; just spin
2050	>	else if (UNSAFE.compareAndSwapInt(this, sizeCtlOffset, sc, -1)) {
2051	>	try {
2052	>	if ((tab = table) == null) {
2053	>	int n = (sc > 0) ? sc : DEFAULT_CAPACITY;
2054	>	tab = table = new Node[n];
2055	>	sc = n - (n >>> 2);
2056		}
2057	<	else if ((n = tab.length) < MAXIMUM_CAPACITY &&
2058	<	(sizeHint <= 0 || n < sizeHint))
650	<	cap = n << 1;
651	<	else
652	<	break;
653	<	threshold = (int)(cap * loadFactor) - THRESHOLD_OFFSET;
654	<	Node[] nextTab = new Node[cap];
655	<	if (tab != null)
656	<	transfer(tab, nextTab);
657	<	table = nextTab;
658	<	if (tab == null || cap >= MAXIMUM_CAPACITY ||
659	<	((sizeHint > 0) ? cap >= sizeHint :
660	<	counter.sum() < threshold))
661	<	break;
2057	>	} finally {
2058	>	sizeCtl = sc;
2059		}
2060	<	} finally {
664	<	resizing = 0;
2060	>	break;
2061		}
2062		}
2063	<	else if (table == null)
668	<	Thread.yield(); // lost initialization race; just spin
669	<	return table;
2063	>	return tab;
2064		}
2065
2066		/**
2067	<	* Implementation for putAll and constructor with Map
2068	<	* argument. Tries to first override initial capacity or grow
2069	<	* based on map size to pre-allocate table space.
2067	>	* If table is too small and not already resizing, creates next
2068	>	* table and transfers bins.  Rechecks occupancy after a transfer
2069	>	* to see if another resize is already needed because resizings
2070	>	* are lagging additions.
2071		*/
2072	<	private final void internalPutAll(Map<? extends K, ? extends V> m) {
2073	<	int s = m.size();
2074	<	grow((s >= (MAXIMUM_CAPACITY >>> 1)) ? s : s + (s >>> 1));
2075	<	for (Map.Entry<? extends K, ? extends V> e : m.entrySet()) {
2076	<	Object k = e.getKey();
2077	<	Object v = e.getValue();
2078	<	if (k == null || v == null)
2079	<	throw new NullPointerException();
2080	<	internalPut(k, v, true);
2072	>	private final void checkForResize() {
2073	>	Node[] tab; int n, sc;
2074	>	while ((tab = table) != null &&
2075	>	(n = tab.length) < MAXIMUM_CAPACITY &&
2076	>	(sc = sizeCtl) >= 0 && counter.sum() >= (long)sc &&
2077	>	UNSAFE.compareAndSwapInt(this, sizeCtlOffset, sc, -1)) {
2078	>	try {
2079	>	if (tab == table) {
2080	>	table = rebuild(tab);
2081	>	sc = (n << 1) - (n >>> 1);
2082	>	}
2083	>	} finally {
2084	>	sizeCtl = sc;
2085	>	}
2086		}
2087		}
2088
2089		/**
2090	<	* Implementation for clear. Steps through each bin, removing all nodes.
2090	>	* Tries to presize table to accommodate the given number of elements.
2091	>	*
2092	>	* @param size number of elements (doesn't need to be perfectly accurate)
2093		*/
2094	<	private final void internalClear() {
2095	<	long deletions = 0L;
2096	<	int i = 0;
2097	<	Node[] tab = table;
2098	<	while (tab != null && i < tab.length) {
2099	<	Node e = tabAt(tab, i);
2100	<	if (e == null)
2101	<	++i;
2102	<	else if (e.hash < 0)
2103	<	tab = (Node[])e.key;
2104	<	else {
2094	>	private final void tryPresize(int size) {
2095	>	int c = (size >= (MAXIMUM_CAPACITY >>> 1)) ? MAXIMUM_CAPACITY :
2096	>	tableSizeFor(size + (size >>> 1) + 1);
2097	>	int sc;
2098	>	while ((sc = sizeCtl) >= 0) {
2099	>	Node[] tab = table; int n;
2100	>	if (tab == null || (n = tab.length) == 0) {
2101	>	n = (sc > c) ? sc : c;
2102	>	if (UNSAFE.compareAndSwapInt(this, sizeCtlOffset, sc, -1)) {
2103	>	try {
2104	>	if (table == tab) {
2105	>	table = new Node[n];
2106	>	sc = n - (n >>> 2);
2107	>	}
2108	>	} finally {
2109	>	sizeCtl = sc;
2110	>	}
2111	>	}
2112	>	}
2113	>	else if (c <= sc || n >= MAXIMUM_CAPACITY)
2114	>	break;
2115	>	else if (UNSAFE.compareAndSwapInt(this, sizeCtlOffset, sc, -1)) {
2116	>	try {
2117	>	if (table == tab) {
2118	>	table = rebuild(tab);
2119	>	sc = (n << 1) - (n >>> 1);
2120	>	}
2121	>	} finally {
2122	>	sizeCtl = sc;
2123	>	}
2124	>	}
2125	>	}
2126	>	}
2127	>
2128	>	/*
2129	>	* Moves and/or copies the nodes in each bin to new table. See
2130	>	* above for explanation.
2131	>	*
2132	>	* @return the new table
2133	>	*/
2134	>	private static final Node[] rebuild(Node[] tab) {
2135	>	int n = tab.length;
2136	>	Node[] nextTab = new Node[n << 1];
2137	>	Node fwd = new Node(MOVED, nextTab, null, null);
2138	>	int[] buffer = null;       // holds bins to revisit; null until needed
2139	>	Node rev = null;           // reverse forwarder; null until needed
2140	>	int nbuffered = 0;         // the number of bins in buffer list
2141	>	int bufferIndex = 0;       // buffer index of current buffered bin
2142	>	int bin = n - 1;           // current non-buffered bin or -1 if none
2143	>
2144	>	for (int i = bin;;) {      // start upwards sweep
2145	>	int fh; Node f;
2146	>	if ((f = tabAt(tab, i)) == null) {
2147	>	if (bin >= 0) {    // Unbuffered; no lock needed (or available)
2148	>	if (!casTabAt(tab, i, f, fwd))
2149	>	continue;
2150	>	}
2151	>	else {             // transiently use a locked forwarding node
2152	>	Node g = new Node(MOVED|LOCKED, nextTab, null, null);
2153	>	if (!casTabAt(tab, i, f, g))
2154	>	continue;
2155	>	setTabAt(nextTab, i, null);
2156	>	setTabAt(nextTab, i + n, null);
2157	>	setTabAt(tab, i, fwd);
2158	>	if (!g.casHash(MOVED|LOCKED, MOVED)) {
2159	>	g.hash = MOVED;
2160	>	synchronized (g) { g.notifyAll(); }
2161	>	}
2162	>	}
2163	>	}
2164	>	else if ((fh = f.hash) == MOVED) {
2165	>	Object fk = f.key;
2166	>	if (fk instanceof TreeBin) {
2167	>	TreeBin t = (TreeBin)fk;
2168	>	boolean validated = false;
2169	>	t.acquire(0);
2170	>	try {
2171	>	if (tabAt(tab, i) == f) {
2172	>	validated = true;
2173	>	splitTreeBin(nextTab, i, t);
2174	>	setTabAt(tab, i, fwd);
2175	>	}
2176	>	} finally {
2177	>	t.release(0);
2178	>	}
2179	>	if (!validated)
2180	>	continue;
2181	>	}
2182	>	}
2183	>	else if ((fh & LOCKED) == 0 && f.casHash(fh, fh|LOCKED)) {
2184		boolean validated = false;
2185	<	synchronized (e) {
2186	<	if (tabAt(tab, i) == e) {
2185	>	try {              // split to lo and hi lists; copying as needed
2186	>	if (tabAt(tab, i) == f) {
2187		validated = true;
2188	<	do {
2189	<	if (e.val != null) {
709	<	e.val = null;
710	<	++deletions;
711	<	}
712	<	} while ((e = e.next) != null);
713	<	setTabAt(tab, i, null);
2188	>	splitBin(nextTab, i, f);
2189	>	setTabAt(tab, i, fwd);
2190		}
2191	<	}
2192	<	if (validated) {
2193	<	++i;
2194	<	if (deletions > THRESHOLD_OFFSET) { // bound lag in counts
719	<	counter.add(-deletions);
720	<	deletions = 0L;
2191	>	} finally {
2192	>	if (!f.casHash(fh | LOCKED, fh)) {
2193	>	f.hash = fh;
2194	>	synchronized (f) { f.notifyAll(); };
2195		}
2196		}
2197	+	if (!validated)
2198	+	continue;
2199	+	}
2200	+	else {
2201	+	if (buffer == null) // initialize buffer for revisits
2202	+	buffer = new int[TRANSFER_BUFFER_SIZE];
2203	+	if (bin < 0 && bufferIndex > 0) {
2204	+	int j = buffer[--bufferIndex];
2205	+	buffer[bufferIndex] = i;
2206	+	i = j;         // swap with another bin
2207	+	continue;
2208	+	}
2209	+	if (bin < 0 || nbuffered >= TRANSFER_BUFFER_SIZE) {
2210	+	f.tryAwaitLock(tab, i);
2211	+	continue;      // no other options -- block
2212	+	}
2213	+	if (rev == null)   // initialize reverse-forwarder
2214	+	rev = new Node(MOVED, tab, null, null);
2215	+	if (tabAt(tab, i) != f || (f.hash & LOCKED) == 0)
2216	+	continue;      // recheck before adding to list
2217	+	buffer[nbuffered++] = i;
2218	+	setTabAt(nextTab, i, rev);     // install place-holders
2219	+	setTabAt(nextTab, i + n, rev);
2220	+	}
2221	+
2222	+	if (bin > 0)
2223	+	i = --bin;
2224	+	else if (buffer != null && nbuffered > 0) {
2225	+	bin = -1;
2226	+	i = buffer[bufferIndex = --nbuffered];
2227		}
2228	+	else
2229	+	return nextTab;
2230		}
725	–	if (deletions != 0L)
726	–	counter.add(-deletions);
2231		}
2232
2233		/**
2234	<	* Base class for key, value, and entry iterators, plus internal
2235	<	* implementations of public traversal-based methods, to avoid
732	<	* duplicating traversal code.
2234	>	* Splits a normal bin with list headed by e into lo and hi parts;
2235	>	* installs in given table.
2236		*/
2237	<	class HashIterator {
2238	<	private Node next;          // the next entry to return
2239	<	private Node[] tab;         // current table; updated if resized
2240	<	private Node lastReturned;  // the last entry returned, for remove
2241	<	private Object nextVal;     // cached value of next
2242	<	private int index;          // index of bin to use next
2243	<	private int baseIndex;      // current index of initial table
2244	<	private final int baseSize; // initial table size
2237	>	private static void splitBin(Node[] nextTab, int i, Node e) {
2238	>	int bit = nextTab.length >>> 1; // bit to split on
2239	>	int runBit = e.hash & bit;
2240	>	Node lastRun = e, lo = null, hi = null;
2241	>	for (Node p = e.next; p != null; p = p.next) {
2242	>	int b = p.hash & bit;
2243	>	if (b != runBit) {
2244	>	runBit = b;
2245	>	lastRun = p;
2246	>	}
2247	>	}
2248	>	if (runBit == 0)
2249	>	lo = lastRun;
2250	>	else
2251	>	hi = lastRun;
2252	>	for (Node p = e; p != lastRun; p = p.next) {
2253	>	int ph = p.hash & HASH_BITS;
2254	>	Object pk = p.key, pv = p.val;
2255	>	if ((ph & bit) == 0)
2256	>	lo = new Node(ph, pk, pv, lo);
2257	>	else
2258	>	hi = new Node(ph, pk, pv, hi);
2259	>	}
2260	>	setTabAt(nextTab, i, lo);
2261	>	setTabAt(nextTab, i + bit, hi);
2262	>	}
2263
2264	<	HashIterator() {
2265	<	Node[] t = tab = table;
2266	<	if (t == null)
2267	<	baseSize = 0;
2264	>	/**
2265	>	* Splits a tree bin into lo and hi parts; installs in given table.
2266	>	*/
2267	>	private static void splitTreeBin(Node[] nextTab, int i, TreeBin t) {
2268	>	int bit = nextTab.length >>> 1;
2269	>	TreeBin lt = new TreeBin();
2270	>	TreeBin ht = new TreeBin();
2271	>	int lc = 0, hc = 0;
2272	>	for (Node e = t.first; e != null; e = e.next) {
2273	>	int h = e.hash & HASH_BITS;
2274	>	Object k = e.key, v = e.val;
2275	>	if ((h & bit) == 0) {
2276	>	++lc;
2277	>	lt.putTreeNode(h, k, v);
2278	>	}
2279		else {
2280	<	baseSize = t.length;
2281	<	advance(null);
2280	>	++hc;
2281	>	ht.putTreeNode(h, k, v);
2282		}
2283		}
2284	+	Node ln, hn; // throw away trees if too small
2285	+	if (lc <= (TREE_THRESHOLD >>> 1)) {
2286	+	ln = null;
2287	+	for (Node p = lt.first; p != null; p = p.next)
2288	+	ln = new Node(p.hash, p.key, p.val, ln);
2289	+	}
2290	+	else
2291	+	ln = new Node(MOVED, lt, null, null);
2292	+	setTabAt(nextTab, i, ln);
2293	+	if (hc <= (TREE_THRESHOLD >>> 1)) {
2294	+	hn = null;
2295	+	for (Node p = ht.first; p != null; p = p.next)
2296	+	hn = new Node(p.hash, p.key, p.val, hn);
2297	+	}
2298	+	else
2299	+	hn = new Node(MOVED, ht, null, null);
2300	+	setTabAt(nextTab, i + bit, hn);
2301	+	}
2302
2303	<	public final boolean hasNext()         { return next != null; }
2304	<	public final boolean hasMoreElements() { return next != null; }
2305	<
2306	<	/**
2307	<	* Advances next.  Normally, iteration proceeds bin-by-bin
2308	<	* traversing lists.  However, if the table has been resized,
2309	<	* then all future steps must traverse both the bin at the
2310	<	* current index as well as at (index + baseSize); and so on
2311	<	* for further resizings. To paranoically cope with potential
2312	<	* (improper) sharing of iterators across threads, table reads
2313	<	* are bounds-checked.
2314	<	*/
2315	<	final void advance(Node e) {
2316	<	for (;;) {
2317	<	Node[] t; int i; // for bounds checks
2318	<	if (e != null) {
2319	<	Object ek = e.key, ev = e.val;
2320	<	if (ev != null && ek != null) {
2321	<	nextVal = ev;
2322	<	next = e;
2323	<	break;
2303	>	/**
2304	>	* Implementation for clear. Steps through each bin, removing all
2305	>	* nodes.
2306	>	*/
2307	>	private final void internalClear() {
2308	>	long delta = 0L; // negative number of deletions
2309	>	int i = 0;
2310	>	Node[] tab = table;
2311	>	while (tab != null && i < tab.length) {
2312	>	int fh; Object fk;
2313	>	Node f = tabAt(tab, i);
2314	>	if (f == null)
2315	>	++i;
2316	>	else if ((fh = f.hash) == MOVED) {
2317	>	if ((fk = f.key) instanceof TreeBin) {
2318	>	TreeBin t = (TreeBin)fk;
2319	>	t.acquire(0);
2320	>	try {
2321	>	if (tabAt(tab, i) == f) {
2322	>	for (Node p = t.first; p != null; p = p.next) {
2323	>	if (p.val != null) { // (currently always true)
2324	>	p.val = null;
2325	>	--delta;
2326	>	}
2327	>	}
2328	>	t.first = null;
2329	>	t.root = null;
2330	>	++i;
2331	>	}
2332	>	} finally {
2333	>	t.release(0);
2334		}
775	–	e = e.next;
2335		}
2336	<	else if (baseIndex < baseSize && (t = tab) != null &&
2337	<	t.length > (i = index) && i >= 0) {
2338	<	if ((e = tabAt(t, i)) != null && e.hash < 0) {
2339	<	tab = (Node[])e.key;
2340	<	e = null;
2336	>	else
2337	>	tab = (Node[])fk;
2338	>	}
2339	>	else if ((fh & LOCKED) != 0) {
2340	>	counter.add(delta); // opportunistically update count
2341	>	delta = 0L;
2342	>	f.tryAwaitLock(tab, i);
2343	>	}
2344	>	else if (f.casHash(fh, fh | LOCKED)) {
2345	>	try {
2346	>	if (tabAt(tab, i) == f) {
2347	>	for (Node e = f; e != null; e = e.next) {
2348	>	if (e.val != null) {  // (currently always true)
2349	>	e.val = null;
2350	>	--delta;
2351	>	}
2352	>	}
2353	>	setTabAt(tab, i, null);
2354	>	++i;
2355	>	}
2356	>	} finally {
2357	>	if (!f.casHash(fh | LOCKED, fh)) {
2358	>	f.hash = fh;
2359	>	synchronized (f) { f.notifyAll(); };
2360		}
783	–	else if (i + baseSize < t.length)
784	–	index += baseSize;    // visit forwarded upper slots
785	–	else
786	–	index = ++baseIndex;
787	–	}
788	–	else {
789	–	next = null;
790	–	break;
2361		}
2362		}
2363		}
2364	+	if (delta != 0)
2365	+	counter.add(delta);
2366	+	}
2367
2368	<	final Object nextKey() {
796	<	Node e = next;
797	<	if (e == null)
798	<	throw new NoSuchElementException();
799	<	Object k = e.key;
800	<	advance((lastReturned = e).next);
801	<	return k;
802	<	}
2368	>	/* ----------------Table Traversal -------------- */
2369
2370	<	final Object nextValue() {
2371	<	Node e = next;
2372	<	if (e == null)
2373	<	throw new NoSuchElementException();
2374	<	Object v = nextVal;
2375	<	advance((lastReturned = e).next);
2376	<	return v;
2370	>	/**
2371	>	* Encapsulates traversal for methods such as containsValue; also
2372	>	* serves as a base class for other iterators and bulk tasks.
2373	>	*
2374	>	* At each step, the iterator snapshots the key ("nextKey") and
2375	>	* value ("nextVal") of a valid node (i.e., one that, at point of
2376	>	* snapshot, has a non-null user value). Because val fields can
2377	>	* change (including to null, indicating deletion), field nextVal
2378	>	* might not be accurate at point of use, but still maintains the
2379	>	* weak consistency property of holding a value that was once
2380	>	* valid. To support iterator.remove, the nextKey field is not
2381	>	* updated (nulled out) when the iterator cannot advance.
2382	>	*
2383	>	* Internal traversals directly access these fields, as in:
2384	>	* {@code while (it.advance() != null) { process(it.nextKey); }}
2385	>	*
2386	>	* Exported iterators must track whether the iterator has advanced
2387	>	* (in hasNext vs next) (by setting/checking/nulling field
2388	>	* nextVal), and then extract key, value, or key-value pairs as
2389	>	* return values of next().
2390	>	*
2391	>	* The iterator visits once each still-valid node that was
2392	>	* reachable upon iterator construction. It might miss some that
2393	>	* were added to a bin after the bin was visited, which is OK wrt
2394	>	* consistency guarantees. Maintaining this property in the face
2395	>	* of possible ongoing resizes requires a fair amount of
2396	>	* bookkeeping state that is difficult to optimize away amidst
2397	>	* volatile accesses.  Even so, traversal maintains reasonable
2398	>	* throughput.
2399	>	*
2400	>	* Normally, iteration proceeds bin-by-bin traversing lists.
2401	>	* However, if the table has been resized, then all future steps
2402	>	* must traverse both the bin at the current index as well as at
2403	>	* (index + baseSize); and so on for further resizings. To
2404	>	* paranoically cope with potential sharing by users of iterators
2405	>	* across threads, iteration terminates if a bounds checks fails
2406	>	* for a table read.
2407	>	*
2408	>	* This class extends CountedCompleter to streamline parallel
2409	>	* iteration in bulk operations. This adds only a few fields of
2410	>	* space overhead, which is small enough in cases where it is not
2411	>	* needed to not worry about it.  Because CountedCompleter is
2412	>	* Serializable, but iterators need not be, we need to add warning
2413	>	* suppressions.
2414	>	*/
2415	>	@SuppressWarnings("serial") static class Traverser<K,V,R> extends CountedCompleter<R> {
2416	>	final ConcurrentHashMapV8<K, V> map;
2417	>	Node next;           // the next entry to use
2418	>	Object nextKey;      // cached key field of next
2419	>	Object nextVal;      // cached val field of next
2420	>	Node[] tab;          // current table; updated if resized
2421	>	int index;           // index of bin to use next
2422	>	int baseIndex;       // current index of initial table
2423	>	int baseLimit;       // index bound for initial table
2424	>	int baseSize;        // initial table size
2425	>	int batch;           // split control
2426	>
2427	>	/** Creates iterator for all entries in the table. */
2428	>	Traverser(ConcurrentHashMapV8<K, V> map) {
2429	>	this.map = map;
2430	>	}
2431	>
2432	>	/** Creates iterator for split() methods and task constructors */
2433	>	Traverser(ConcurrentHashMapV8<K,V> map, Traverser<K,V,?> it, int batch) {
2434	>	super(it);
2435	>	this.batch = batch;
2436	>	if ((this.map = map) != null && it != null) { // split parent
2437	>	Node[] t;
2438	>	if ((t = it.tab) == null &&
2439	>	(t = it.tab = map.table) != null)
2440	>	it.baseLimit = it.baseSize = t.length;
2441	>	this.tab = t;
2442	>	this.baseSize = it.baseSize;
2443	>	int hi = this.baseLimit = it.baseLimit;
2444	>	it.baseLimit = this.index = this.baseIndex =
2445	>	(hi + it.baseIndex + 1) >>> 1;
2446	>	}
2447		}
2448
2449	<	final WriteThroughEntry nextEntry() {
2449	>	/**
2450	>	* Advances next; returns nextVal or null if terminated.
2451	>	* See above for explanation.
2452	>	*/
2453	>	final Object advance() {
2454		Node e = next;
2455	<	if (e == null)
2456	<	throw new NoSuchElementException();
2457	<	WriteThroughEntry entry =
2458	<	new WriteThroughEntry(e.key, nextVal);
2459	<	advance((lastReturned = e).next);
2460	<	return entry;
2455	>	Object ev = null;
2456	>	outer: do {
2457	>	if (e != null)                  // advance past used/skipped node
2458	>	e = e.next;
2459	>	while (e == null) {             // get to next non-null bin
2460	>	ConcurrentHashMapV8<K, V> m;
2461	>	Node[] t; int b, i, n; Object ek; // checks must use locals
2462	>	if ((t = tab) != null)
2463	>	n = t.length;
2464	>	else if ((m = map) != null && (t = tab = m.table) != null)
2465	>	n = baseLimit = baseSize = t.length;
2466	>	else
2467	>	break outer;
2468	>	if ((b = baseIndex) >= baseLimit ||
2469	>	(i = index) < 0 || i >= n)
2470	>	break outer;
2471	>	if ((e = tabAt(t, i)) != null && e.hash == MOVED) {
2472	>	if ((ek = e.key) instanceof TreeBin)
2473	>	e = ((TreeBin)ek).first;
2474	>	else {
2475	>	tab = (Node[])ek;
2476	>	continue;           // restarts due to null val
2477	>	}
2478	>	}                           // visit upper slots if present
2479	>	index = (i += baseSize) < n ? i : (baseIndex = b + 1);
2480	>	}
2481	>	nextKey = e.key;
2482	>	} while ((ev = e.val) == null);    // skip deleted or special nodes
2483	>	next = e;
2484	>	return nextVal = ev;
2485		}
2486
2487		public final void remove() {
2488	<	if (lastReturned == null)
2488	>	Object k = nextKey;
2489	>	if (k == null && (advance() == null || (k = nextKey) == null))
2490		throw new IllegalStateException();
2491	<	ConcurrentHashMapV8.this.remove(lastReturned.key);
827	<	lastReturned = null;
2491	>	map.internalReplace(k, null, null);
2492		}
2493
2494	<	/** Helper for serialization */
2495	<	final void writeEntries(java.io.ObjectOutputStream s)
832	<	throws java.io.IOException {
833	<	Node e;
834	<	while ((e = next) != null) {
835	<	s.writeObject(e.key);
836	<	s.writeObject(nextVal);
837	<	advance(e.next);
838	<	}
2494	>	public final boolean hasNext() {
2495	>	return nextVal != null || advance() != null;
2496		}
2497
2498	<	/** Helper for containsValue */
842	<	final boolean containsVal(Object value) {
843	<	if (value != null) {
844	<	Node e;
845	<	while ((e = next) != null) {
846	<	Object v = nextVal;
847	<	if (value == v || value.equals(v))
848	<	return true;
849	<	advance(e.next);
850	<	}
851	<	}
852	<	return false;
853	<	}
2498	>	public final boolean hasMoreElements() { return hasNext(); }
2499
2500	<	/** Helper for Map.hashCode */
856	<	final int mapHashCode() {
857	<	int h = 0;
858	<	Node e;
859	<	while ((e = next) != null) {
860	<	h += e.key.hashCode() ^ nextVal.hashCode();
861	<	advance(e.next);
862	<	}
863	<	return h;
864	<	}
2500	>	public void compute() { } // default no-op CountedCompleter body
2501
2502	<	/** Helper for Map.toString */
2503	<	final String mapToString() {
2504	<	Node e = next;
2505	<	if (e == null)
2506	<	return "{}";
2507	<	StringBuilder sb = new StringBuilder();
2508	<	sb.append('{');
2509	<	for (;;) {
2510	<	sb.append(e.key   == this ? "(this Map)" : e.key);
2511	<	sb.append('=');
2512	<	sb.append(nextVal == this ? "(this Map)" : nextVal);
2513	<	advance(e.next);
2514	<	if ((e = next) != null)
2515	<	sb.append(',').append(' ');
2516	<	else
2517	<	return sb.append('}').toString();
2502	>	/**
2503	>	* Returns a batch value > 0 if this task should (and must) be
2504	>	* split, if so, adding to pending count, and in any case
2505	>	* updating batch value. The initial batch value is approx
2506	>	* exp2 of the number of times (minus one) to split task by
2507	>	* two before executing leaf action. This value is faster to
2508	>	* compute and more convenient to use as a guide to splitting
2509	>	* than is the depth, since it is used while dividing by two
2510	>	* anyway.
2511	>	*/
2512	>	final int preSplit() {
2513	>	ConcurrentHashMapV8<K, V> m; int b; Node[] t;  ForkJoinPool pool;
2514	>	if ((b = batch) < 0 && (m = map) != null) { // force initialization
2515	>	if ((t = tab) == null && (t = tab = m.table) != null)
2516	>	baseLimit = baseSize = t.length;
2517	>	if (t != null) {
2518	>	long n = m.counter.sum();
2519	>	int par = ((pool = getPool()) == null) ?
2520	>	ForkJoinPool.getCommonPoolParallelism() :
2521	>	pool.getParallelism();
2522	>	int sp = par << 3; // slack of 8
2523	>	b = (n <= 0L) ? 0 : (n < (long)sp) ? (int)n : sp;
2524	>	}
2525		}
2526	+	b = (b <= 1 || baseIndex == baseLimit) ? 0 : (b >>> 1);
2527	+	if ((batch = b) > 0)
2528	+	addToPendingCount(1);
2529	+	return b;
2530		}
2531	+
2532		}
2533
2534		/* ---------------- Public operations -------------- */
2535
2536		/**
2537	<	* Creates a new, empty map with the specified initial
890	<	* capacity, load factor and concurrency level.
891	<	*
892	<	* @param initialCapacity the initial capacity. The implementation
893	<	* performs internal sizing to accommodate this many elements.
894	<	* @param loadFactor  the load factor threshold, used to control resizing.
895	<	* Resizing may be performed when the average number of elements per
896	<	* bin exceeds this threshold.
897	<	* @param concurrencyLevel the estimated number of concurrently
898	<	* updating threads. The implementation may use this value as
899	<	* a sizing hint.
900	<	* @throws IllegalArgumentException if the initial capacity is
901	<	* negative or the load factor or concurrencyLevel are
902	<	* nonpositive.
2537	>	* Creates a new, empty map with the default initial table size (16).
2538		*/
2539	<	public ConcurrentHashMapV8(int initialCapacity,
905	<	float loadFactor, int concurrencyLevel) {
906	<	if (!(loadFactor > 0) || initialCapacity < 0 || concurrencyLevel <= 0)
907	<	throw new IllegalArgumentException();
908	<	this.initCap = initialCapacity;
909	<	this.loadFactor = loadFactor;
2539	>	public ConcurrentHashMapV8() {
2540		this.counter = new LongAdder();
2541		}
2542
2543		/**
2544	<	* Creates a new, empty map with the specified initial capacity
2545	<	* and load factor and with the default concurrencyLevel (16).
2544	>	* Creates a new, empty map with an initial table size
2545	>	* accommodating the specified number of elements without the need
2546	>	* to dynamically resize.
2547		*
2548		* @param initialCapacity The implementation performs internal
2549		* sizing to accommodate this many elements.
2550	<	* @param loadFactor  the load factor threshold, used to control resizing.
2551	<	* Resizing may be performed when the average number of elements per
2552	<	* bin exceeds this threshold.
2550	>	* @throws IllegalArgumentException if the initial capacity of
2551	>	* elements is negative
2552	>	*/
2553	>	public ConcurrentHashMapV8(int initialCapacity) {
2554	>	if (initialCapacity < 0)
2555	>	throw new IllegalArgumentException();
2556	>	int cap = ((initialCapacity >= (MAXIMUM_CAPACITY >>> 1)) ?
2557	>	MAXIMUM_CAPACITY :
2558	>	tableSizeFor(initialCapacity + (initialCapacity >>> 1) + 1));
2559	>	this.counter = new LongAdder();
2560	>	this.sizeCtl = cap;
2561	>	}
2562	>
2563	>	/**
2564	>	* Creates a new map with the same mappings as the given map.
2565	>	*
2566	>	* @param m the map
2567	>	*/
2568	>	public ConcurrentHashMapV8(Map<? extends K, ? extends V> m) {
2569	>	this.counter = new LongAdder();
2570	>	this.sizeCtl = DEFAULT_CAPACITY;
2571	>	internalPutAll(m);
2572	>	}
2573	>
2574	>	/**
2575	>	* Creates a new, empty map with an initial table size based on
2576	>	* the given number of elements ({@code initialCapacity}) and
2577	>	* initial table density ({@code loadFactor}).
2578	>	*
2579	>	* @param initialCapacity the initial capacity. The implementation
2580	>	* performs internal sizing to accommodate this many elements,
2581	>	* given the specified load factor.
2582	>	* @param loadFactor the load factor (table density) for
2583	>	* establishing the initial table size
2584		* @throws IllegalArgumentException if the initial capacity of
2585		* elements is negative or the load factor is nonpositive
2586		*
2587		* @since 1.6
2588		*/
2589		public ConcurrentHashMapV8(int initialCapacity, float loadFactor) {
2590	<	this(initialCapacity, loadFactor, DEFAULT_CONCURRENCY_LEVEL);
2590	>	this(initialCapacity, loadFactor, 1);
2591		}
2592
2593		/**
2594	<	* Creates a new, empty map with the specified initial capacity,
2595	<	* and with default load factor (0.75) and concurrencyLevel (16).
2594	>	* Creates a new, empty map with an initial table size based on
2595	>	* the given number of elements ({@code initialCapacity}), table
2596	>	* density ({@code loadFactor}), and number of concurrently
2597	>	* updating threads ({@code concurrencyLevel}).
2598		*
2599		* @param initialCapacity the initial capacity. The implementation
2600	<	* performs internal sizing to accommodate this many elements.
2601	<	* @throws IllegalArgumentException if the initial capacity of
2602	<	* elements is negative.
2600	>	* performs internal sizing to accommodate this many elements,
2601	>	* given the specified load factor.
2602	>	* @param loadFactor the load factor (table density) for
2603	>	* establishing the initial table size
2604	>	* @param concurrencyLevel the estimated number of concurrently
2605	>	* updating threads. The implementation may use this value as
2606	>	* a sizing hint.
2607	>	* @throws IllegalArgumentException if the initial capacity is
2608	>	* negative or the load factor or concurrencyLevel are
2609	>	* nonpositive
2610		*/
2611	<	public ConcurrentHashMapV8(int initialCapacity) {
2612	<	this(initialCapacity, DEFAULT_LOAD_FACTOR, DEFAULT_CONCURRENCY_LEVEL);
2611	>	public ConcurrentHashMapV8(int initialCapacity,
2612	>	float loadFactor, int concurrencyLevel) {
2613	>	if (!(loadFactor > 0.0f) || initialCapacity < 0 || concurrencyLevel <= 0)
2614	>	throw new IllegalArgumentException();
2615	>	if (initialCapacity < concurrencyLevel)   // Use at least as many bins
2616	>	initialCapacity = concurrencyLevel;   // as estimated threads
2617	>	long size = (long)(1.0 + (long)initialCapacity / loadFactor);
2618	>	int cap = (size >= (long)MAXIMUM_CAPACITY) ?
2619	>	MAXIMUM_CAPACITY : tableSizeFor((int)size);
2620	>	this.counter = new LongAdder();
2621	>	this.sizeCtl = cap;
2622		}
2623
2624		/**
2625	<	* Creates a new, empty map with a default initial capacity (16),
2626	<	* load factor (0.75) and concurrencyLevel (16).
2625	>	* Creates a new {@link Set} backed by a ConcurrentHashMapV8
2626	>	* from the given type to {@code Boolean.TRUE}.
2627	>	*
2628	>	* @return the new set
2629		*/
2630	<	public ConcurrentHashMapV8() {
2631	<	this(DEFAULT_CAPACITY, DEFAULT_LOAD_FACTOR, DEFAULT_CONCURRENCY_LEVEL);
2630	>	public static <K> KeySetView<K,Boolean> newKeySet() {
2631	>	return new KeySetView<K,Boolean>(new ConcurrentHashMapV8<K,Boolean>(),
2632	>	Boolean.TRUE);
2633		}
2634
2635		/**
2636	<	* Creates a new map with the same mappings as the given map.
2637	<	* The map is created with a capacity of 1.5 times the number
955	<	* of mappings in the given map or 16 (whichever is greater),
956	<	* and a default load factor (0.75) and concurrencyLevel (16).
2636	>	* Creates a new {@link Set} backed by a ConcurrentHashMapV8
2637	>	* from the given type to {@code Boolean.TRUE}.
2638		*
2639	<	* @param m the map
2639	>	* @param initialCapacity The implementation performs internal
2640	>	* sizing to accommodate this many elements.
2641	>	* @throws IllegalArgumentException if the initial capacity of
2642	>	* elements is negative
2643	>	* @return the new set
2644		*/
2645	<	public ConcurrentHashMapV8(Map<? extends K, ? extends V> m) {
2646	<	this(DEFAULT_CAPACITY, DEFAULT_LOAD_FACTOR, DEFAULT_CONCURRENCY_LEVEL);
2647	<	if (m == null)
963	<	throw new NullPointerException();
964	<	internalPutAll(m);
2645	>	public static <K> KeySetView<K,Boolean> newKeySet(int initialCapacity) {
2646	>	return new KeySetView<K,Boolean>(new ConcurrentHashMapV8<K,Boolean>(initialCapacity),
2647	>	Boolean.TRUE);
2648		}
2649
2650		/**
2651	<	* Returns {@code true} if this map contains no key-value mappings.
969	<	*
970	<	* @return {@code true} if this map contains no key-value mappings
2651	>	* {@inheritDoc}
2652		*/
2653		public boolean isEmpty() {
2654		return counter.sum() <= 0L; // ignore transient negative values
2655		}
2656
2657		/**
2658	<	* Returns the number of key-value mappings in this map.  If the
978	<	* map contains more than {@code Integer.MAX_VALUE} elements, returns
979	<	* {@code Integer.MAX_VALUE}.
980	<	*
981	<	* @return the number of key-value mappings in this map
2658	>	* {@inheritDoc}
2659		*/
2660		public int size() {
2661		long n = counter.sum();
2662	<	return ((n >>> 31) == 0) ? (int)n : (n < 0L) ? 0 : Integer.MAX_VALUE;
2662	>	return ((n < 0L) ? 0 :
2663	>	(n > (long)Integer.MAX_VALUE) ? Integer.MAX_VALUE :
2664	>	(int)n);
2665	>	}
2666	>
2667	>	/**
2668	>	* Returns the number of mappings. This method should be used
2669	>	* instead of {@link #size} because a ConcurrentHashMapV8 may
2670	>	* contain more mappings than can be represented as an int. The
2671	>	* value returned is an estimate; the actual count may differ if
2672	>	* there are concurrent insertions or removals.
2673	>	*
2674	>	* @return the number of mappings
2675	>	*/
2676	>	public long mappingCount() {
2677	>	long n = counter.sum();
2678	>	return (n < 0L) ? 0L : n; // ignore transient negative values
2679		}
2680
2681		/**
#	Line 996 | Line 2689 | public class ConcurrentHashMapV8<K, V>
2689		*
2690		* @throws NullPointerException if the specified key is null
2691		*/
2692	<	@SuppressWarnings("unchecked")
1000	<	public V get(Object key) {
2692	>	@SuppressWarnings("unchecked") public V get(Object key) {
2693		if (key == null)
2694		throw new NullPointerException();
2695		return (V)internalGet(key);
2696		}
2697
2698		/**
2699	+	* Returns the value to which the specified key is mapped,
2700	+	* or the given defaultValue if this map contains no mapping for the key.
2701	+	*
2702	+	* @param key the key
2703	+	* @param defaultValue the value to return if this map contains
2704	+	* no mapping for the given key
2705	+	* @return the mapping for the key, if present; else the defaultValue
2706	+	* @throws NullPointerException if the specified key is null
2707	+	*/
2708	+	@SuppressWarnings("unchecked") public V getValueOrDefault(Object key, V defaultValue) {
2709	+	if (key == null)
2710	+	throw new NullPointerException();
2711	+	V v = (V) internalGet(key);
2712	+	return v == null ? defaultValue : v;
2713	+	}
2714	+
2715	+	/**
2716		* Tests if the specified object is a key in this table.
2717		*
2718		* @param  key   possible key
2719		* @return {@code true} if and only if the specified object
2720		*         is a key in this table, as determined by the
2721	<	*         {@code equals} method; {@code false} otherwise.
2721	>	*         {@code equals} method; {@code false} otherwise
2722		* @throws NullPointerException if the specified key is null
2723		*/
2724		public boolean containsKey(Object key) {
#	Line 1020 | Line 2729 | public class ConcurrentHashMapV8<K, V>
2729
2730		/**
2731		* Returns {@code true} if this map maps one or more keys to the
2732	<	* specified value. Note: This method requires a full internal
2733	<	* traversal of the hash table, and so is much slower than
1025	<	* method {@code containsKey}.
2732	>	* specified value. Note: This method may require a full traversal
2733	>	* of the map, and is much slower than method {@code containsKey}.
2734		*
2735		* @param value value whose presence in this map is to be tested
2736		* @return {@code true} if this map maps one or more keys to the
#	Line 1032 | Line 2740 | public class ConcurrentHashMapV8<K, V>
2740		public boolean containsValue(Object value) {
2741		if (value == null)
2742		throw new NullPointerException();
2743	<	return new HashIterator().containsVal(value);
2743	>	Object v;
2744	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
2745	>	while ((v = it.advance()) != null) {
2746	>	if (v == value || value.equals(v))
2747	>	return true;
2748	>	}
2749	>	return false;
2750		}
2751
2752		/**
#	Line 1058 | Line 2772 | public class ConcurrentHashMapV8<K, V>
2772		* Maps the specified key to the specified value in this table.
2773		* Neither the key nor the value can be null.
2774		*
2775	<	* <p> The value can be retrieved by calling the {@code get} method
2775	>	* <p>The value can be retrieved by calling the {@code get} method
2776		* with a key that is equal to the original key.
2777		*
2778		* @param key key with which the specified value is to be associated
#	Line 1067 | Line 2781 | public class ConcurrentHashMapV8<K, V>
2781		*         {@code null} if there was no mapping for {@code key}
2782		* @throws NullPointerException if the specified key or value is null
2783		*/
2784	<	@SuppressWarnings("unchecked")
1071	<	public V put(K key, V value) {
2784	>	@SuppressWarnings("unchecked") public V put(K key, V value) {
2785		if (key == null || value == null)
2786		throw new NullPointerException();
2787	<	return (V)internalPut(key, value, true);
2787	>	return (V)internalPut(key, value);
2788		}
2789
2790		/**
#	Line 1081 | Line 2794 | public class ConcurrentHashMapV8<K, V>
2794		*         or {@code null} if there was no mapping for the key
2795		* @throws NullPointerException if the specified key or value is null
2796		*/
2797	<	@SuppressWarnings("unchecked")
1085	<	public V putIfAbsent(K key, V value) {
2797	>	@SuppressWarnings("unchecked") public V putIfAbsent(K key, V value) {
2798		if (key == null || value == null)
2799		throw new NullPointerException();
2800	<	return (V)internalPut(key, value, false);
2800	>	return (V)internalPutIfAbsent(key, value);
2801		}
2802
2803		/**
#	Line 1096 | Line 2808 | public class ConcurrentHashMapV8<K, V>
2808		* @param m mappings to be stored in this map
2809		*/
2810		public void putAll(Map<? extends K, ? extends V> m) {
1099	–	if (m == null)
1100	–	throw new NullPointerException();
2811		internalPutAll(m);
2812		}
2813
2814		/**
2815		* If the specified key is not already associated with a value,
2816	<	* computes its value using the given mappingFunction, and if
2817	<	* non-null, enters it into the map.  This is equivalent to
2818	<	*  <pre> {@code
2816	>	* computes its value using the given mappingFunction and enters
2817	>	* it into the map unless null.  This is equivalent to
2818	>	* <pre> {@code
2819		* if (map.containsKey(key))
2820		*   return map.get(key);
2821	<	* value = mappingFunction.map(key);
2821	>	* value = mappingFunction.apply(key);
2822		* if (value != null)
2823		*   map.put(key, value);
2824		* return value;}</pre>
2825		*
2826	<	* except that the action is performed atomically.  Some attempted
2827	<	* update operations on this map by other threads may be blocked
2828	<	* while computation is in progress, so the computation should be
2829	<	* short and simple, and must not attempt to update any other
2830	<	* mappings of this Map. The most appropriate usage is to
2826	>	* except that the action is performed atomically.  If the
2827	>	* function returns {@code null} no mapping is recorded. If the
2828	>	* function itself throws an (unchecked) exception, the exception
2829	>	* is rethrown to its caller, and no mapping is recorded.  Some
2830	>	* attempted update operations on this map by other threads may be
2831	>	* blocked while computation is in progress, so the computation
2832	>	* should be short and simple, and must not attempt to update any
2833	>	* other mappings of this Map. The most appropriate usage is to
2834		* construct a new object serving as an initial mapped value, or
2835		* memoized result, as in:
2836	+	*
2837		*  <pre> {@code
2838	<	* map.computeIfAbsent(key, new MappingFunction<K, V>() {
2838	>	* map.computeIfAbsent(key, new Fun<K, V>() {
2839		*   public V map(K k) { return new Value(f(k)); }});}</pre>
2840		*
2841		* @param key key with which the specified value is to be associated
2842		* @param mappingFunction the function to compute a value
2843		* @return the current (existing or computed) value associated with
2844	<	*         the specified key, or {@code null} if the computation
1131	<	*         returned {@code null}.
2844	>	*         the specified key, or null if the computed value is null
2845		* @throws NullPointerException if the specified key or mappingFunction
2846	<	*         is null,
2846	>	*         is null
2847		* @throws IllegalStateException if the computation detectably
2848		*         attempts a recursive update to this map that would
2849	<	*         otherwise never complete.
2849	>	*         otherwise never complete
2850		* @throws RuntimeException or Error if the mappingFunction does so,
2851	<	*         in which case the mapping is left unestablished.
2851	>	*         in which case the mapping is left unestablished
2852		*/
2853	<	public V computeIfAbsent(K key, MappingFunction<? super K, ? extends V> mappingFunction) {
2853	>	@SuppressWarnings("unchecked") public V computeIfAbsent
2854	>	(K key, Fun<? super K, ? extends V> mappingFunction) {
2855		if (key == null || mappingFunction == null)
2856		throw new NullPointerException();
2857	<	return internalCompute(key, mappingFunction, false);
2857	>	return (V)internalComputeIfAbsent(key, mappingFunction);
2858		}
2859
2860		/**
2861	<	* Computes the value associated with the given key using the given
2862	<	* mappingFunction, and if non-null, enters it into the map.  This
1149	<	* is equivalent to
2861	>	* If the given key is present, computes a new mapping value given a key and
2862	>	* its current mapped value. This is equivalent to
2863		*  <pre> {@code
2864	<	* value = mappingFunction.map(key);
2865	<	* if (value != null)
2866	<	*   map.put(key, value);
2867	<	* else
2868	<	*   value = map.get(key);
2869	<	* return value;}</pre>
2864	>	*   if (map.containsKey(key)) {
2865	>	*     value = remappingFunction.apply(key, map.get(key));
2866	>	*     if (value != null)
2867	>	*       map.put(key, value);
2868	>	*     else
2869	>	*       map.remove(key);
2870	>	*   }
2871	>	* }</pre>
2872	>	*
2873	>	* except that the action is performed atomically.  If the
2874	>	* function returns {@code null}, the mapping is removed.  If the
2875	>	* function itself throws an (unchecked) exception, the exception
2876	>	* is rethrown to its caller, and the current mapping is left
2877	>	* unchanged.  Some attempted update operations on this map by
2878	>	* other threads may be blocked while computation is in progress,
2879	>	* so the computation should be short and simple, and must not
2880	>	* attempt to update any other mappings of this Map. For example,
2881	>	* to either create or append new messages to a value mapping:
2882		*
2883	<	* except that the action is performed atomically.  Some attempted
2884	<	* update operations on this map by other threads may be blocked
2885	<	* while computation is in progress, so the computation should be
2886	<	* short and simple, and must not attempt to update any other
2887	<	* mappings of this Map.
2883	>	* @param key key with which the specified value is to be associated
2884	>	* @param remappingFunction the function to compute a value
2885	>	* @return the new value associated with the specified key, or null if none
2886	>	* @throws NullPointerException if the specified key or remappingFunction
2887	>	*         is null
2888	>	* @throws IllegalStateException if the computation detectably
2889	>	*         attempts a recursive update to this map that would
2890	>	*         otherwise never complete
2891	>	* @throws RuntimeException or Error if the remappingFunction does so,
2892	>	*         in which case the mapping is unchanged
2893	>	*/
2894	>	@SuppressWarnings("unchecked") public V computeIfPresent
2895	>	(K key, BiFun<? super K, ? super V, ? extends V> remappingFunction) {
2896	>	if (key == null || remappingFunction == null)
2897	>	throw new NullPointerException();
2898	>	return (V)internalCompute(key, true, remappingFunction);
2899	>	}
2900	>
2901	>	/**
2902	>	* Computes a new mapping value given a key and
2903	>	* its current mapped value (or {@code null} if there is no current
2904	>	* mapping). This is equivalent to
2905	>	*  <pre> {@code
2906	>	*   value = remappingFunction.apply(key, map.get(key));
2907	>	*   if (value != null)
2908	>	*     map.put(key, value);
2909	>	*   else
2910	>	*     map.remove(key);
2911	>	* }</pre>
2912	>	*
2913	>	* except that the action is performed atomically.  If the
2914	>	* function returns {@code null}, the mapping is removed.  If the
2915	>	* function itself throws an (unchecked) exception, the exception
2916	>	* is rethrown to its caller, and the current mapping is left
2917	>	* unchanged.  Some attempted update operations on this map by
2918	>	* other threads may be blocked while computation is in progress,
2919	>	* so the computation should be short and simple, and must not
2920	>	* attempt to update any other mappings of this Map. For example,
2921	>	* to either create or append new messages to a value mapping:
2922	>	*
2923	>	* <pre> {@code
2924	>	* Map<Key, String> map = ...;
2925	>	* final String msg = ...;
2926	>	* map.compute(key, new BiFun<Key, String, String>() {
2927	>	*   public String apply(Key k, String v) {
2928	>	*    return (v == null) ? msg : v + msg;});}}</pre>
2929		*
2930		* @param key key with which the specified value is to be associated
2931	<	* @param mappingFunction the function to compute a value
2932	<	* @return the current value associated with
2933	<	*         the specified key, or {@code null} if the computation
2934	<	*         returned {@code null} and the value was not otherwise present.
1169	<	* @throws NullPointerException if the specified key or mappingFunction
1170	<	*         is null,
2931	>	* @param remappingFunction the function to compute a value
2932	>	* @return the new value associated with the specified key, or null if none
2933	>	* @throws NullPointerException if the specified key or remappingFunction
2934	>	*         is null
2935		* @throws IllegalStateException if the computation detectably
2936		*         attempts a recursive update to this map that would
2937	<	*         otherwise never complete.
2938	<	* @throws RuntimeException or Error if the mappingFunction does so,
2939	<	*         in which case the mapping is unchanged.
2940	<	*/
2941	<	public V compute(K key, MappingFunction<? super K, ? extends V> mappingFunction) {
2942	<	if (key == null || mappingFunction == null)
2937	>	*         otherwise never complete
2938	>	* @throws RuntimeException or Error if the remappingFunction does so,
2939	>	*         in which case the mapping is unchanged
2940	>	*/
2941	>	@SuppressWarnings("unchecked") public V compute
2942	>	(K key, BiFun<? super K, ? super V, ? extends V> remappingFunction) {
2943	>	if (key == null || remappingFunction == null)
2944	>	throw new NullPointerException();
2945	>	return (V)internalCompute(key, false, remappingFunction);
2946	>	}
2947	>
2948	>	/**
2949	>	* If the specified key is not already associated
2950	>	* with a value, associate it with the given value.
2951	>	* Otherwise, replace the value with the results of
2952	>	* the given remapping function. This is equivalent to:
2953	>	*  <pre> {@code
2954	>	*   if (!map.containsKey(key))
2955	>	*     map.put(value);
2956	>	*   else {
2957	>	*     newValue = remappingFunction.apply(map.get(key), value);
2958	>	*     if (value != null)
2959	>	*       map.put(key, value);
2960	>	*     else
2961	>	*       map.remove(key);
2962	>	*   }
2963	>	* }</pre>
2964	>	* except that the action is performed atomically.  If the
2965	>	* function returns {@code null}, the mapping is removed.  If the
2966	>	* function itself throws an (unchecked) exception, the exception
2967	>	* is rethrown to its caller, and the current mapping is left
2968	>	* unchanged.  Some attempted update operations on this map by
2969	>	* other threads may be blocked while computation is in progress,
2970	>	* so the computation should be short and simple, and must not
2971	>	* attempt to update any other mappings of this Map.
2972	>	*/
2973	>	@SuppressWarnings("unchecked") public V merge
2974	>	(K key, V value, BiFun<? super V, ? super V, ? extends V> remappingFunction) {
2975	>	if (key == null || value == null || remappingFunction == null)
2976		throw new NullPointerException();
2977	<	return internalCompute(key, mappingFunction, true);
2977	>	return (V)internalMerge(key, value, remappingFunction);
2978		}
2979
2980		/**
#	Line 1189 | Line 2986 | public class ConcurrentHashMapV8<K, V>
2986		*         {@code null} if there was no mapping for {@code key}
2987		* @throws NullPointerException if the specified key is null
2988		*/
2989	<	@SuppressWarnings("unchecked")
1193	<	public V remove(Object key) {
2989	>	@SuppressWarnings("unchecked") public V remove(Object key) {
2990		if (key == null)
2991		throw new NullPointerException();
2992		return (V)internalReplace(key, null, null);
#	Line 1227 | Line 3023 | public class ConcurrentHashMapV8<K, V>
3023		*         or {@code null} if there was no mapping for the key
3024		* @throws NullPointerException if the specified key or value is null
3025		*/
3026	<	@SuppressWarnings("unchecked")
1231	<	public V replace(K key, V value) {
3026	>	@SuppressWarnings("unchecked") public V replace(K key, V value) {
3027		if (key == null || value == null)
3028		throw new NullPointerException();
3029		return (V)internalReplace(key, value, null);
#	Line 1244 | Line 3039 | public class ConcurrentHashMapV8<K, V>
3039		/**
3040		* Returns a {@link Set} view of the keys contained in this map.
3041		* The set is backed by the map, so changes to the map are
3042	<	* reflected in the set, and vice-versa.  The set supports element
1248	<	* removal, which removes the corresponding mapping from this map,
1249	<	* via the {@code Iterator.remove}, {@code Set.remove},
1250	<	* {@code removeAll}, {@code retainAll}, and {@code clear}
1251	<	* operations.  It does not support the {@code add} or
1252	<	* {@code addAll} operations.
3042	>	* reflected in the set, and vice-versa.
3043		*
3044	<	* <p>The view's {@code iterator} is a "weakly consistent" iterator
1255	<	* that will never throw {@link ConcurrentModificationException},
1256	<	* and guarantees to traverse elements as they existed upon
1257	<	* construction of the iterator, and may (but is not guaranteed to)
1258	<	* reflect any modifications subsequent to construction.
3044	>	* @return the set view
3045		*/
3046	<	public Set<K> keySet() {
3047	<	Set<K> ks = keySet;
3048	<	return (ks != null) ? ks : (keySet = new KeySet());
3046	>	public KeySetView<K,V> keySet() {
3047	>	KeySetView<K,V> ks = keySet;
3048	>	return (ks != null) ? ks : (keySet = new KeySetView<K,V>(this, null));
3049	>	}
3050	>
3051	>	/**
3052	>	* Returns a {@link Set} view of the keys in this map, using the
3053	>	* given common mapped value for any additions (i.e., {@link
3054	>	* Collection#add} and {@link Collection#addAll}). This is of
3055	>	* course only appropriate if it is acceptable to use the same
3056	>	* value for all additions from this view.
3057	>	*
3058	>	* @param mappedValue the mapped value to use for any
3059	>	* additions.
3060	>	* @return the set view
3061	>	* @throws NullPointerException if the mappedValue is null
3062	>	*/
3063	>	public KeySetView<K,V> keySet(V mappedValue) {
3064	>	if (mappedValue == null)
3065	>	throw new NullPointerException();
3066	>	return new KeySetView<K,V>(this, mappedValue);
3067		}
3068
3069		/**
3070		* Returns a {@link Collection} view of the values contained in this map.
3071		* The collection is backed by the map, so changes to the map are
3072	<	* reflected in the collection, and vice-versa.  The collection
1269	<	* supports element removal, which removes the corresponding
1270	<	* mapping from this map, via the {@code Iterator.remove},
1271	<	* {@code Collection.remove}, {@code removeAll},
1272	<	* {@code retainAll}, and {@code clear} operations.  It does not
1273	<	* support the {@code add} or {@code addAll} operations.
1274	<	*
1275	<	* <p>The view's {@code iterator} is a "weakly consistent" iterator
1276	<	* that will never throw {@link ConcurrentModificationException},
1277	<	* and guarantees to traverse elements as they existed upon
1278	<	* construction of the iterator, and may (but is not guaranteed to)
1279	<	* reflect any modifications subsequent to construction.
3072	>	* reflected in the collection, and vice-versa.
3073		*/
3074	<	public Collection<V> values() {
3075	<	Collection<V> vs = values;
3076	<	return (vs != null) ? vs : (values = new Values());
3074	>	public ValuesView<K,V> values() {
3075	>	ValuesView<K,V> vs = values;
3076	>	return (vs != null) ? vs : (values = new ValuesView<K,V>(this));
3077		}
3078
3079		/**
#	Line 1300 | Line 3093 | public class ConcurrentHashMapV8<K, V>
3093		* reflect any modifications subsequent to construction.
3094		*/
3095		public Set<Map.Entry<K,V>> entrySet() {
3096	<	Set<Map.Entry<K,V>> es = entrySet;
3097	<	return (es != null) ? es : (entrySet = new EntrySet());
3096	>	EntrySetView<K,V> es = entrySet;
3097	>	return (es != null) ? es : (entrySet = new EntrySetView<K,V>(this));
3098		}
3099
3100		/**
#	Line 1311 | Line 3104 | public class ConcurrentHashMapV8<K, V>
3104		* @see #keySet()
3105		*/
3106		public Enumeration<K> keys() {
3107	<	return new KeyIterator();
3107	>	return new KeyIterator<K,V>(this);
3108		}
3109
3110		/**
#	Line 1321 | Line 3114 | public class ConcurrentHashMapV8<K, V>
3114		* @see #values()
3115		*/
3116		public Enumeration<V> elements() {
3117	<	return new ValueIterator();
3117	>	return new ValueIterator<K,V>(this);
3118	>	}
3119	>
3120	>	/**
3121	>	* Returns a partitionable iterator of the keys in this map.
3122	>	*
3123	>	* @return a partitionable iterator of the keys in this map
3124	>	*/
3125	>	public Spliterator<K> keySpliterator() {
3126	>	return new KeyIterator<K,V>(this);
3127	>	}
3128	>
3129	>	/**
3130	>	* Returns a partitionable iterator of the values in this map.
3131	>	*
3132	>	* @return a partitionable iterator of the values in this map
3133	>	*/
3134	>	public Spliterator<V> valueSpliterator() {
3135	>	return new ValueIterator<K,V>(this);
3136	>	}
3137	>
3138	>	/**
3139	>	* Returns a partitionable iterator of the entries in this map.
3140	>	*
3141	>	* @return a partitionable iterator of the entries in this map
3142	>	*/
3143	>	public Spliterator<Map.Entry<K,V>> entrySpliterator() {
3144	>	return new EntryIterator<K,V>(this);
3145		}
3146
3147		/**
#	Line 1332 | Line 3152 | public class ConcurrentHashMapV8<K, V>
3152		* @return the hash code value for this map
3153		*/
3154		public int hashCode() {
3155	<	return new HashIterator().mapHashCode();
3155	>	int h = 0;
3156	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
3157	>	Object v;
3158	>	while ((v = it.advance()) != null) {
3159	>	h += it.nextKey.hashCode() ^ v.hashCode();
3160	>	}
3161	>	return h;
3162		}
3163
3164		/**
#	Line 1347 | Line 3173 | public class ConcurrentHashMapV8<K, V>
3173		* @return a string representation of this map
3174		*/
3175		public String toString() {
3176	<	return new HashIterator().mapToString();
3176	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
3177	>	StringBuilder sb = new StringBuilder();
3178	>	sb.append('{');
3179	>	Object v;
3180	>	if ((v = it.advance()) != null) {
3181	>	for (;;) {
3182	>	Object k = it.nextKey;
3183	>	sb.append(k == this ? "(this Map)" : k);
3184	>	sb.append('=');
3185	>	sb.append(v == this ? "(this Map)" : v);
3186	>	if ((v = it.advance()) == null)
3187	>	break;
3188	>	sb.append(',').append(' ');
3189	>	}
3190	>	}
3191	>	return sb.append('}').toString();
3192		}
3193
3194		/**
#	Line 1361 | Line 3202 | public class ConcurrentHashMapV8<K, V>
3202		* @return {@code true} if the specified object is equal to this map
3203		*/
3204		public boolean equals(Object o) {
3205	<	if (o == this)
3206	<	return true;
3207	<	if (!(o instanceof Map))
3208	<	return false;
3209	<	Map<?,?> m = (Map<?,?>) o;
3210	<	try {
3211	<	for (Map.Entry<K,V> e : this.entrySet())
3212	<	if (! e.getValue().equals(m.get(e.getKey())))
3205	>	if (o != this) {
3206	>	if (!(o instanceof Map))
3207	>	return false;
3208	>	Map<?,?> m = (Map<?,?>) o;
3209	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
3210	>	Object val;
3211	>	while ((val = it.advance()) != null) {
3212	>	Object v = m.get(it.nextKey);
3213	>	if (v == null || (v != val && !v.equals(val)))
3214		return false;
3215	+	}
3216		for (Map.Entry<?,?> e : m.entrySet()) {
3217	<	Object k = e.getKey();
3218	<	Object v = e.getValue();
3219	<	if (k == null || v == null || !v.equals(get(k)))
3217	>	Object mk, mv, v;
3218	>	if ((mk = e.getKey()) == null ||
3219	>	(mv = e.getValue()) == null ||
3220	>	(v = internalGet(mk)) == null ||
3221	>	(mv != v && !mv.equals(v)))
3222		return false;
3223		}
1379	–	return true;
1380	–	} catch (ClassCastException unused) {
1381	–	return false;
1382	–	} catch (NullPointerException unused) {
1383	–	return false;
3224		}
3225	+	return true;
3226		}
3227
3228	<	/**
1388	<	* Custom Entry class used by EntryIterator.next(), that relays
1389	<	* setValue changes to the underlying map.
1390	<	*/
1391	<	final class WriteThroughEntry extends AbstractMap.SimpleEntry<K,V> {
1392	<	@SuppressWarnings("unchecked")
1393	<	WriteThroughEntry(Object k, Object v) {
1394	<	super((K)k, (V)v);
1395	<	}
3228	>	/* ----------------Iterators -------------- */
3229
3230	<	/**
3231	<	* Sets our entry's value and writes through to the map. The
3232	<	* value to return is somewhat arbitrary here. Since a
3233	<	* WriteThroughEntry does not necessarily track asynchronous
3234	<	* changes, the most recent "previous" value could be
3235	<	* different from what we return (or could even have been
3236	<	* removed in which case the put will re-establish). We do not
3237	<	* and cannot guarantee more.
3238	<	*/
3239	<	public V setValue(V value) {
3240	<	if (value == null) throw new NullPointerException();
3241	<	V v = super.setValue(value);
3242	<	ConcurrentHashMapV8.this.put(getKey(), value);
3243	<	return v;
3230	>	@SuppressWarnings("serial") static final class KeyIterator<K,V> extends Traverser<K,V,Object>
3231	>	implements Spliterator<K>, Enumeration<K> {
3232	>	KeyIterator(ConcurrentHashMapV8<K, V> map) { super(map); }
3233	>	KeyIterator(ConcurrentHashMapV8<K, V> map, Traverser<K,V,Object> it) {
3234	>	super(map, it, -1);
3235	>	}
3236	>	public KeyIterator<K,V> split() {
3237	>	if (nextKey != null)
3238	>	throw new IllegalStateException();
3239	>	return new KeyIterator<K,V>(map, this);
3240	>	}
3241	>	@SuppressWarnings("unchecked") public final K next() {
3242	>	if (nextVal == null && advance() == null)
3243	>	throw new NoSuchElementException();
3244	>	Object k = nextKey;
3245	>	nextVal = null;
3246	>	return (K) k;
3247		}
1412	–	}
1413	–
1414	–	final class KeyIterator extends HashIterator
1415	–	implements Iterator<K>, Enumeration<K> {
1416	–	@SuppressWarnings("unchecked")
1417	–	public final K next()        { return (K)super.nextKey(); }
1418	–	@SuppressWarnings("unchecked")
1419	–	public final K nextElement() { return (K)super.nextKey(); }
1420	–	}
1421	–
1422	–	final class ValueIterator extends HashIterator
1423	–	implements Iterator<V>, Enumeration<V> {
1424	–	@SuppressWarnings("unchecked")
1425	–	public final V next()        { return (V)super.nextValue(); }
1426	–	@SuppressWarnings("unchecked")
1427	–	public final V nextElement() { return (V)super.nextValue(); }
1428	–	}
3248
3249	<	final class EntryIterator extends HashIterator
1431	<	implements Iterator<Entry<K,V>> {
1432	<	public final Map.Entry<K,V> next() { return super.nextEntry(); }
3249	>	public final K nextElement() { return next(); }
3250		}
3251
3252	<	final class KeySet extends AbstractSet<K> {
3253	<	public int size() {
3254	<	return ConcurrentHashMapV8.this.size();
3255	<	}
3256	<	public boolean isEmpty() {
1440	<	return ConcurrentHashMapV8.this.isEmpty();
1441	<	}
1442	<	public void clear() {
1443	<	ConcurrentHashMapV8.this.clear();
1444	<	}
1445	<	public Iterator<K> iterator() {
1446	<	return new KeyIterator();
3252	>	@SuppressWarnings("serial") static final class ValueIterator<K,V> extends Traverser<K,V,Object>
3253	>	implements Spliterator<V>, Enumeration<V> {
3254	>	ValueIterator(ConcurrentHashMapV8<K, V> map) { super(map); }
3255	>	ValueIterator(ConcurrentHashMapV8<K, V> map, Traverser<K,V,Object> it) {
3256	>	super(map, it, -1);
3257		}
3258	<	public boolean contains(Object o) {
3259	<	return ConcurrentHashMapV8.this.containsKey(o);
3258	>	public ValueIterator<K,V> split() {
3259	>	if (nextKey != null)
3260	>	throw new IllegalStateException();
3261	>	return new ValueIterator<K,V>(map, this);
3262		}
3263	<	public boolean remove(Object o) {
3264	<	return ConcurrentHashMapV8.this.remove(o) != null;
3263	>
3264	>	@SuppressWarnings("unchecked") public final V next() {
3265	>	Object v;
3266	>	if ((v = nextVal) == null && (v = advance()) == null)
3267	>	throw new NoSuchElementException();
3268	>	nextVal = null;
3269	>	return (V) v;
3270		}
3271	+
3272	+	public final V nextElement() { return next(); }
3273		}
3274
3275	<	final class Values extends AbstractCollection<V> {
3276	<	public int size() {
3277	<	return ConcurrentHashMapV8.this.size();
3278	<	}
3279	<	public boolean isEmpty() {
1461	<	return ConcurrentHashMapV8.this.isEmpty();
1462	<	}
1463	<	public void clear() {
1464	<	ConcurrentHashMapV8.this.clear();
3275	>	@SuppressWarnings("serial") static final class EntryIterator<K,V> extends Traverser<K,V,Object>
3276	>	implements Spliterator<Map.Entry<K,V>> {
3277	>	EntryIterator(ConcurrentHashMapV8<K, V> map) { super(map); }
3278	>	EntryIterator(ConcurrentHashMapV8<K, V> map, Traverser<K,V,Object> it) {
3279	>	super(map, it, -1);
3280		}
3281	<	public Iterator<V> iterator() {
3282	<	return new ValueIterator();
3281	>	public EntryIterator<K,V> split() {
3282	>	if (nextKey != null)
3283	>	throw new IllegalStateException();
3284	>	return new EntryIterator<K,V>(map, this);
3285		}
3286	<	public boolean contains(Object o) {
3287	<	return ConcurrentHashMapV8.this.containsValue(o);
3286	>
3287	>	@SuppressWarnings("unchecked") public final Map.Entry<K,V> next() {
3288	>	Object v;
3289	>	if ((v = nextVal) == null && (v = advance()) == null)
3290	>	throw new NoSuchElementException();
3291	>	Object k = nextKey;
3292	>	nextVal = null;
3293	>	return new MapEntry<K,V>((K)k, (V)v, map);
3294		}
3295		}
3296
3297	<	final class EntrySet extends AbstractSet<Map.Entry<K,V>> {
3298	<	public int size() {
3299	<	return ConcurrentHashMapV8.this.size();
3300	<	}
3301	<	public boolean isEmpty() {
3302	<	return ConcurrentHashMapV8.this.isEmpty();
3303	<	}
3304	<	public void clear() {
3305	<	ConcurrentHashMapV8.this.clear();
3306	<	}
3307	<	public Iterator<Map.Entry<K,V>> iterator() {
1485	<	return new EntryIterator();
3297	>	/**
3298	>	* Exported Entry for iterators
3299	>	*/
3300	>	static final class MapEntry<K,V> implements Map.Entry<K, V> {
3301	>	final K key; // non-null
3302	>	V val;       // non-null
3303	>	final ConcurrentHashMapV8<K, V> map;
3304	>	MapEntry(K key, V val, ConcurrentHashMapV8<K, V> map) {
3305	>	this.key = key;
3306	>	this.val = val;
3307	>	this.map = map;
3308		}
3309	<	public boolean contains(Object o) {
3310	<	if (!(o instanceof Map.Entry))
3311	<	return false;
3312	<	Map.Entry<?,?> e = (Map.Entry<?,?>)o;
3313	<	V v = ConcurrentHashMapV8.this.get(e.getKey());
3314	<	return v != null && v.equals(e.getValue());
3309	>	public final K getKey()       { return key; }
3310	>	public final V getValue()     { return val; }
3311	>	public final int hashCode()   { return key.hashCode() ^ val.hashCode(); }
3312	>	public final String toString(){ return key + "=" + val; }
3313	>
3314	>	public final boolean equals(Object o) {
3315	>	Object k, v; Map.Entry<?,?> e;
3316	>	return ((o instanceof Map.Entry) &&
3317	>	(k = (e = (Map.Entry<?,?>)o).getKey()) != null &&
3318	>	(v = e.getValue()) != null &&
3319	>	(k == key || k.equals(key)) &&
3320	>	(v == val || v.equals(val)));
3321		}
3322	<	public boolean remove(Object o) {
3323	<	if (!(o instanceof Map.Entry))
3324	<	return false;
3325	<	Map.Entry<?,?> e = (Map.Entry<?,?>)o;
3326	<	return ConcurrentHashMapV8.this.remove(e.getKey(), e.getValue());
3322	>
3323	>	/**
3324	>	* Sets our entry's value and writes through to the map. The
3325	>	* value to return is somewhat arbitrary here. Since we do not
3326	>	* necessarily track asynchronous changes, the most recent
3327	>	* "previous" value could be different from what we return (or
3328	>	* could even have been removed in which case the put will
3329	>	* re-establish). We do not and cannot guarantee more.
3330	>	*/
3331	>	public final V setValue(V value) {
3332	>	if (value == null) throw new NullPointerException();
3333	>	V v = val;
3334	>	val = value;
3335	>	map.put(key, value);
3336	>	return v;
3337		}
3338		}
3339
3340	+	/**
3341	+	* Returns exportable snapshot entry for the given key and value
3342	+	* when write-through can't or shouldn't be used.
3343	+	*/
3344	+	static <K,V> AbstractMap.SimpleEntry<K,V> entryFor(K k, V v) {
3345	+	return new AbstractMap.SimpleEntry<K,V>(k, v);
3346	+	}
3347	+
3348		/* ---------------- Serialization Support -------------- */
3349
3350		/**
3351	<	* Helper class used in previous version, declared for the sake of
3352	<	* serialization compatibility
3351	>	* Stripped-down version of helper class used in previous version,
3352	>	* declared for the sake of serialization compatibility
3353		*/
3354	<	static class Segment<K,V> extends java.util.concurrent.locks.ReentrantLock
1509	<	implements Serializable {
3354	>	static class Segment<K,V> implements Serializable {
3355		private static final long serialVersionUID = 2249069246763182397L;
3356		final float loadFactor;
3357		Segment(float lf) { this.loadFactor = lf; }
#	Line 1521 | Line 3366 | public class ConcurrentHashMapV8<K, V>
3366		* for each key-value mapping, followed by a null pair.
3367		* The key-value mappings are emitted in no particular order.
3368		*/
3369	<	@SuppressWarnings("unchecked")
3370	<	private void writeObject(java.io.ObjectOutputStream s)
1526	<	throws java.io.IOException {
3369	>	@SuppressWarnings("unchecked") private void writeObject(java.io.ObjectOutputStream s)
3370	>	throws java.io.IOException {
3371		if (segments == null) { // for serialization compatibility
3372		segments = (Segment<K,V>[])
3373		new Segment<?,?>[DEFAULT_CONCURRENCY_LEVEL];
3374		for (int i = 0; i < segments.length; ++i)
3375	<	segments[i] = new Segment<K,V>(loadFactor);
3375	>	segments[i] = new Segment<K,V>(LOAD_FACTOR);
3376		}
3377		s.defaultWriteObject();
3378	<	new HashIterator().writeEntries(s);
3378	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
3379	>	Object v;
3380	>	while ((v = it.advance()) != null) {
3381	>	s.writeObject(it.nextKey);
3382	>	s.writeObject(v);
3383	>	}
3384		s.writeObject(null);
3385		s.writeObject(null);
3386		segments = null; // throw away
#	Line 1541 | Line 3390 | public class ConcurrentHashMapV8<K, V>
3390		* Reconstitutes the instance from a stream (that is, deserializes it).
3391		* @param s the stream
3392		*/
3393	<	@SuppressWarnings("unchecked")
3394	<	private void readObject(java.io.ObjectInputStream s)
1546	<	throws java.io.IOException, ClassNotFoundException {
3393	>	@SuppressWarnings("unchecked") private void readObject(java.io.ObjectInputStream s)
3394	>	throws java.io.IOException, ClassNotFoundException {
3395		s.defaultReadObject();
1548	–	// find load factor in a segment, if one exists
1549	–	if (segments != null && segments.length != 0)
1550	–	this.loadFactor = segments[0].loadFactor;
1551	–	else
1552	–	this.loadFactor = DEFAULT_LOAD_FACTOR;
1553	–	this.initCap = DEFAULT_CAPACITY;
1554	–	LongAdder ct = new LongAdder(); // force final field write
1555	–	UNSAFE.putObjectVolatile(this, counterOffset, ct);
3396		this.segments = null; // unneeded
3397	+	// initialize transient final field
3398	+	UNSAFE.putObjectVolatile(this, counterOffset, new LongAdder());
3399
3400	<	// Read the keys and values, and put the mappings in the table
3400	>	// Create all nodes, then place in table once size is known
3401	>	long size = 0L;
3402	>	Node p = null;
3403		for (;;) {
3404	<	K key = (K) s.readObject();
3405	<	V value = (V) s.readObject();
3406	<	if (key == null)
3404	>	K k = (K) s.readObject();
3405	>	V v = (V) s.readObject();
3406	>	if (k != null && v != null) {
3407	>	int h = spread(k.hashCode());
3408	>	p = new Node(h, k, v, p);
3409	>	++size;
3410	>	}
3411	>	else
3412		break;
3413	<	put(key, value);
3413	>	}
3414	>	if (p != null) {
3415	>	boolean init = false;
3416	>	int n;
3417	>	if (size >= (long)(MAXIMUM_CAPACITY >>> 1))
3418	>	n = MAXIMUM_CAPACITY;
3419	>	else {
3420	>	int sz = (int)size;
3421	>	n = tableSizeFor(sz + (sz >>> 1) + 1);
3422	>	}
3423	>	int sc = sizeCtl;
3424	>	boolean collide = false;
3425	>	if (n > sc &&
3426	>	UNSAFE.compareAndSwapInt(this, sizeCtlOffset, sc, -1)) {
3427	>	try {
3428	>	if (table == null) {
3429	>	init = true;
3430	>	Node[] tab = new Node[n];
3431	>	int mask = n - 1;
3432	>	while (p != null) {
3433	>	int j = p.hash & mask;
3434	>	Node next = p.next;
3435	>	Node q = p.next = tabAt(tab, j);
3436	>	setTabAt(tab, j, p);
3437	>	if (!collide && q != null && q.hash == p.hash)
3438	>	collide = true;
3439	>	p = next;
3440	>	}
3441	>	table = tab;
3442	>	counter.add(size);
3443	>	sc = n - (n >>> 2);
3444	>	}
3445	>	} finally {
3446	>	sizeCtl = sc;
3447	>	}
3448	>	if (collide) { // rescan and convert to TreeBins
3449	>	Node[] tab = table;
3450	>	for (int i = 0; i < tab.length; ++i) {
3451	>	int c = 0;
3452	>	for (Node e = tabAt(tab, i); e != null; e = e.next) {
3453	>	if (++c > TREE_THRESHOLD &&
3454	>	(e.key instanceof Comparable)) {
3455	>	replaceWithTreeBin(tab, i, e.key);
3456	>	break;
3457	>	}
3458	>	}
3459	>	}
3460	>	}
3461	>	}
3462	>	if (!init) { // Can only happen if unsafely published.
3463	>	while (p != null) {
3464	>	internalPut(p.key, p.val);
3465	>	p = p.next;
3466	>	}
3467	>	}
3468	>	}
3469	>	}
3470	>
3471	>
3472	>	// -------------------------------------------------------
3473	>
3474	>	// Sams
3475	>	/** Interface describing a void action of one argument */
3476	>	public interface Action<A> { void apply(A a); }
3477	>	/** Interface describing a void action of two arguments */
3478	>	public interface BiAction<A,B> { void apply(A a, B b); }
3479	>	/** Interface describing a function of one argument */
3480	>	public interface Fun<A,T> { T apply(A a); }
3481	>	/** Interface describing a function of two arguments */
3482	>	public interface BiFun<A,B,T> { T apply(A a, B b); }
3483	>	/** Interface describing a function of no arguments */
3484	>	public interface Generator<T> { T apply(); }
3485	>	/** Interface describing a function mapping its argument to a double */
3486	>	public interface ObjectToDouble<A> { double apply(A a); }
3487	>	/** Interface describing a function mapping its argument to a long */
3488	>	public interface ObjectToLong<A> { long apply(A a); }
3489	>	/** Interface describing a function mapping its argument to an int */
3490	>	public interface ObjectToInt<A> {int apply(A a); }
3491	>	/** Interface describing a function mapping two arguments to a double */
3492	>	public interface ObjectByObjectToDouble<A,B> { double apply(A a, B b); }
3493	>	/** Interface describing a function mapping two arguments to a long */
3494	>	public interface ObjectByObjectToLong<A,B> { long apply(A a, B b); }
3495	>	/** Interface describing a function mapping two arguments to an int */
3496	>	public interface ObjectByObjectToInt<A,B> {int apply(A a, B b); }
3497	>	/** Interface describing a function mapping a double to a double */
3498	>	public interface DoubleToDouble { double apply(double a); }
3499	>	/** Interface describing a function mapping a long to a long */
3500	>	public interface LongToLong { long apply(long a); }
3501	>	/** Interface describing a function mapping an int to an int */
3502	>	public interface IntToInt { int apply(int a); }
3503	>	/** Interface describing a function mapping two doubles to a double */
3504	>	public interface DoubleByDoubleToDouble { double apply(double a, double b); }
3505	>	/** Interface describing a function mapping two longs to a long */
3506	>	public interface LongByLongToLong { long apply(long a, long b); }
3507	>	/** Interface describing a function mapping two ints to an int */
3508	>	public interface IntByIntToInt { int apply(int a, int b); }
3509	>
3510	>
3511	>	// -------------------------------------------------------
3512	>
3513	>	/**
3514	>	* Performs the given action for each (key, value).
3515	>	*
3516	>	* @param action the action
3517	>	*/
3518	>	public void forEach(BiAction<K,V> action) {
3519	>	ForkJoinTasks.forEach
3520	>	(this, action).invoke();
3521	>	}
3522	>
3523	>	/**
3524	>	* Performs the given action for each non-null transformation
3525	>	* of each (key, value).
3526	>	*
3527	>	* @param transformer a function returning the transformation
3528	>	* for an element, or null of there is no transformation (in
3529	>	* which case the action is not applied).
3530	>	* @param action the action
3531	>	*/
3532	>	public <U> void forEach(BiFun<? super K, ? super V, ? extends U> transformer,
3533	>	Action<U> action) {
3534	>	ForkJoinTasks.forEach
3535	>	(this, transformer, action).invoke();
3536	>	}
3537	>
3538	>	/**
3539	>	* Returns a non-null result from applying the given search
3540	>	* function on each (key, value), or null if none.  Upon
3541	>	* success, further element processing is suppressed and the
3542	>	* results of any other parallel invocations of the search
3543	>	* function are ignored.
3544	>	*
3545	>	* @param searchFunction a function returning a non-null
3546	>	* result on success, else null
3547	>	* @return a non-null result from applying the given search
3548	>	* function on each (key, value), or null if none
3549	>	*/
3550	>	public <U> U search(BiFun<? super K, ? super V, ? extends U> searchFunction) {
3551	>	return ForkJoinTasks.search
3552	>	(this, searchFunction).invoke();
3553	>	}
3554	>
3555	>	/**
3556	>	* Returns the result of accumulating the given transformation
3557	>	* of all (key, value) pairs using the given reducer to
3558	>	* combine values, or null if none.
3559	>	*
3560	>	* @param transformer a function returning the transformation
3561	>	* for an element, or null of there is no transformation (in
3562	>	* which case it is not combined).
3563	>	* @param reducer a commutative associative combining function
3564	>	* @return the result of accumulating the given transformation
3565	>	* of all (key, value) pairs
3566	>	*/
3567	>	public <U> U reduce(BiFun<? super K, ? super V, ? extends U> transformer,
3568	>	BiFun<? super U, ? super U, ? extends U> reducer) {
3569	>	return ForkJoinTasks.reduce
3570	>	(this, transformer, reducer).invoke();
3571	>	}
3572	>
3573	>	/**
3574	>	* Returns the result of accumulating the given transformation
3575	>	* of all (key, value) pairs using the given reducer to
3576	>	* combine values, and the given basis as an identity value.
3577	>	*
3578	>	* @param transformer a function returning the transformation
3579	>	* for an element
3580	>	* @param basis the identity (initial default value) for the reduction
3581	>	* @param reducer a commutative associative combining function
3582	>	* @return the result of accumulating the given transformation
3583	>	* of all (key, value) pairs
3584	>	*/
3585	>	public double reduceToDouble(ObjectByObjectToDouble<? super K, ? super V> transformer,
3586	>	double basis,
3587	>	DoubleByDoubleToDouble reducer) {
3588	>	return ForkJoinTasks.reduceToDouble
3589	>	(this, transformer, basis, reducer).invoke();
3590	>	}
3591	>
3592	>	/**
3593	>	* Returns the result of accumulating the given transformation
3594	>	* of all (key, value) pairs using the given reducer to
3595	>	* combine values, and the given basis as an identity value.
3596	>	*
3597	>	* @param transformer a function returning the transformation
3598	>	* for an element
3599	>	* @param basis the identity (initial default value) for the reduction
3600	>	* @param reducer a commutative associative combining function
3601	>	* @return the result of accumulating the given transformation
3602	>	* of all (key, value) pairs
3603	>	*/
3604	>	public long reduceToLong(ObjectByObjectToLong<? super K, ? super V> transformer,
3605	>	long basis,
3606	>	LongByLongToLong reducer) {
3607	>	return ForkJoinTasks.reduceToLong
3608	>	(this, transformer, basis, reducer).invoke();
3609	>	}
3610	>
3611	>	/**
3612	>	* Returns the result of accumulating the given transformation
3613	>	* of all (key, value) pairs using the given reducer to
3614	>	* combine values, and the given basis as an identity value.
3615	>	*
3616	>	* @param transformer a function returning the transformation
3617	>	* for an element
3618	>	* @param basis the identity (initial default value) for the reduction
3619	>	* @param reducer a commutative associative combining function
3620	>	* @return the result of accumulating the given transformation
3621	>	* of all (key, value) pairs
3622	>	*/
3623	>	public int reduceToInt(ObjectByObjectToInt<? super K, ? super V> transformer,
3624	>	int basis,
3625	>	IntByIntToInt reducer) {
3626	>	return ForkJoinTasks.reduceToInt
3627	>	(this, transformer, basis, reducer).invoke();
3628	>	}
3629	>
3630	>	/**
3631	>	* Performs the given action for each key.
3632	>	*
3633	>	* @param action the action
3634	>	*/
3635	>	public void forEachKey(Action<K> action) {
3636	>	ForkJoinTasks.forEachKey
3637	>	(this, action).invoke();
3638	>	}
3639	>
3640	>	/**
3641	>	* Performs the given action for each non-null transformation
3642	>	* of each key.
3643	>	*
3644	>	* @param transformer a function returning the transformation
3645	>	* for an element, or null of there is no transformation (in
3646	>	* which case the action is not applied).
3647	>	* @param action the action
3648	>	*/
3649	>	public <U> void forEachKey(Fun<? super K, ? extends U> transformer,
3650	>	Action<U> action) {
3651	>	ForkJoinTasks.forEachKey
3652	>	(this, transformer, action).invoke();
3653	>	}
3654	>
3655	>	/**
3656	>	* Returns a non-null result from applying the given search
3657	>	* function on each key, or null if none. Upon success,
3658	>	* further element processing is suppressed and the results of
3659	>	* any other parallel invocations of the search function are
3660	>	* ignored.
3661	>	*
3662	>	* @param searchFunction a function returning a non-null
3663	>	* result on success, else null
3664	>	* @return a non-null result from applying the given search
3665	>	* function on each key, or null if none
3666	>	*/
3667	>	public <U> U searchKeys(Fun<? super K, ? extends U> searchFunction) {
3668	>	return ForkJoinTasks.searchKeys
3669	>	(this, searchFunction).invoke();
3670	>	}
3671	>
3672	>	/**
3673	>	* Returns the result of accumulating all keys using the given
3674	>	* reducer to combine values, or null if none.
3675	>	*
3676	>	* @param reducer a commutative associative combining function
3677	>	* @return the result of accumulating all keys using the given
3678	>	* reducer to combine values, or null if none
3679	>	*/
3680	>	public K reduceKeys(BiFun<? super K, ? super K, ? extends K> reducer) {
3681	>	return ForkJoinTasks.reduceKeys
3682	>	(this, reducer).invoke();
3683	>	}
3684	>
3685	>	/**
3686	>	* Returns the result of accumulating the given transformation
3687	>	* of all keys using the given reducer to combine values, or
3688	>	* null if none.
3689	>	*
3690	>	* @param transformer a function returning the transformation
3691	>	* for an element, or null of there is no transformation (in
3692	>	* which case it is not combined).
3693	>	* @param reducer a commutative associative combining function
3694	>	* @return the result of accumulating the given transformation
3695	>	* of all keys
3696	>	*/
3697	>	public <U> U reduceKeys(Fun<? super K, ? extends U> transformer,
3698	>	BiFun<? super U, ? super U, ? extends U> reducer) {
3699	>	return ForkJoinTasks.reduceKeys
3700	>	(this, transformer, reducer).invoke();
3701	>	}
3702	>
3703	>	/**
3704	>	* Returns the result of accumulating the given transformation
3705	>	* of all keys using the given reducer to combine values, and
3706	>	* the given basis as an identity value.
3707	>	*
3708	>	* @param transformer a function returning the transformation
3709	>	* for an element
3710	>	* @param basis the identity (initial default value) for the reduction
3711	>	* @param reducer a commutative associative combining function
3712	>	* @return  the result of accumulating the given transformation
3713	>	* of all keys
3714	>	*/
3715	>	public double reduceKeysToDouble(ObjectToDouble<? super K> transformer,
3716	>	double basis,
3717	>	DoubleByDoubleToDouble reducer) {
3718	>	return ForkJoinTasks.reduceKeysToDouble
3719	>	(this, transformer, basis, reducer).invoke();
3720	>	}
3721	>
3722	>	/**
3723	>	* Returns the result of accumulating the given transformation
3724	>	* of all keys using the given reducer to combine values, and
3725	>	* the given basis as an identity value.
3726	>	*
3727	>	* @param transformer a function returning the transformation
3728	>	* for an element
3729	>	* @param basis the identity (initial default value) for the reduction
3730	>	* @param reducer a commutative associative combining function
3731	>	* @return the result of accumulating the given transformation
3732	>	* of all keys
3733	>	*/
3734	>	public long reduceKeysToLong(ObjectToLong<? super K> transformer,
3735	>	long basis,
3736	>	LongByLongToLong reducer) {
3737	>	return ForkJoinTasks.reduceKeysToLong
3738	>	(this, transformer, basis, reducer).invoke();
3739	>	}
3740	>
3741	>	/**
3742	>	* Returns the result of accumulating the given transformation
3743	>	* of all keys using the given reducer to combine values, and
3744	>	* the given basis as an identity value.
3745	>	*
3746	>	* @param transformer a function returning the transformation
3747	>	* for an element
3748	>	* @param basis the identity (initial default value) for the reduction
3749	>	* @param reducer a commutative associative combining function
3750	>	* @return the result of accumulating the given transformation
3751	>	* of all keys
3752	>	*/
3753	>	public int reduceKeysToInt(ObjectToInt<? super K> transformer,
3754	>	int basis,
3755	>	IntByIntToInt reducer) {
3756	>	return ForkJoinTasks.reduceKeysToInt
3757	>	(this, transformer, basis, reducer).invoke();
3758	>	}
3759	>
3760	>	/**
3761	>	* Performs the given action for each value.
3762	>	*
3763	>	* @param action the action
3764	>	*/
3765	>	public void forEachValue(Action<V> action) {
3766	>	ForkJoinTasks.forEachValue
3767	>	(this, action).invoke();
3768	>	}
3769	>
3770	>	/**
3771	>	* Performs the given action for each non-null transformation
3772	>	* of each value.
3773	>	*
3774	>	* @param transformer a function returning the transformation
3775	>	* for an element, or null of there is no transformation (in
3776	>	* which case the action is not applied).
3777	>	*/
3778	>	public <U> void forEachValue(Fun<? super V, ? extends U> transformer,
3779	>	Action<U> action) {
3780	>	ForkJoinTasks.forEachValue
3781	>	(this, transformer, action).invoke();
3782	>	}
3783	>
3784	>	/**
3785	>	* Returns a non-null result from applying the given search
3786	>	* function on each value, or null if none.  Upon success,
3787	>	* further element processing is suppressed and the results of
3788	>	* any other parallel invocations of the search function are
3789	>	* ignored.
3790	>	*
3791	>	* @param searchFunction a function returning a non-null
3792	>	* result on success, else null
3793	>	* @return a non-null result from applying the given search
3794	>	* function on each value, or null if none
3795	>	*
3796	>	*/
3797	>	public <U> U searchValues(Fun<? super V, ? extends U> searchFunction) {
3798	>	return ForkJoinTasks.searchValues
3799	>	(this, searchFunction).invoke();
3800	>	}
3801	>
3802	>	/**
3803	>	* Returns the result of accumulating all values using the
3804	>	* given reducer to combine values, or null if none.
3805	>	*
3806	>	* @param reducer a commutative associative combining function
3807	>	* @return  the result of accumulating all values
3808	>	*/
3809	>	public V reduceValues(BiFun<? super V, ? super V, ? extends V> reducer) {
3810	>	return ForkJoinTasks.reduceValues
3811	>	(this, reducer).invoke();
3812	>	}
3813	>
3814	>	/**
3815	>	* Returns the result of accumulating the given transformation
3816	>	* of all values using the given reducer to combine values, or
3817	>	* null if none.
3818	>	*
3819	>	* @param transformer a function returning the transformation
3820	>	* for an element, or null of there is no transformation (in
3821	>	* which case it is not combined).
3822	>	* @param reducer a commutative associative combining function
3823	>	* @return the result of accumulating the given transformation
3824	>	* of all values
3825	>	*/
3826	>	public <U> U reduceValues(Fun<? super V, ? extends U> transformer,
3827	>	BiFun<? super U, ? super U, ? extends U> reducer) {
3828	>	return ForkJoinTasks.reduceValues
3829	>	(this, transformer, reducer).invoke();
3830	>	}
3831	>
3832	>	/**
3833	>	* Returns the result of accumulating the given transformation
3834	>	* of all values using the given reducer to combine values,
3835	>	* and the given basis as an identity value.
3836	>	*
3837	>	* @param transformer a function returning the transformation
3838	>	* for an element
3839	>	* @param basis the identity (initial default value) for the reduction
3840	>	* @param reducer a commutative associative combining function
3841	>	* @return the result of accumulating the given transformation
3842	>	* of all values
3843	>	*/
3844	>	public double reduceValuesToDouble(ObjectToDouble<? super V> transformer,
3845	>	double basis,
3846	>	DoubleByDoubleToDouble reducer) {
3847	>	return ForkJoinTasks.reduceValuesToDouble
3848	>	(this, transformer, basis, reducer).invoke();
3849	>	}
3850	>
3851	>	/**
3852	>	* Returns the result of accumulating the given transformation
3853	>	* of all values using the given reducer to combine values,
3854	>	* and the given basis as an identity value.
3855	>	*
3856	>	* @param transformer a function returning the transformation
3857	>	* for an element
3858	>	* @param basis the identity (initial default value) for the reduction
3859	>	* @param reducer a commutative associative combining function
3860	>	* @return the result of accumulating the given transformation
3861	>	* of all values
3862	>	*/
3863	>	public long reduceValuesToLong(ObjectToLong<? super V> transformer,
3864	>	long basis,
3865	>	LongByLongToLong reducer) {
3866	>	return ForkJoinTasks.reduceValuesToLong
3867	>	(this, transformer, basis, reducer).invoke();
3868	>	}
3869	>
3870	>	/**
3871	>	* Returns the result of accumulating the given transformation
3872	>	* of all values using the given reducer to combine values,
3873	>	* and the given basis as an identity value.
3874	>	*
3875	>	* @param transformer a function returning the transformation
3876	>	* for an element
3877	>	* @param basis the identity (initial default value) for the reduction
3878	>	* @param reducer a commutative associative combining function
3879	>	* @return the result of accumulating the given transformation
3880	>	* of all values
3881	>	*/
3882	>	public int reduceValuesToInt(ObjectToInt<? super V> transformer,
3883	>	int basis,
3884	>	IntByIntToInt reducer) {
3885	>	return ForkJoinTasks.reduceValuesToInt
3886	>	(this, transformer, basis, reducer).invoke();
3887	>	}
3888	>
3889	>	/**
3890	>	* Performs the given action for each entry.
3891	>	*
3892	>	* @param action the action
3893	>	*/
3894	>	public void forEachEntry(Action<Map.Entry<K,V>> action) {
3895	>	ForkJoinTasks.forEachEntry
3896	>	(this, action).invoke();
3897	>	}
3898	>
3899	>	/**
3900	>	* Performs the given action for each non-null transformation
3901	>	* of each entry.
3902	>	*
3903	>	* @param transformer a function returning the transformation
3904	>	* for an element, or null of there is no transformation (in
3905	>	* which case the action is not applied).
3906	>	* @param action the action
3907	>	*/
3908	>	public <U> void forEachEntry(Fun<Map.Entry<K,V>, ? extends U> transformer,
3909	>	Action<U> action) {
3910	>	ForkJoinTasks.forEachEntry
3911	>	(this, transformer, action).invoke();
3912	>	}
3913	>
3914	>	/**
3915	>	* Returns a non-null result from applying the given search
3916	>	* function on each entry, or null if none.  Upon success,
3917	>	* further element processing is suppressed and the results of
3918	>	* any other parallel invocations of the search function are
3919	>	* ignored.
3920	>	*
3921	>	* @param searchFunction a function returning a non-null
3922	>	* result on success, else null
3923	>	* @return a non-null result from applying the given search
3924	>	* function on each entry, or null if none
3925	>	*/
3926	>	public <U> U searchEntries(Fun<Map.Entry<K,V>, ? extends U> searchFunction) {
3927	>	return ForkJoinTasks.searchEntries
3928	>	(this, searchFunction).invoke();
3929	>	}
3930	>
3931	>	/**
3932	>	* Returns the result of accumulating all entries using the
3933	>	* given reducer to combine values, or null if none.
3934	>	*
3935	>	* @param reducer a commutative associative combining function
3936	>	* @return the result of accumulating all entries
3937	>	*/
3938	>	public Map.Entry<K,V> reduceEntries(BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer) {
3939	>	return ForkJoinTasks.reduceEntries
3940	>	(this, reducer).invoke();
3941	>	}
3942	>
3943	>	/**
3944	>	* Returns the result of accumulating the given transformation
3945	>	* of all entries using the given reducer to combine values,
3946	>	* or null if none.
3947	>	*
3948	>	* @param transformer a function returning the transformation
3949	>	* for an element, or null of there is no transformation (in
3950	>	* which case it is not combined).
3951	>	* @param reducer a commutative associative combining function
3952	>	* @return the result of accumulating the given transformation
3953	>	* of all entries
3954	>	*/
3955	>	public <U> U reduceEntries(Fun<Map.Entry<K,V>, ? extends U> transformer,
3956	>	BiFun<? super U, ? super U, ? extends U> reducer) {
3957	>	return ForkJoinTasks.reduceEntries
3958	>	(this, transformer, reducer).invoke();
3959	>	}
3960	>
3961	>	/**
3962	>	* Returns the result of accumulating the given transformation
3963	>	* of all entries using the given reducer to combine values,
3964	>	* and the given basis as an identity value.
3965	>	*
3966	>	* @param transformer a function returning the transformation
3967	>	* for an element
3968	>	* @param basis the identity (initial default value) for the reduction
3969	>	* @param reducer a commutative associative combining function
3970	>	* @return the result of accumulating the given transformation
3971	>	* of all entries
3972	>	*/
3973	>	public double reduceEntriesToDouble(ObjectToDouble<Map.Entry<K,V>> transformer,
3974	>	double basis,
3975	>	DoubleByDoubleToDouble reducer) {
3976	>	return ForkJoinTasks.reduceEntriesToDouble
3977	>	(this, transformer, basis, reducer).invoke();
3978	>	}
3979	>
3980	>	/**
3981	>	* Returns the result of accumulating the given transformation
3982	>	* of all entries using the given reducer to combine values,
3983	>	* and the given basis as an identity value.
3984	>	*
3985	>	* @param transformer a function returning the transformation
3986	>	* for an element
3987	>	* @param basis the identity (initial default value) for the reduction
3988	>	* @param reducer a commutative associative combining function
3989	>	* @return  the result of accumulating the given transformation
3990	>	* of all entries
3991	>	*/
3992	>	public long reduceEntriesToLong(ObjectToLong<Map.Entry<K,V>> transformer,
3993	>	long basis,
3994	>	LongByLongToLong reducer) {
3995	>	return ForkJoinTasks.reduceEntriesToLong
3996	>	(this, transformer, basis, reducer).invoke();
3997	>	}
3998	>
3999	>	/**
4000	>	* Returns the result of accumulating the given transformation
4001	>	* of all entries using the given reducer to combine values,
4002	>	* and the given basis as an identity value.
4003	>	*
4004	>	* @param transformer a function returning the transformation
4005	>	* for an element
4006	>	* @param basis the identity (initial default value) for the reduction
4007	>	* @param reducer a commutative associative combining function
4008	>	* @return the result of accumulating the given transformation
4009	>	* of all entries
4010	>	*/
4011	>	public int reduceEntriesToInt(ObjectToInt<Map.Entry<K,V>> transformer,
4012	>	int basis,
4013	>	IntByIntToInt reducer) {
4014	>	return ForkJoinTasks.reduceEntriesToInt
4015	>	(this, transformer, basis, reducer).invoke();
4016	>	}
4017	>
4018	>	/* ----------------Views -------------- */
4019	>
4020	>	/**
4021	>	* Base class for views.
4022	>	*/
4023	>	static abstract class CHMView<K, V> {
4024	>	final ConcurrentHashMapV8<K, V> map;
4025	>	CHMView(ConcurrentHashMapV8<K, V> map)  { this.map = map; }
4026	>
4027	>	/**
4028	>	* Returns the map backing this view.
4029	>	*
4030	>	* @return the map backing this view
4031	>	*/
4032	>	public ConcurrentHashMapV8<K,V> getMap() { return map; }
4033	>
4034	>	public final int size()                 { return map.size(); }
4035	>	public final boolean isEmpty()          { return map.isEmpty(); }
4036	>	public final void clear()               { map.clear(); }
4037	>
4038	>	// implementations below rely on concrete classes supplying these
4039	>	abstract public Iterator<?> iterator();
4040	>	abstract public boolean contains(Object o);
4041	>	abstract public boolean remove(Object o);
4042	>
4043	>	private static final String oomeMsg = "Required array size too large";
4044	>
4045	>	public final Object[] toArray() {
4046	>	long sz = map.mappingCount();
4047	>	if (sz > (long)(MAX_ARRAY_SIZE))
4048	>	throw new OutOfMemoryError(oomeMsg);
4049	>	int n = (int)sz;
4050	>	Object[] r = new Object[n];
4051	>	int i = 0;
4052	>	Iterator<?> it = iterator();
4053	>	while (it.hasNext()) {
4054	>	if (i == n) {
4055	>	if (n >= MAX_ARRAY_SIZE)
4056	>	throw new OutOfMemoryError(oomeMsg);
4057	>	if (n >= MAX_ARRAY_SIZE - (MAX_ARRAY_SIZE >>> 1) - 1)
4058	>	n = MAX_ARRAY_SIZE;
4059	>	else
4060	>	n += (n >>> 1) + 1;
4061	>	r = Arrays.copyOf(r, n);
4062	>	}
4063	>	r[i++] = it.next();
4064	>	}
4065	>	return (i == n) ? r : Arrays.copyOf(r, i);
4066	>	}
4067	>
4068	>	@SuppressWarnings("unchecked") public final <T> T[] toArray(T[] a) {
4069	>	long sz = map.mappingCount();
4070	>	if (sz > (long)(MAX_ARRAY_SIZE))
4071	>	throw new OutOfMemoryError(oomeMsg);
4072	>	int m = (int)sz;
4073	>	T[] r = (a.length >= m) ? a :
4074	>	(T[])java.lang.reflect.Array
4075	>	.newInstance(a.getClass().getComponentType(), m);
4076	>	int n = r.length;
4077	>	int i = 0;
4078	>	Iterator<?> it = iterator();
4079	>	while (it.hasNext()) {
4080	>	if (i == n) {
4081	>	if (n >= MAX_ARRAY_SIZE)
4082	>	throw new OutOfMemoryError(oomeMsg);
4083	>	if (n >= MAX_ARRAY_SIZE - (MAX_ARRAY_SIZE >>> 1) - 1)
4084	>	n = MAX_ARRAY_SIZE;
4085	>	else
4086	>	n += (n >>> 1) + 1;
4087	>	r = Arrays.copyOf(r, n);
4088	>	}
4089	>	r[i++] = (T)it.next();
4090	>	}
4091	>	if (a == r && i < n) {
4092	>	r[i] = null; // null-terminate
4093	>	return r;
4094	>	}
4095	>	return (i == n) ? r : Arrays.copyOf(r, i);
4096	>	}
4097	>
4098	>	public final int hashCode() {
4099	>	int h = 0;
4100	>	for (Iterator<?> it = iterator(); it.hasNext();)
4101	>	h += it.next().hashCode();
4102	>	return h;
4103	>	}
4104	>
4105	>	public final String toString() {
4106	>	StringBuilder sb = new StringBuilder();
4107	>	sb.append('[');
4108	>	Iterator<?> it = iterator();
4109	>	if (it.hasNext()) {
4110	>	for (;;) {
4111	>	Object e = it.next();
4112	>	sb.append(e == this ? "(this Collection)" : e);
4113	>	if (!it.hasNext())
4114	>	break;
4115	>	sb.append(',').append(' ');
4116	>	}
4117	>	}
4118	>	return sb.append(']').toString();
4119	>	}
4120	>
4121	>	public final boolean containsAll(Collection<?> c) {
4122	>	if (c != this) {
4123	>	for (Iterator<?> it = c.iterator(); it.hasNext();) {
4124	>	Object e = it.next();
4125	>	if (e == null || !contains(e))
4126	>	return false;
4127	>	}
4128	>	}
4129	>	return true;
4130	>	}
4131	>
4132	>	public final boolean removeAll(Collection<?> c) {
4133	>	boolean modified = false;
4134	>	for (Iterator<?> it = iterator(); it.hasNext();) {
4135	>	if (c.contains(it.next())) {
4136	>	it.remove();
4137	>	modified = true;
4138	>	}
4139	>	}
4140	>	return modified;
4141	>	}
4142	>
4143	>	public final boolean retainAll(Collection<?> c) {
4144	>	boolean modified = false;
4145	>	for (Iterator<?> it = iterator(); it.hasNext();) {
4146	>	if (!c.contains(it.next())) {
4147	>	it.remove();
4148	>	modified = true;
4149	>	}
4150	>	}
4151	>	return modified;
4152	>	}
4153	>
4154	>	}
4155	>
4156	>	/**
4157	>	* A view of a ConcurrentHashMapV8 as a {@link Set} of keys, in
4158	>	* which additions may optionally be enabled by mapping to a
4159	>	* common value.  This class cannot be directly instantiated. See
4160	>	* {@link #keySet}, {@link #keySet(Object)}, {@link #newKeySet()},
4161	>	* {@link #newKeySet(int)}.
4162	>	*/
4163	>	public static class KeySetView<K,V> extends CHMView<K,V> implements Set<K>, java.io.Serializable {
4164	>	private static final long serialVersionUID = 7249069246763182397L;
4165	>	private final V value;
4166	>	KeySetView(ConcurrentHashMapV8<K, V> map, V value) {  // non-public
4167	>	super(map);
4168	>	this.value = value;
4169	>	}
4170	>
4171	>	/**
4172	>	* Returns the default mapped value for additions,
4173	>	* or {@code null} if additions are not supported.
4174	>	*
4175	>	* @return the default mapped value for additions, or {@code null}
4176	>	* if not supported.
4177	>	*/
4178	>	public V getMappedValue() { return value; }
4179	>
4180	>	// implement Set API
4181	>
4182	>	public boolean contains(Object o) { return map.containsKey(o); }
4183	>	public boolean remove(Object o)   { return map.remove(o) != null; }
4184	>
4185	>	/**
4186	>	* Returns a "weakly consistent" iterator that will never
4187	>	* throw {@link ConcurrentModificationException}, and
4188	>	* guarantees to traverse elements as they existed upon
4189	>	* construction of the iterator, and may (but is not
4190	>	* guaranteed to) reflect any modifications subsequent to
4191	>	* construction.
4192	>	*
4193	>	* @return an iterator over the keys of this map
4194	>	*/
4195	>	public Iterator<K> iterator()     { return new KeyIterator<K,V>(map); }
4196	>	public boolean add(K e) {
4197	>	V v;
4198	>	if ((v = value) == null)
4199	>	throw new UnsupportedOperationException();
4200	>	if (e == null)
4201	>	throw new NullPointerException();
4202	>	return map.internalPutIfAbsent(e, v) == null;
4203	>	}
4204	>	public boolean addAll(Collection<? extends K> c) {
4205	>	boolean added = false;
4206	>	V v;
4207	>	if ((v = value) == null)
4208	>	throw new UnsupportedOperationException();
4209	>	for (K e : c) {
4210	>	if (e == null)
4211	>	throw new NullPointerException();
4212	>	if (map.internalPutIfAbsent(e, v) == null)
4213	>	added = true;
4214	>	}
4215	>	return added;
4216	>	}
4217	>	public boolean equals(Object o) {
4218	>	Set<?> c;
4219	>	return ((o instanceof Set) &&
4220	>	((c = (Set<?>)o) == this ||
4221	>	(containsAll(c) && c.containsAll(this))));
4222	>	}
4223	>
4224	>	/**
4225	>	* Performs the given action for each key.
4226	>	*
4227	>	* @param action the action
4228	>	*/
4229	>	public void forEach(Action<K> action) {
4230	>	ForkJoinTasks.forEachKey
4231	>	(map, action).invoke();
4232	>	}
4233	>
4234	>	/**
4235	>	* Performs the given action for each non-null transformation
4236	>	* of each key.
4237	>	*
4238	>	* @param transformer a function returning the transformation
4239	>	* for an element, or null of there is no transformation (in
4240	>	* which case the action is not applied).
4241	>	* @param action the action
4242	>	*/
4243	>	public <U> void forEach(Fun<? super K, ? extends U> transformer,
4244	>	Action<U> action) {
4245	>	ForkJoinTasks.forEachKey
4246	>	(map, transformer, action).invoke();
4247	>	}
4248	>
4249	>	/**
4250	>	* Returns a non-null result from applying the given search
4251	>	* function on each key, or null if none. Upon success,
4252	>	* further element processing is suppressed and the results of
4253	>	* any other parallel invocations of the search function are
4254	>	* ignored.
4255	>	*
4256	>	* @param searchFunction a function returning a non-null
4257	>	* result on success, else null
4258	>	* @return a non-null result from applying the given search
4259	>	* function on each key, or null if none
4260	>	*/
4261	>	public <U> U search(Fun<? super K, ? extends U> searchFunction) {
4262	>	return ForkJoinTasks.searchKeys
4263	>	(map, searchFunction).invoke();
4264	>	}
4265	>
4266	>	/**
4267	>	* Returns the result of accumulating all keys using the given
4268	>	* reducer to combine values, or null if none.
4269	>	*
4270	>	* @param reducer a commutative associative combining function
4271	>	* @return the result of accumulating all keys using the given
4272	>	* reducer to combine values, or null if none
4273	>	*/
4274	>	public K reduce(BiFun<? super K, ? super K, ? extends K> reducer) {
4275	>	return ForkJoinTasks.reduceKeys
4276	>	(map, reducer).invoke();
4277	>	}
4278	>
4279	>	/**
4280	>	* Returns the result of accumulating the given transformation
4281	>	* of all keys using the given reducer to combine values, and
4282	>	* the given basis as an identity value.
4283	>	*
4284	>	* @param transformer a function returning the transformation
4285	>	* for an element
4286	>	* @param basis the identity (initial default value) for the reduction
4287	>	* @param reducer a commutative associative combining function
4288	>	* @return  the result of accumulating the given transformation
4289	>	* of all keys
4290	>	*/
4291	>	public double reduceToDouble(ObjectToDouble<? super K> transformer,
4292	>	double basis,
4293	>	DoubleByDoubleToDouble reducer) {
4294	>	return ForkJoinTasks.reduceKeysToDouble
4295	>	(map, transformer, basis, reducer).invoke();
4296	>	}
4297	>
4298	>
4299	>	/**
4300	>	* Returns the result of accumulating the given transformation
4301	>	* of all keys using the given reducer to combine values, and
4302	>	* the given basis as an identity value.
4303	>	*
4304	>	* @param transformer a function returning the transformation
4305	>	* for an element
4306	>	* @param basis the identity (initial default value) for the reduction
4307	>	* @param reducer a commutative associative combining function
4308	>	* @return the result of accumulating the given transformation
4309	>	* of all keys
4310	>	*/
4311	>	public long reduceToLong(ObjectToLong<? super K> transformer,
4312	>	long basis,
4313	>	LongByLongToLong reducer) {
4314	>	return ForkJoinTasks.reduceKeysToLong
4315	>	(map, transformer, basis, reducer).invoke();
4316	>	}
4317	>
4318	>	/**
4319	>	* Returns the result of accumulating the given transformation
4320	>	* of all keys using the given reducer to combine values, and
4321	>	* the given basis as an identity value.
4322	>	*
4323	>	* @param transformer a function returning the transformation
4324	>	* for an element
4325	>	* @param basis the identity (initial default value) for the reduction
4326	>	* @param reducer a commutative associative combining function
4327	>	* @return the result of accumulating the given transformation
4328	>	* of all keys
4329	>	*/
4330	>	public int reduceToInt(ObjectToInt<? super K> transformer,
4331	>	int basis,
4332	>	IntByIntToInt reducer) {
4333	>	return ForkJoinTasks.reduceKeysToInt
4334	>	(map, transformer, basis, reducer).invoke();
4335	>	}
4336	>
4337	>	}
4338	>
4339	>	/**
4340	>	* A view of a ConcurrentHashMapV8 as a {@link Collection} of
4341	>	* values, in which additions are disabled. This class cannot be
4342	>	* directly instantiated. See {@link #values},
4343	>	*
4344	>	* <p>The view's {@code iterator} is a "weakly consistent" iterator
4345	>	* that will never throw {@link ConcurrentModificationException},
4346	>	* and guarantees to traverse elements as they existed upon
4347	>	* construction of the iterator, and may (but is not guaranteed to)
4348	>	* reflect any modifications subsequent to construction.
4349	>	*/
4350	>	public static final class ValuesView<K,V> extends CHMView<K,V>
4351	>	implements Collection<V> {
4352	>	ValuesView(ConcurrentHashMapV8<K, V> map)   { super(map); }
4353	>	public final boolean contains(Object o) { return map.containsValue(o); }
4354	>	public final boolean remove(Object o) {
4355	>	if (o != null) {
4356	>	Iterator<V> it = new ValueIterator<K,V>(map);
4357	>	while (it.hasNext()) {
4358	>	if (o.equals(it.next())) {
4359	>	it.remove();
4360	>	return true;
4361	>	}
4362	>	}
4363	>	}
4364	>	return false;
4365	>	}
4366	>
4367	>	/**
4368	>	* Returns a "weakly consistent" iterator that will never
4369	>	* throw {@link ConcurrentModificationException}, and
4370	>	* guarantees to traverse elements as they existed upon
4371	>	* construction of the iterator, and may (but is not
4372	>	* guaranteed to) reflect any modifications subsequent to
4373	>	* construction.
4374	>	*
4375	>	* @return an iterator over the values of this map
4376	>	*/
4377	>	public final Iterator<V> iterator() {
4378	>	return new ValueIterator<K,V>(map);
4379	>	}
4380	>	public final boolean add(V e) {
4381	>	throw new UnsupportedOperationException();
4382	>	}
4383	>	public final boolean addAll(Collection<? extends V> c) {
4384	>	throw new UnsupportedOperationException();
4385	>	}
4386	>
4387	>	/**
4388	>	* Performs the given action for each value.
4389	>	*
4390	>	* @param action the action
4391	>	*/
4392	>	public void forEach(Action<V> action) {
4393	>	ForkJoinTasks.forEachValue
4394	>	(map, action).invoke();
4395	>	}
4396	>
4397	>	/**
4398	>	* Performs the given action for each non-null transformation
4399	>	* of each value.
4400	>	*
4401	>	* @param transformer a function returning the transformation
4402	>	* for an element, or null of there is no transformation (in
4403	>	* which case the action is not applied).
4404	>	*/
4405	>	public <U> void forEach(Fun<? super V, ? extends U> transformer,
4406	>	Action<U> action) {
4407	>	ForkJoinTasks.forEachValue
4408	>	(map, transformer, action).invoke();
4409	>	}
4410	>
4411	>	/**
4412	>	* Returns a non-null result from applying the given search
4413	>	* function on each value, or null if none.  Upon success,
4414	>	* further element processing is suppressed and the results of
4415	>	* any other parallel invocations of the search function are
4416	>	* ignored.
4417	>	*
4418	>	* @param searchFunction a function returning a non-null
4419	>	* result on success, else null
4420	>	* @return a non-null result from applying the given search
4421	>	* function on each value, or null if none
4422	>	*
4423	>	*/
4424	>	public <U> U search(Fun<? super V, ? extends U> searchFunction) {
4425	>	return ForkJoinTasks.searchValues
4426	>	(map, searchFunction).invoke();
4427	>	}
4428	>
4429	>	/**
4430	>	* Returns the result of accumulating all values using the
4431	>	* given reducer to combine values, or null if none.
4432	>	*
4433	>	* @param reducer a commutative associative combining function
4434	>	* @return  the result of accumulating all values
4435	>	*/
4436	>	public V reduce(BiFun<? super V, ? super V, ? extends V> reducer) {
4437	>	return ForkJoinTasks.reduceValues
4438	>	(map, reducer).invoke();
4439	>	}
4440	>
4441	>	/**
4442	>	* Returns the result of accumulating the given transformation
4443	>	* of all values using the given reducer to combine values, or
4444	>	* null if none.
4445	>	*
4446	>	* @param transformer a function returning the transformation
4447	>	* for an element, or null of there is no transformation (in
4448	>	* which case it is not combined).
4449	>	* @param reducer a commutative associative combining function
4450	>	* @return the result of accumulating the given transformation
4451	>	* of all values
4452	>	*/
4453	>	public <U> U reduce(Fun<? super V, ? extends U> transformer,
4454	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4455	>	return ForkJoinTasks.reduceValues
4456	>	(map, transformer, reducer).invoke();
4457	>	}
4458	>
4459	>	/**
4460	>	* Returns the result of accumulating the given transformation
4461	>	* of all values using the given reducer to combine values,
4462	>	* and the given basis as an identity value.
4463	>	*
4464	>	* @param transformer a function returning the transformation
4465	>	* for an element
4466	>	* @param basis the identity (initial default value) for the reduction
4467	>	* @param reducer a commutative associative combining function
4468	>	* @return the result of accumulating the given transformation
4469	>	* of all values
4470	>	*/
4471	>	public double reduceToDouble(ObjectToDouble<? super V> transformer,
4472	>	double basis,
4473	>	DoubleByDoubleToDouble reducer) {
4474	>	return ForkJoinTasks.reduceValuesToDouble
4475	>	(map, transformer, basis, reducer).invoke();
4476	>	}
4477	>
4478	>	/**
4479	>	* Returns the result of accumulating the given transformation
4480	>	* of all values using the given reducer to combine values,
4481	>	* and the given basis as an identity value.
4482	>	*
4483	>	* @param transformer a function returning the transformation
4484	>	* for an element
4485	>	* @param basis the identity (initial default value) for the reduction
4486	>	* @param reducer a commutative associative combining function
4487	>	* @return the result of accumulating the given transformation
4488	>	* of all values
4489	>	*/
4490	>	public long reduceToLong(ObjectToLong<? super V> transformer,
4491	>	long basis,
4492	>	LongByLongToLong reducer) {
4493	>	return ForkJoinTasks.reduceValuesToLong
4494	>	(map, transformer, basis, reducer).invoke();
4495	>	}
4496	>
4497	>	/**
4498	>	* Returns the result of accumulating the given transformation
4499	>	* of all values using the given reducer to combine values,
4500	>	* and the given basis as an identity value.
4501	>	*
4502	>	* @param transformer a function returning the transformation
4503	>	* for an element
4504	>	* @param basis the identity (initial default value) for the reduction
4505	>	* @param reducer a commutative associative combining function
4506	>	* @return the result of accumulating the given transformation
4507	>	* of all values
4508	>	*/
4509	>	public int reduceToInt(ObjectToInt<? super V> transformer,
4510	>	int basis,
4511	>	IntByIntToInt reducer) {
4512	>	return ForkJoinTasks.reduceValuesToInt
4513	>	(map, transformer, basis, reducer).invoke();
4514	>	}
4515	>
4516	>	}
4517	>
4518	>	/**
4519	>	* A view of a ConcurrentHashMapV8 as a {@link Set} of (key, value)
4520	>	* entries.  This class cannot be directly instantiated. See
4521	>	* {@link #entrySet}.
4522	>	*/
4523	>	public static final class EntrySetView<K,V> extends CHMView<K,V>
4524	>	implements Set<Map.Entry<K,V>> {
4525	>	EntrySetView(ConcurrentHashMapV8<K, V> map) { super(map); }
4526	>	public final boolean contains(Object o) {
4527	>	Object k, v, r; Map.Entry<?,?> e;
4528	>	return ((o instanceof Map.Entry) &&
4529	>	(k = (e = (Map.Entry<?,?>)o).getKey()) != null &&
4530	>	(r = map.get(k)) != null &&
4531	>	(v = e.getValue()) != null &&
4532	>	(v == r || v.equals(r)));
4533	>	}
4534	>	public final boolean remove(Object o) {
4535	>	Object k, v; Map.Entry<?,?> e;
4536	>	return ((o instanceof Map.Entry) &&
4537	>	(k = (e = (Map.Entry<?,?>)o).getKey()) != null &&
4538	>	(v = e.getValue()) != null &&
4539	>	map.remove(k, v));
4540	>	}
4541	>
4542	>	/**
4543	>	* Returns a "weakly consistent" iterator that will never
4544	>	* throw {@link ConcurrentModificationException}, and
4545	>	* guarantees to traverse elements as they existed upon
4546	>	* construction of the iterator, and may (but is not
4547	>	* guaranteed to) reflect any modifications subsequent to
4548	>	* construction.
4549	>	*
4550	>	* @return an iterator over the entries of this map
4551	>	*/
4552	>	public final Iterator<Map.Entry<K,V>> iterator() {
4553	>	return new EntryIterator<K,V>(map);
4554	>	}
4555	>
4556	>	public final boolean add(Entry<K,V> e) {
4557	>	K key = e.getKey();
4558	>	V value = e.getValue();
4559	>	if (key == null || value == null)
4560	>	throw new NullPointerException();
4561	>	return map.internalPut(key, value) == null;
4562	>	}
4563	>	public final boolean addAll(Collection<? extends Entry<K,V>> c) {
4564	>	boolean added = false;
4565	>	for (Entry<K,V> e : c) {
4566	>	if (add(e))
4567	>	added = true;
4568	>	}
4569	>	return added;
4570	>	}
4571	>	public boolean equals(Object o) {
4572	>	Set<?> c;
4573	>	return ((o instanceof Set) &&
4574	>	((c = (Set<?>)o) == this ||
4575	>	(containsAll(c) && c.containsAll(this))));
4576	>	}
4577	>
4578	>	/**
4579	>	* Performs the given action for each entry.
4580	>	*
4581	>	* @param action the action
4582	>	*/
4583	>	public void forEach(Action<Map.Entry<K,V>> action) {
4584	>	ForkJoinTasks.forEachEntry
4585	>	(map, action).invoke();
4586	>	}
4587	>
4588	>	/**
4589	>	* Performs the given action for each non-null transformation
4590	>	* of each entry.
4591	>	*
4592	>	* @param transformer a function returning the transformation
4593	>	* for an element, or null of there is no transformation (in
4594	>	* which case the action is not applied).
4595	>	* @param action the action
4596	>	*/
4597	>	public <U> void forEach(Fun<Map.Entry<K,V>, ? extends U> transformer,
4598	>	Action<U> action) {
4599	>	ForkJoinTasks.forEachEntry
4600	>	(map, transformer, action).invoke();
4601	>	}
4602	>
4603	>	/**
4604	>	* Returns a non-null result from applying the given search
4605	>	* function on each entry, or null if none.  Upon success,
4606	>	* further element processing is suppressed and the results of
4607	>	* any other parallel invocations of the search function are
4608	>	* ignored.
4609	>	*
4610	>	* @param searchFunction a function returning a non-null
4611	>	* result on success, else null
4612	>	* @return a non-null result from applying the given search
4613	>	* function on each entry, or null if none
4614	>	*/
4615	>	public <U> U search(Fun<Map.Entry<K,V>, ? extends U> searchFunction) {
4616	>	return ForkJoinTasks.searchEntries
4617	>	(map, searchFunction).invoke();
4618	>	}
4619	>
4620	>	/**
4621	>	* Returns the result of accumulating all entries using the
4622	>	* given reducer to combine values, or null if none.
4623	>	*
4624	>	* @param reducer a commutative associative combining function
4625	>	* @return the result of accumulating all entries
4626	>	*/
4627	>	public Map.Entry<K,V> reduce(BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer) {
4628	>	return ForkJoinTasks.reduceEntries
4629	>	(map, reducer).invoke();
4630	>	}
4631	>
4632	>	/**
4633	>	* Returns the result of accumulating the given transformation
4634	>	* of all entries using the given reducer to combine values,
4635	>	* or null if none.
4636	>	*
4637	>	* @param transformer a function returning the transformation
4638	>	* for an element, or null of there is no transformation (in
4639	>	* which case it is not combined).
4640	>	* @param reducer a commutative associative combining function
4641	>	* @return the result of accumulating the given transformation
4642	>	* of all entries
4643	>	*/
4644	>	public <U> U reduce(Fun<Map.Entry<K,V>, ? extends U> transformer,
4645	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4646	>	return ForkJoinTasks.reduceEntries
4647	>	(map, transformer, reducer).invoke();
4648	>	}
4649	>
4650	>	/**
4651	>	* Returns the result of accumulating the given transformation
4652	>	* of all entries using the given reducer to combine values,
4653	>	* and the given basis as an identity value.
4654	>	*
4655	>	* @param transformer a function returning the transformation
4656	>	* for an element
4657	>	* @param basis the identity (initial default value) for the reduction
4658	>	* @param reducer a commutative associative combining function
4659	>	* @return the result of accumulating the given transformation
4660	>	* of all entries
4661	>	*/
4662	>	public double reduceToDouble(ObjectToDouble<Map.Entry<K,V>> transformer,
4663	>	double basis,
4664	>	DoubleByDoubleToDouble reducer) {
4665	>	return ForkJoinTasks.reduceEntriesToDouble
4666	>	(map, transformer, basis, reducer).invoke();
4667	>	}
4668	>
4669	>	/**
4670	>	* Returns the result of accumulating the given transformation
4671	>	* of all entries using the given reducer to combine values,
4672	>	* and the given basis as an identity value.
4673	>	*
4674	>	* @param transformer a function returning the transformation
4675	>	* for an element
4676	>	* @param basis the identity (initial default value) for the reduction
4677	>	* @param reducer a commutative associative combining function
4678	>	* @return  the result of accumulating the given transformation
4679	>	* of all entries
4680	>	*/
4681	>	public long reduceToLong(ObjectToLong<Map.Entry<K,V>> transformer,
4682	>	long basis,
4683	>	LongByLongToLong reducer) {
4684	>	return ForkJoinTasks.reduceEntriesToLong
4685	>	(map, transformer, basis, reducer).invoke();
4686	>	}
4687	>
4688	>	/**
4689	>	* Returns the result of accumulating the given transformation
4690	>	* of all entries using the given reducer to combine values,
4691	>	* and the given basis as an identity value.
4692	>	*
4693	>	* @param transformer a function returning the transformation
4694	>	* for an element
4695	>	* @param basis the identity (initial default value) for the reduction
4696	>	* @param reducer a commutative associative combining function
4697	>	* @return the result of accumulating the given transformation
4698	>	* of all entries
4699	>	*/
4700	>	public int reduceToInt(ObjectToInt<Map.Entry<K,V>> transformer,
4701	>	int basis,
4702	>	IntByIntToInt reducer) {
4703	>	return ForkJoinTasks.reduceEntriesToInt
4704	>	(map, transformer, basis, reducer).invoke();
4705	>	}
4706	>
4707	>	}
4708	>
4709	>	// ---------------------------------------------------------------------
4710	>
4711	>	/**
4712	>	* Predefined tasks for performing bulk parallel operations on
4713	>	* ConcurrentHashMapV8s. These tasks follow the forms and rules used
4714	>	* for bulk operations. Each method has the same name, but returns
4715	>	* a task rather than invoking it. These methods may be useful in
4716	>	* custom applications such as submitting a task without waiting
4717	>	* for completion, using a custom pool, or combining with other
4718	>	* tasks.
4719	>	*/
4720	>	public static class ForkJoinTasks {
4721	>	private ForkJoinTasks() {}
4722	>
4723	>	/**
4724	>	* Returns a task that when invoked, performs the given
4725	>	* action for each (key, value)
4726	>	*
4727	>	* @param map the map
4728	>	* @param action the action
4729	>	* @return the task
4730	>	*/
4731	>	public static <K,V> ForkJoinTask<Void> forEach
4732	>	(ConcurrentHashMapV8<K,V> map,
4733	>	BiAction<K,V> action) {
4734	>	if (action == null) throw new NullPointerException();
4735	>	return new ForEachMappingTask<K,V>(map, null, -1, action);
4736	>	}
4737	>
4738	>	/**
4739	>	* Returns a task that when invoked, performs the given
4740	>	* action for each non-null transformation of each (key, value)
4741	>	*
4742	>	* @param map the map
4743	>	* @param transformer a function returning the transformation
4744	>	* for an element, or null if there is no transformation (in
4745	>	* which case the action is not applied)
4746	>	* @param action the action
4747	>	* @return the task
4748	>	*/
4749	>	public static <K,V,U> ForkJoinTask<Void> forEach
4750	>	(ConcurrentHashMapV8<K,V> map,
4751	>	BiFun<? super K, ? super V, ? extends U> transformer,
4752	>	Action<U> action) {
4753	>	if (transformer == null || action == null)
4754	>	throw new NullPointerException();
4755	>	return new ForEachTransformedMappingTask<K,V,U>
4756	>	(map, null, -1, transformer, action);
4757	>	}
4758	>
4759	>	/**
4760	>	* Returns a task that when invoked, returns a non-null result
4761	>	* from applying the given search function on each (key,
4762	>	* value), or null if none. Upon success, further element
4763	>	* processing is suppressed and the results of any other
4764	>	* parallel invocations of the search function are ignored.
4765	>	*
4766	>	* @param map the map
4767	>	* @param searchFunction a function returning a non-null
4768	>	* result on success, else null
4769	>	* @return the task
4770	>	*/
4771	>	public static <K,V,U> ForkJoinTask<U> search
4772	>	(ConcurrentHashMapV8<K,V> map,
4773	>	BiFun<? super K, ? super V, ? extends U> searchFunction) {
4774	>	if (searchFunction == null) throw new NullPointerException();
4775	>	return new SearchMappingsTask<K,V,U>
4776	>	(map, null, -1, searchFunction,
4777	>	new AtomicReference<U>());
4778	>	}
4779	>
4780	>	/**
4781	>	* Returns a task that when invoked, returns the result of
4782	>	* accumulating the given transformation of all (key, value) pairs
4783	>	* using the given reducer to combine values, or null if none.
4784	>	*
4785	>	* @param map the map
4786	>	* @param transformer a function returning the transformation
4787	>	* for an element, or null if there is no transformation (in
4788	>	* which case it is not combined).
4789	>	* @param reducer a commutative associative combining function
4790	>	* @return the task
4791	>	*/
4792	>	public static <K,V,U> ForkJoinTask<U> reduce
4793	>	(ConcurrentHashMapV8<K,V> map,
4794	>	BiFun<? super K, ? super V, ? extends U> transformer,
4795	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4796	>	if (transformer == null || reducer == null)
4797	>	throw new NullPointerException();
4798	>	return new MapReduceMappingsTask<K,V,U>
4799	>	(map, null, -1, null, transformer, reducer);
4800	>	}
4801	>
4802	>	/**
4803	>	* Returns a task that when invoked, returns the result of
4804	>	* accumulating the given transformation of all (key, value) pairs
4805	>	* using the given reducer to combine values, and the given
4806	>	* basis as an identity value.
4807	>	*
4808	>	* @param map the map
4809	>	* @param transformer a function returning the transformation
4810	>	* for an element
4811	>	* @param basis the identity (initial default value) for the reduction
4812	>	* @param reducer a commutative associative combining function
4813	>	* @return the task
4814	>	*/
4815	>	public static <K,V> ForkJoinTask<Double> reduceToDouble
4816	>	(ConcurrentHashMapV8<K,V> map,
4817	>	ObjectByObjectToDouble<? super K, ? super V> transformer,
4818	>	double basis,
4819	>	DoubleByDoubleToDouble reducer) {
4820	>	if (transformer == null || reducer == null)
4821	>	throw new NullPointerException();
4822	>	return new MapReduceMappingsToDoubleTask<K,V>
4823	>	(map, null, -1, null, transformer, basis, reducer);
4824	>	}
4825	>
4826	>	/**
4827	>	* Returns a task that when invoked, returns the result of
4828	>	* accumulating the given transformation of all (key, value) pairs
4829	>	* using the given reducer to combine values, and the given
4830	>	* basis as an identity value.
4831	>	*
4832	>	* @param map the map
4833	>	* @param transformer a function returning the transformation
4834	>	* for an element
4835	>	* @param basis the identity (initial default value) for the reduction
4836	>	* @param reducer a commutative associative combining function
4837	>	* @return the task
4838	>	*/
4839	>	public static <K,V> ForkJoinTask<Long> reduceToLong
4840	>	(ConcurrentHashMapV8<K,V> map,
4841	>	ObjectByObjectToLong<? super K, ? super V> transformer,
4842	>	long basis,
4843	>	LongByLongToLong reducer) {
4844	>	if (transformer == null || reducer == null)
4845	>	throw new NullPointerException();
4846	>	return new MapReduceMappingsToLongTask<K,V>
4847	>	(map, null, -1, null, transformer, basis, reducer);
4848	>	}
4849	>
4850	>	/**
4851	>	* Returns a task that when invoked, returns the result of
4852	>	* accumulating the given transformation of all (key, value) pairs
4853	>	* using the given reducer to combine values, and the given
4854	>	* basis as an identity value.
4855	>	*
4856	>	* @param transformer a function returning the transformation
4857	>	* for an element
4858	>	* @param basis the identity (initial default value) for the reduction
4859	>	* @param reducer a commutative associative combining function
4860	>	* @return the task
4861	>	*/
4862	>	public static <K,V> ForkJoinTask<Integer> reduceToInt
4863	>	(ConcurrentHashMapV8<K,V> map,
4864	>	ObjectByObjectToInt<? super K, ? super V> transformer,
4865	>	int basis,
4866	>	IntByIntToInt reducer) {
4867	>	if (transformer == null || reducer == null)
4868	>	throw new NullPointerException();
4869	>	return new MapReduceMappingsToIntTask<K,V>
4870	>	(map, null, -1, null, transformer, basis, reducer);
4871	>	}
4872	>
4873	>	/**
4874	>	* Returns a task that when invoked, performs the given action
4875	>	* for each key.
4876	>	*
4877	>	* @param map the map
4878	>	* @param action the action
4879	>	* @return the task
4880	>	*/
4881	>	public static <K,V> ForkJoinTask<Void> forEachKey
4882	>	(ConcurrentHashMapV8<K,V> map,
4883	>	Action<K> action) {
4884	>	if (action == null) throw new NullPointerException();
4885	>	return new ForEachKeyTask<K,V>(map, null, -1, action);
4886	>	}
4887	>
4888	>	/**
4889	>	* Returns a task that when invoked, performs the given action
4890	>	* for each non-null transformation of each key.
4891	>	*
4892	>	* @param map the map
4893	>	* @param transformer a function returning the transformation
4894	>	* for an element, or null if there is no transformation (in
4895	>	* which case the action is not applied)
4896	>	* @param action the action
4897	>	* @return the task
4898	>	*/
4899	>	public static <K,V,U> ForkJoinTask<Void> forEachKey
4900	>	(ConcurrentHashMapV8<K,V> map,
4901	>	Fun<? super K, ? extends U> transformer,
4902	>	Action<U> action) {
4903	>	if (transformer == null || action == null)
4904	>	throw new NullPointerException();
4905	>	return new ForEachTransformedKeyTask<K,V,U>
4906	>	(map, null, -1, transformer, action);
4907	>	}
4908	>
4909	>	/**
4910	>	* Returns a task that when invoked, returns a non-null result
4911	>	* from applying the given search function on each key, or
4912	>	* null if none.  Upon success, further element processing is
4913	>	* suppressed and the results of any other parallel
4914	>	* invocations of the search function are ignored.
4915	>	*
4916	>	* @param map the map
4917	>	* @param searchFunction a function returning a non-null
4918	>	* result on success, else null
4919	>	* @return the task
4920	>	*/
4921	>	public static <K,V,U> ForkJoinTask<U> searchKeys
4922	>	(ConcurrentHashMapV8<K,V> map,
4923	>	Fun<? super K, ? extends U> searchFunction) {
4924	>	if (searchFunction == null) throw new NullPointerException();
4925	>	return new SearchKeysTask<K,V,U>
4926	>	(map, null, -1, searchFunction,
4927	>	new AtomicReference<U>());
4928	>	}
4929	>
4930	>	/**
4931	>	* Returns a task that when invoked, returns the result of
4932	>	* accumulating all keys using the given reducer to combine
4933	>	* values, or null if none.
4934	>	*
4935	>	* @param map the map
4936	>	* @param reducer a commutative associative combining function
4937	>	* @return the task
4938	>	*/
4939	>	public static <K,V> ForkJoinTask<K> reduceKeys
4940	>	(ConcurrentHashMapV8<K,V> map,
4941	>	BiFun<? super K, ? super K, ? extends K> reducer) {
4942	>	if (reducer == null) throw new NullPointerException();
4943	>	return new ReduceKeysTask<K,V>
4944	>	(map, null, -1, null, reducer);
4945	>	}
4946	>
4947	>	/**
4948	>	* Returns a task that when invoked, returns the result of
4949	>	* accumulating the given transformation of all keys using the given
4950	>	* reducer to combine values, or null if none.
4951	>	*
4952	>	* @param map the map
4953	>	* @param transformer a function returning the transformation
4954	>	* for an element, or null if there is no transformation (in
4955	>	* which case it is not combined).
4956	>	* @param reducer a commutative associative combining function
4957	>	* @return the task
4958	>	*/
4959	>	public static <K,V,U> ForkJoinTask<U> reduceKeys
4960	>	(ConcurrentHashMapV8<K,V> map,
4961	>	Fun<? super K, ? extends U> transformer,
4962	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4963	>	if (transformer == null || reducer == null)
4964	>	throw new NullPointerException();
4965	>	return new MapReduceKeysTask<K,V,U>
4966	>	(map, null, -1, null, transformer, reducer);
4967	>	}
4968	>
4969	>	/**
4970	>	* Returns a task that when invoked, returns the result of
4971	>	* accumulating the given transformation of all keys using the given
4972	>	* reducer to combine values, and the given basis as an
4973	>	* identity value.
4974	>	*
4975	>	* @param map the map
4976	>	* @param transformer a function returning the transformation
4977	>	* for an element
4978	>	* @param basis the identity (initial default value) for the reduction
4979	>	* @param reducer a commutative associative combining function
4980	>	* @return the task
4981	>	*/
4982	>	public static <K,V> ForkJoinTask<Double> reduceKeysToDouble
4983	>	(ConcurrentHashMapV8<K,V> map,
4984	>	ObjectToDouble<? super K> transformer,
4985	>	double basis,
4986	>	DoubleByDoubleToDouble reducer) {
4987	>	if (transformer == null || reducer == null)
4988	>	throw new NullPointerException();
4989	>	return new MapReduceKeysToDoubleTask<K,V>
4990	>	(map, null, -1, null, transformer, basis, reducer);
4991	>	}
4992	>
4993	>	/**
4994	>	* Returns a task that when invoked, returns the result of
4995	>	* accumulating the given transformation of all keys using the given
4996	>	* reducer to combine values, and the given basis as an
4997	>	* identity value.
4998	>	*
4999	>	* @param map the map
5000	>	* @param transformer a function returning the transformation
5001	>	* for an element
5002	>	* @param basis the identity (initial default value) for the reduction
5003	>	* @param reducer a commutative associative combining function
5004	>	* @return the task
5005	>	*/
5006	>	public static <K,V> ForkJoinTask<Long> reduceKeysToLong
5007	>	(ConcurrentHashMapV8<K,V> map,
5008	>	ObjectToLong<? super K> transformer,
5009	>	long basis,
5010	>	LongByLongToLong reducer) {
5011	>	if (transformer == null || reducer == null)
5012	>	throw new NullPointerException();
5013	>	return new MapReduceKeysToLongTask<K,V>
5014	>	(map, null, -1, null, transformer, basis, reducer);
5015	>	}
5016	>
5017	>	/**
5018	>	* Returns a task that when invoked, returns the result of
5019	>	* accumulating the given transformation of all keys using the given
5020	>	* reducer to combine values, and the given basis as an
5021	>	* identity value.
5022	>	*
5023	>	* @param map the map
5024	>	* @param transformer a function returning the transformation
5025	>	* for an element
5026	>	* @param basis the identity (initial default value) for the reduction
5027	>	* @param reducer a commutative associative combining function
5028	>	* @return the task
5029	>	*/
5030	>	public static <K,V> ForkJoinTask<Integer> reduceKeysToInt
5031	>	(ConcurrentHashMapV8<K,V> map,
5032	>	ObjectToInt<? super K> transformer,
5033	>	int basis,
5034	>	IntByIntToInt reducer) {
5035	>	if (transformer == null || reducer == null)
5036	>	throw new NullPointerException();
5037	>	return new MapReduceKeysToIntTask<K,V>
5038	>	(map, null, -1, null, transformer, basis, reducer);
5039	>	}
5040	>
5041	>	/**
5042	>	* Returns a task that when invoked, performs the given action
5043	>	* for each value.
5044	>	*
5045	>	* @param map the map
5046	>	* @param action the action
5047	>	*/
5048	>	public static <K,V> ForkJoinTask<Void> forEachValue
5049	>	(ConcurrentHashMapV8<K,V> map,
5050	>	Action<V> action) {
5051	>	if (action == null) throw new NullPointerException();
5052	>	return new ForEachValueTask<K,V>(map, null, -1, action);
5053	>	}
5054	>
5055	>	/**
5056	>	* Returns a task that when invoked, performs the given action
5057	>	* for each non-null transformation of each value.
5058	>	*
5059	>	* @param map the map
5060	>	* @param transformer a function returning the transformation
5061	>	* for an element, or null if there is no transformation (in
5062	>	* which case the action is not applied)
5063	>	* @param action the action
5064	>	*/
5065	>	public static <K,V,U> ForkJoinTask<Void> forEachValue
5066	>	(ConcurrentHashMapV8<K,V> map,
5067	>	Fun<? super V, ? extends U> transformer,
5068	>	Action<U> action) {
5069	>	if (transformer == null || action == null)
5070	>	throw new NullPointerException();
5071	>	return new ForEachTransformedValueTask<K,V,U>
5072	>	(map, null, -1, transformer, action);
5073	>	}
5074	>
5075	>	/**
5076	>	* Returns a task that when invoked, returns a non-null result
5077	>	* from applying the given search function on each value, or
5078	>	* null if none.  Upon success, further element processing is
5079	>	* suppressed and the results of any other parallel
5080	>	* invocations of the search function are ignored.
5081	>	*
5082	>	* @param map the map
5083	>	* @param searchFunction a function returning a non-null
5084	>	* result on success, else null
5085	>	* @return the task
5086	>	*/
5087	>	public static <K,V,U> ForkJoinTask<U> searchValues
5088	>	(ConcurrentHashMapV8<K,V> map,
5089	>	Fun<? super V, ? extends U> searchFunction) {
5090	>	if (searchFunction == null) throw new NullPointerException();
5091	>	return new SearchValuesTask<K,V,U>
5092	>	(map, null, -1, searchFunction,
5093	>	new AtomicReference<U>());
5094	>	}
5095	>
5096	>	/**
5097	>	* Returns a task that when invoked, returns the result of
5098	>	* accumulating all values using the given reducer to combine
5099	>	* values, or null if none.
5100	>	*
5101	>	* @param map the map
5102	>	* @param reducer a commutative associative combining function
5103	>	* @return the task
5104	>	*/
5105	>	public static <K,V> ForkJoinTask<V> reduceValues
5106	>	(ConcurrentHashMapV8<K,V> map,
5107	>	BiFun<? super V, ? super V, ? extends V> reducer) {
5108	>	if (reducer == null) throw new NullPointerException();
5109	>	return new ReduceValuesTask<K,V>
5110	>	(map, null, -1, null, reducer);
5111	>	}
5112	>
5113	>	/**
5114	>	* Returns a task that when invoked, returns the result of
5115	>	* accumulating the given transformation of all values using the
5116	>	* given reducer to combine values, or null if none.
5117	>	*
5118	>	* @param map the map
5119	>	* @param transformer a function returning the transformation
5120	>	* for an element, or null if there is no transformation (in
5121	>	* which case it is not combined).
5122	>	* @param reducer a commutative associative combining function
5123	>	* @return the task
5124	>	*/
5125	>	public static <K,V,U> ForkJoinTask<U> reduceValues
5126	>	(ConcurrentHashMapV8<K,V> map,
5127	>	Fun<? super V, ? extends U> transformer,
5128	>	BiFun<? super U, ? super U, ? extends U> reducer) {
5129	>	if (transformer == null || reducer == null)
5130	>	throw new NullPointerException();
5131	>	return new MapReduceValuesTask<K,V,U>
5132	>	(map, null, -1, null, transformer, reducer);
5133	>	}
5134	>
5135	>	/**
5136	>	* Returns a task that when invoked, returns the result of
5137	>	* accumulating the given transformation of all values using the
5138	>	* given reducer to combine values, and the given basis as an
5139	>	* identity value.
5140	>	*
5141	>	* @param map the map
5142	>	* @param transformer a function returning the transformation
5143	>	* for an element
5144	>	* @param basis the identity (initial default value) for the reduction
5145	>	* @param reducer a commutative associative combining function
5146	>	* @return the task
5147	>	*/
5148	>	public static <K,V> ForkJoinTask<Double> reduceValuesToDouble
5149	>	(ConcurrentHashMapV8<K,V> map,
5150	>	ObjectToDouble<? super V> transformer,
5151	>	double basis,
5152	>	DoubleByDoubleToDouble reducer) {
5153	>	if (transformer == null || reducer == null)
5154	>	throw new NullPointerException();
5155	>	return new MapReduceValuesToDoubleTask<K,V>
5156	>	(map, null, -1, null, transformer, basis, reducer);
5157	>	}
5158	>
5159	>	/**
5160	>	* Returns a task that when invoked, returns the result of
5161	>	* accumulating the given transformation of all values using the
5162	>	* given reducer to combine values, and the given basis as an
5163	>	* identity value.
5164	>	*
5165	>	* @param map the map
5166	>	* @param transformer a function returning the transformation
5167	>	* for an element
5168	>	* @param basis the identity (initial default value) for the reduction
5169	>	* @param reducer a commutative associative combining function
5170	>	* @return the task
5171	>	*/
5172	>	public static <K,V> ForkJoinTask<Long> reduceValuesToLong
5173	>	(ConcurrentHashMapV8<K,V> map,
5174	>	ObjectToLong<? super V> transformer,
5175	>	long basis,
5176	>	LongByLongToLong reducer) {
5177	>	if (transformer == null || reducer == null)
5178	>	throw new NullPointerException();
5179	>	return new MapReduceValuesToLongTask<K,V>
5180	>	(map, null, -1, null, transformer, basis, reducer);
5181	>	}
5182	>
5183	>	/**
5184	>	* Returns a task that when invoked, returns the result of
5185	>	* accumulating the given transformation of all values using the
5186	>	* given reducer to combine values, and the given basis as an
5187	>	* identity value.
5188	>	*
5189	>	* @param map the map
5190	>	* @param transformer a function returning the transformation
5191	>	* for an element
5192	>	* @param basis the identity (initial default value) for the reduction
5193	>	* @param reducer a commutative associative combining function
5194	>	* @return the task
5195	>	*/
5196	>	public static <K,V> ForkJoinTask<Integer> reduceValuesToInt
5197	>	(ConcurrentHashMapV8<K,V> map,
5198	>	ObjectToInt<? super V> transformer,
5199	>	int basis,
5200	>	IntByIntToInt reducer) {
5201	>	if (transformer == null || reducer == null)
5202	>	throw new NullPointerException();
5203	>	return new MapReduceValuesToIntTask<K,V>
5204	>	(map, null, -1, null, transformer, basis, reducer);
5205	>	}
5206	>
5207	>	/**
5208	>	* Returns a task that when invoked, perform the given action
5209	>	* for each entry.
5210	>	*
5211	>	* @param map the map
5212	>	* @param action the action
5213	>	*/
5214	>	public static <K,V> ForkJoinTask<Void> forEachEntry
5215	>	(ConcurrentHashMapV8<K,V> map,
5216	>	Action<Map.Entry<K,V>> action) {
5217	>	if (action == null) throw new NullPointerException();
5218	>	return new ForEachEntryTask<K,V>(map, null, -1, action);
5219	>	}
5220	>
5221	>	/**
5222	>	* Returns a task that when invoked, perform the given action
5223	>	* for each non-null transformation of each entry.
5224	>	*
5225	>	* @param map the map
5226	>	* @param transformer a function returning the transformation
5227	>	* for an element, or null if there is no transformation (in
5228	>	* which case the action is not applied)
5229	>	* @param action the action
5230	>	*/
5231	>	public static <K,V,U> ForkJoinTask<Void> forEachEntry
5232	>	(ConcurrentHashMapV8<K,V> map,
5233	>	Fun<Map.Entry<K,V>, ? extends U> transformer,
5234	>	Action<U> action) {
5235	>	if (transformer == null || action == null)
5236	>	throw new NullPointerException();
5237	>	return new ForEachTransformedEntryTask<K,V,U>
5238	>	(map, null, -1, transformer, action);
5239	>	}
5240	>
5241	>	/**
5242	>	* Returns a task that when invoked, returns a non-null result
5243	>	* from applying the given search function on each entry, or
5244	>	* null if none.  Upon success, further element processing is
5245	>	* suppressed and the results of any other parallel
5246	>	* invocations of the search function are ignored.
5247	>	*
5248	>	* @param map the map
5249	>	* @param searchFunction a function returning a non-null
5250	>	* result on success, else null
5251	>	* @return the task
5252	>	*/
5253	>	public static <K,V,U> ForkJoinTask<U> searchEntries
5254	>	(ConcurrentHashMapV8<K,V> map,
5255	>	Fun<Map.Entry<K,V>, ? extends U> searchFunction) {
5256	>	if (searchFunction == null) throw new NullPointerException();
5257	>	return new SearchEntriesTask<K,V,U>
5258	>	(map, null, -1, searchFunction,
5259	>	new AtomicReference<U>());
5260	>	}
5261	>
5262	>	/**
5263	>	* Returns a task that when invoked, returns the result of
5264	>	* accumulating all entries using the given reducer to combine
5265	>	* values, or null if none.
5266	>	*
5267	>	* @param map the map
5268	>	* @param reducer a commutative associative combining function
5269	>	* @return the task
5270	>	*/
5271	>	public static <K,V> ForkJoinTask<Map.Entry<K,V>> reduceEntries
5272	>	(ConcurrentHashMapV8<K,V> map,
5273	>	BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer) {
5274	>	if (reducer == null) throw new NullPointerException();
5275	>	return new ReduceEntriesTask<K,V>
5276	>	(map, null, -1, null, reducer);
5277	>	}
5278	>
5279	>	/**
5280	>	* Returns a task that when invoked, returns the result of
5281	>	* accumulating the given transformation of all entries using the
5282	>	* given reducer to combine values, or null if none.
5283	>	*
5284	>	* @param map the map
5285	>	* @param transformer a function returning the transformation
5286	>	* for an element, or null if there is no transformation (in
5287	>	* which case it is not combined).
5288	>	* @param reducer a commutative associative combining function
5289	>	* @return the task
5290	>	*/
5291	>	public static <K,V,U> ForkJoinTask<U> reduceEntries
5292	>	(ConcurrentHashMapV8<K,V> map,
5293	>	Fun<Map.Entry<K,V>, ? extends U> transformer,
5294	>	BiFun<? super U, ? super U, ? extends U> reducer) {
5295	>	if (transformer == null || reducer == null)
5296	>	throw new NullPointerException();
5297	>	return new MapReduceEntriesTask<K,V,U>
5298	>	(map, null, -1, null, transformer, reducer);
5299	>	}
5300	>
5301	>	/**
5302	>	* Returns a task that when invoked, returns the result of
5303	>	* accumulating the given transformation of all entries using the
5304	>	* given reducer to combine values, and the given basis as an
5305	>	* identity value.
5306	>	*
5307	>	* @param map the map
5308	>	* @param transformer a function returning the transformation
5309	>	* for an element
5310	>	* @param basis the identity (initial default value) for the reduction
5311	>	* @param reducer a commutative associative combining function
5312	>	* @return the task
5313	>	*/
5314	>	public static <K,V> ForkJoinTask<Double> reduceEntriesToDouble
5315	>	(ConcurrentHashMapV8<K,V> map,
5316	>	ObjectToDouble<Map.Entry<K,V>> transformer,
5317	>	double basis,
5318	>	DoubleByDoubleToDouble reducer) {
5319	>	if (transformer == null || reducer == null)
5320	>	throw new NullPointerException();
5321	>	return new MapReduceEntriesToDoubleTask<K,V>
5322	>	(map, null, -1, null, transformer, basis, reducer);
5323	>	}
5324	>
5325	>	/**
5326	>	* Returns a task that when invoked, returns the result of
5327	>	* accumulating the given transformation of all entries using the
5328	>	* given reducer to combine values, and the given basis as an
5329	>	* identity value.
5330	>	*
5331	>	* @param map the map
5332	>	* @param transformer a function returning the transformation
5333	>	* for an element
5334	>	* @param basis the identity (initial default value) for the reduction
5335	>	* @param reducer a commutative associative combining function
5336	>	* @return the task
5337	>	*/
5338	>	public static <K,V> ForkJoinTask<Long> reduceEntriesToLong
5339	>	(ConcurrentHashMapV8<K,V> map,
5340	>	ObjectToLong<Map.Entry<K,V>> transformer,
5341	>	long basis,
5342	>	LongByLongToLong reducer) {
5343	>	if (transformer == null || reducer == null)
5344	>	throw new NullPointerException();
5345	>	return new MapReduceEntriesToLongTask<K,V>
5346	>	(map, null, -1, null, transformer, basis, reducer);
5347	>	}
5348	>
5349	>	/**
5350	>	* Returns a task that when invoked, returns the result of
5351	>	* accumulating the given transformation of all entries using the
5352	>	* given reducer to combine values, and the given basis as an
5353	>	* identity value.
5354	>	*
5355	>	* @param map the map
5356	>	* @param transformer a function returning the transformation
5357	>	* for an element
5358	>	* @param basis the identity (initial default value) for the reduction
5359	>	* @param reducer a commutative associative combining function
5360	>	* @return the task
5361	>	*/
5362	>	public static <K,V> ForkJoinTask<Integer> reduceEntriesToInt
5363	>	(ConcurrentHashMapV8<K,V> map,
5364	>	ObjectToInt<Map.Entry<K,V>> transformer,
5365	>	int basis,
5366	>	IntByIntToInt reducer) {
5367	>	if (transformer == null || reducer == null)
5368	>	throw new NullPointerException();
5369	>	return new MapReduceEntriesToIntTask<K,V>
5370	>	(map, null, -1, null, transformer, basis, reducer);
5371	>	}
5372	>	}
5373	>
5374	>	// -------------------------------------------------------
5375	>
5376	>	/*
5377	>	* Task classes. Coded in a regular but ugly format/style to
5378	>	* simplify checks that each variant differs in the right way from
5379	>	* others.
5380	>	*/
5381	>
5382	>	@SuppressWarnings("serial") static final class ForEachKeyTask<K,V>
5383	>	extends Traverser<K,V,Void> {
5384	>	final Action<K> action;
5385	>	ForEachKeyTask
5386	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5387	>	Action<K> action) {
5388	>	super(m, p, b);
5389	>	this.action = action;
5390	>	}
5391	>	@SuppressWarnings("unchecked") public final void compute() {
5392	>	final Action<K> action;
5393	>	if ((action = this.action) == null)
5394	>	throw new NullPointerException();
5395	>	for (int b; (b = preSplit()) > 0;)
5396	>	new ForEachKeyTask<K,V>(map, this, b, action).fork();
5397	>	while (advance() != null)
5398	>	action.apply((K)nextKey);
5399	>	propagateCompletion();
5400	>	}
5401	>	}
5402	>
5403	>	@SuppressWarnings("serial") static final class ForEachValueTask<K,V>
5404	>	extends Traverser<K,V,Void> {
5405	>	final Action<V> action;
5406	>	ForEachValueTask
5407	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5408	>	Action<V> action) {
5409	>	super(m, p, b);
5410	>	this.action = action;
5411	>	}
5412	>	@SuppressWarnings("unchecked") public final void compute() {
5413	>	final Action<V> action;
5414	>	if ((action = this.action) == null)
5415	>	throw new NullPointerException();
5416	>	for (int b; (b = preSplit()) > 0;)
5417	>	new ForEachValueTask<K,V>(map, this, b, action).fork();
5418	>	Object v;
5419	>	while ((v = advance()) != null)
5420	>	action.apply((V)v);
5421	>	propagateCompletion();
5422	>	}
5423	>	}
5424	>
5425	>	@SuppressWarnings("serial") static final class ForEachEntryTask<K,V>
5426	>	extends Traverser<K,V,Void> {
5427	>	final Action<Entry<K,V>> action;
5428	>	ForEachEntryTask
5429	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5430	>	Action<Entry<K,V>> action) {
5431	>	super(m, p, b);
5432	>	this.action = action;
5433	>	}
5434	>	@SuppressWarnings("unchecked") public final void compute() {
5435	>	final Action<Entry<K,V>> action;
5436	>	if ((action = this.action) == null)
5437	>	throw new NullPointerException();
5438	>	for (int b; (b = preSplit()) > 0;)
5439	>	new ForEachEntryTask<K,V>(map, this, b, action).fork();
5440	>	Object v;
5441	>	while ((v = advance()) != null)
5442	>	action.apply(entryFor((K)nextKey, (V)v));
5443	>	propagateCompletion();
5444	>	}
5445	>	}
5446	>
5447	>	@SuppressWarnings("serial") static final class ForEachMappingTask<K,V>
5448	>	extends Traverser<K,V,Void> {
5449	>	final BiAction<K,V> action;
5450	>	ForEachMappingTask
5451	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5452	>	BiAction<K,V> action) {
5453	>	super(m, p, b);
5454	>	this.action = action;
5455	>	}
5456	>	@SuppressWarnings("unchecked") public final void compute() {
5457	>	final BiAction<K,V> action;
5458	>	if ((action = this.action) == null)
5459	>	throw new NullPointerException();
5460	>	for (int b; (b = preSplit()) > 0;)
5461	>	new ForEachMappingTask<K,V>(map, this, b, action).fork();
5462	>	Object v;
5463	>	while ((v = advance()) != null)
5464	>	action.apply((K)nextKey, (V)v);
5465	>	propagateCompletion();
5466	>	}
5467	>	}
5468	>
5469	>	@SuppressWarnings("serial") static final class ForEachTransformedKeyTask<K,V,U>
5470	>	extends Traverser<K,V,Void> {
5471	>	final Fun<? super K, ? extends U> transformer;
5472	>	final Action<U> action;
5473	>	ForEachTransformedKeyTask
5474	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5475	>	Fun<? super K, ? extends U> transformer, Action<U> action) {
5476	>	super(m, p, b);
5477	>	this.transformer = transformer; this.action = action;
5478	>	}
5479	>	@SuppressWarnings("unchecked") public final void compute() {
5480	>	final Fun<? super K, ? extends U> transformer;
5481	>	final Action<U> action;
5482	>	if ((transformer = this.transformer) == null ||
5483	>	(action = this.action) == null)
5484	>	throw new NullPointerException();
5485	>	for (int b; (b = preSplit()) > 0;)
5486	>	new ForEachTransformedKeyTask<K,V,U>
5487	>	(map, this, b, transformer, action).fork();
5488	>	U u;
5489	>	while (advance() != null) {
5490	>	if ((u = transformer.apply((K)nextKey)) != null)
5491	>	action.apply(u);
5492	>	}
5493	>	propagateCompletion();
5494	>	}
5495	>	}
5496	>
5497	>	@SuppressWarnings("serial") static final class ForEachTransformedValueTask<K,V,U>
5498	>	extends Traverser<K,V,Void> {
5499	>	final Fun<? super V, ? extends U> transformer;
5500	>	final Action<U> action;
5501	>	ForEachTransformedValueTask
5502	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5503	>	Fun<? super V, ? extends U> transformer, Action<U> action) {
5504	>	super(m, p, b);
5505	>	this.transformer = transformer; this.action = action;
5506	>	}
5507	>	@SuppressWarnings("unchecked") public final void compute() {
5508	>	final Fun<? super V, ? extends U> transformer;
5509	>	final Action<U> action;
5510	>	if ((transformer = this.transformer) == null ||
5511	>	(action = this.action) == null)
5512	>	throw new NullPointerException();
5513	>	for (int b; (b = preSplit()) > 0;)
5514	>	new ForEachTransformedValueTask<K,V,U>
5515	>	(map, this, b, transformer, action).fork();
5516	>	Object v; U u;
5517	>	while ((v = advance()) != null) {
5518	>	if ((u = transformer.apply((V)v)) != null)
5519	>	action.apply(u);
5520	>	}
5521	>	propagateCompletion();
5522	>	}
5523	>	}
5524	>
5525	>	@SuppressWarnings("serial") static final class ForEachTransformedEntryTask<K,V,U>
5526	>	extends Traverser<K,V,Void> {
5527	>	final Fun<Map.Entry<K,V>, ? extends U> transformer;
5528	>	final Action<U> action;
5529	>	ForEachTransformedEntryTask
5530	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5531	>	Fun<Map.Entry<K,V>, ? extends U> transformer, Action<U> action) {
5532	>	super(m, p, b);
5533	>	this.transformer = transformer; this.action = action;
5534	>	}
5535	>	@SuppressWarnings("unchecked") public final void compute() {
5536	>	final Fun<Map.Entry<K,V>, ? extends U> transformer;
5537	>	final Action<U> action;
5538	>	if ((transformer = this.transformer) == null ||
5539	>	(action = this.action) == null)
5540	>	throw new NullPointerException();
5541	>	for (int b; (b = preSplit()) > 0;)
5542	>	new ForEachTransformedEntryTask<K,V,U>
5543	>	(map, this, b, transformer, action).fork();
5544	>	Object v; U u;
5545	>	while ((v = advance()) != null) {
5546	>	if ((u = transformer.apply(entryFor((K)nextKey, (V)v))) != null)
5547	>	action.apply(u);
5548	>	}
5549	>	propagateCompletion();
5550	>	}
5551	>	}
5552	>
5553	>	@SuppressWarnings("serial") static final class ForEachTransformedMappingTask<K,V,U>
5554	>	extends Traverser<K,V,Void> {
5555	>	final BiFun<? super K, ? super V, ? extends U> transformer;
5556	>	final Action<U> action;
5557	>	ForEachTransformedMappingTask
5558	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5559	>	BiFun<? super K, ? super V, ? extends U> transformer,
5560	>	Action<U> action) {
5561	>	super(m, p, b);
5562	>	this.transformer = transformer; this.action = action;
5563	>	}
5564	>	@SuppressWarnings("unchecked") public final void compute() {
5565	>	final BiFun<? super K, ? super V, ? extends U> transformer;
5566	>	final Action<U> action;
5567	>	if ((transformer = this.transformer) == null ||
5568	>	(action = this.action) == null)
5569	>	throw new NullPointerException();
5570	>	for (int b; (b = preSplit()) > 0;)
5571	>	new ForEachTransformedMappingTask<K,V,U>
5572	>	(map, this, b, transformer, action).fork();
5573	>	Object v; U u;
5574	>	while ((v = advance()) != null) {
5575	>	if ((u = transformer.apply((K)nextKey, (V)v)) != null)
5576	>	action.apply(u);
5577	>	}
5578	>	propagateCompletion();
5579	>	}
5580	>	}
5581	>
5582	>	@SuppressWarnings("serial") static final class SearchKeysTask<K,V,U>
5583	>	extends Traverser<K,V,U> {
5584	>	final Fun<? super K, ? extends U> searchFunction;
5585	>	final AtomicReference<U> result;
5586	>	SearchKeysTask
5587	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5588	>	Fun<? super K, ? extends U> searchFunction,
5589	>	AtomicReference<U> result) {
5590	>	super(m, p, b);
5591	>	this.searchFunction = searchFunction; this.result = result;
5592	>	}
5593	>	public final U getRawResult() { return result.get(); }
5594	>	@SuppressWarnings("unchecked") public final void compute() {
5595	>	final Fun<? super K, ? extends U> searchFunction;
5596	>	final AtomicReference<U> result;
5597	>	if ((searchFunction = this.searchFunction) == null ||
5598	>	(result = this.result) == null)
5599	>	throw new NullPointerException();
5600	>	for (int b;;) {
5601	>	if (result.get() != null)
5602	>	return;
5603	>	if ((b = preSplit()) <= 0)
5604	>	break;
5605	>	new SearchKeysTask<K,V,U>
5606	>	(map, this, b, searchFunction, result).fork();
5607	>	}
5608	>	while (result.get() == null) {
5609	>	U u;
5610	>	if (advance() == null) {
5611	>	propagateCompletion();
5612	>	break;
5613	>	}
5614	>	if ((u = searchFunction.apply((K)nextKey)) != null) {
5615	>	if (result.compareAndSet(null, u))
5616	>	quietlyCompleteRoot();
5617	>	break;
5618	>	}
5619	>	}
5620	>	}
5621	>	}
5622	>
5623	>	@SuppressWarnings("serial") static final class SearchValuesTask<K,V,U>
5624	>	extends Traverser<K,V,U> {
5625	>	final Fun<? super V, ? extends U> searchFunction;
5626	>	final AtomicReference<U> result;
5627	>	SearchValuesTask
5628	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5629	>	Fun<? super V, ? extends U> searchFunction,
5630	>	AtomicReference<U> result) {
5631	>	super(m, p, b);
5632	>	this.searchFunction = searchFunction; this.result = result;
5633	>	}
5634	>	public final U getRawResult() { return result.get(); }
5635	>	@SuppressWarnings("unchecked") public final void compute() {
5636	>	final Fun<? super V, ? extends U> searchFunction;
5637	>	final AtomicReference<U> result;
5638	>	if ((searchFunction = this.searchFunction) == null ||
5639	>	(result = this.result) == null)
5640	>	throw new NullPointerException();
5641	>	for (int b;;) {
5642	>	if (result.get() != null)
5643	>	return;
5644	>	if ((b = preSplit()) <= 0)
5645	>	break;
5646	>	new SearchValuesTask<K,V,U>
5647	>	(map, this, b, searchFunction, result).fork();
5648	>	}
5649	>	while (result.get() == null) {
5650	>	Object v; U u;
5651	>	if ((v = advance()) == null) {
5652	>	propagateCompletion();
5653	>	break;
5654	>	}
5655	>	if ((u = searchFunction.apply((V)v)) != null) {
5656	>	if (result.compareAndSet(null, u))
5657	>	quietlyCompleteRoot();
5658	>	break;
5659	>	}
5660	>	}
5661	>	}
5662	>	}
5663	>
5664	>	@SuppressWarnings("serial") static final class SearchEntriesTask<K,V,U>
5665	>	extends Traverser<K,V,U> {
5666	>	final Fun<Entry<K,V>, ? extends U> searchFunction;
5667	>	final AtomicReference<U> result;
5668	>	SearchEntriesTask
5669	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5670	>	Fun<Entry<K,V>, ? extends U> searchFunction,
5671	>	AtomicReference<U> result) {
5672	>	super(m, p, b);
5673	>	this.searchFunction = searchFunction; this.result = result;
5674	>	}
5675	>	public final U getRawResult() { return result.get(); }
5676	>	@SuppressWarnings("unchecked") public final void compute() {
5677	>	final Fun<Entry<K,V>, ? extends U> searchFunction;
5678	>	final AtomicReference<U> result;
5679	>	if ((searchFunction = this.searchFunction) == null ||
5680	>	(result = this.result) == null)
5681	>	throw new NullPointerException();
5682	>	for (int b;;) {
5683	>	if (result.get() != null)
5684	>	return;
5685	>	if ((b = preSplit()) <= 0)
5686	>	break;
5687	>	new SearchEntriesTask<K,V,U>
5688	>	(map, this, b, searchFunction, result).fork();
5689	>	}
5690	>	while (result.get() == null) {
5691	>	Object v; U u;
5692	>	if ((v = advance()) == null) {
5693	>	propagateCompletion();
5694	>	break;
5695	>	}
5696	>	if ((u = searchFunction.apply(entryFor((K)nextKey, (V)v))) != null) {
5697	>	if (result.compareAndSet(null, u))
5698	>	quietlyCompleteRoot();
5699	>	return;
5700	>	}
5701	>	}
5702	>	}
5703	>	}
5704	>
5705	>	@SuppressWarnings("serial") static final class SearchMappingsTask<K,V,U>
5706	>	extends Traverser<K,V,U> {
5707	>	final BiFun<? super K, ? super V, ? extends U> searchFunction;
5708	>	final AtomicReference<U> result;
5709	>	SearchMappingsTask
5710	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5711	>	BiFun<? super K, ? super V, ? extends U> searchFunction,
5712	>	AtomicReference<U> result) {
5713	>	super(m, p, b);
5714	>	this.searchFunction = searchFunction; this.result = result;
5715	>	}
5716	>	public final U getRawResult() { return result.get(); }
5717	>	@SuppressWarnings("unchecked") public final void compute() {
5718	>	final BiFun<? super K, ? super V, ? extends U> searchFunction;
5719	>	final AtomicReference<U> result;
5720	>	if ((searchFunction = this.searchFunction) == null ||
5721	>	(result = this.result) == null)
5722	>	throw new NullPointerException();
5723	>	for (int b;;) {
5724	>	if (result.get() != null)
5725	>	return;
5726	>	if ((b = preSplit()) <= 0)
5727	>	break;
5728	>	new SearchMappingsTask<K,V,U>
5729	>	(map, this, b, searchFunction, result).fork();
5730	>	}
5731	>	while (result.get() == null) {
5732	>	Object v; U u;
5733	>	if ((v = advance()) == null) {
5734	>	propagateCompletion();
5735	>	break;
5736	>	}
5737	>	if ((u = searchFunction.apply((K)nextKey, (V)v)) != null) {
5738	>	if (result.compareAndSet(null, u))
5739	>	quietlyCompleteRoot();
5740	>	break;
5741	>	}
5742	>	}
5743	>	}
5744	>	}
5745	>
5746	>	@SuppressWarnings("serial") static final class ReduceKeysTask<K,V>
5747	>	extends Traverser<K,V,K> {
5748	>	final BiFun<? super K, ? super K, ? extends K> reducer;
5749	>	K result;
5750	>	ReduceKeysTask<K,V> rights, nextRight;
5751	>	ReduceKeysTask
5752	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5753	>	ReduceKeysTask<K,V> nextRight,
5754	>	BiFun<? super K, ? super K, ? extends K> reducer) {
5755	>	super(m, p, b); this.nextRight = nextRight;
5756	>	this.reducer = reducer;
5757	>	}
5758	>	public final K getRawResult() { return result; }
5759	>	@SuppressWarnings("unchecked") public final void compute() {
5760	>	final BiFun<? super K, ? super K, ? extends K> reducer =
5761	>	this.reducer;
5762	>	if (reducer == null)
5763	>	throw new NullPointerException();
5764	>	for (int b; (b = preSplit()) > 0;)
5765	>	(rights = new ReduceKeysTask<K,V>
5766	>	(map, this, b, rights, reducer)).fork();
5767	>	K r = null;
5768	>	while (advance() != null) {
5769	>	K u = (K)nextKey;
5770	>	r = (r == null) ? u : reducer.apply(r, u);
5771	>	}
5772	>	result = r;
5773	>	CountedCompleter<?> c;
5774	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5775	>	ReduceKeysTask<K,V>
5776	>	t = (ReduceKeysTask<K,V>)c,
5777	>	s = t.rights;
5778	>	while (s != null) {
5779	>	K tr, sr;
5780	>	if ((sr = s.result) != null)
5781	>	t.result = (((tr = t.result) == null) ? sr :
5782	>	reducer.apply(tr, sr));
5783	>	s = t.rights = s.nextRight;
5784	>	}
5785	>	}
5786	>	}
5787	>	}
5788	>
5789	>	@SuppressWarnings("serial") static final class ReduceValuesTask<K,V>
5790	>	extends Traverser<K,V,V> {
5791	>	final BiFun<? super V, ? super V, ? extends V> reducer;
5792	>	V result;
5793	>	ReduceValuesTask<K,V> rights, nextRight;
5794	>	ReduceValuesTask
5795	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5796	>	ReduceValuesTask<K,V> nextRight,
5797	>	BiFun<? super V, ? super V, ? extends V> reducer) {
5798	>	super(m, p, b); this.nextRight = nextRight;
5799	>	this.reducer = reducer;
5800	>	}
5801	>	public final V getRawResult() { return result; }
5802	>	@SuppressWarnings("unchecked") public final void compute() {
5803	>	final BiFun<? super V, ? super V, ? extends V> reducer =
5804	>	this.reducer;
5805	>	if (reducer == null)
5806	>	throw new NullPointerException();
5807	>	for (int b; (b = preSplit()) > 0;)
5808	>	(rights = new ReduceValuesTask<K,V>
5809	>	(map, this, b, rights, reducer)).fork();
5810	>	V r = null;
5811	>	Object v;
5812	>	while ((v = advance()) != null) {
5813	>	V u = (V)v;
5814	>	r = (r == null) ? u : reducer.apply(r, u);
5815	>	}
5816	>	result = r;
5817	>	CountedCompleter<?> c;
5818	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5819	>	ReduceValuesTask<K,V>
5820	>	t = (ReduceValuesTask<K,V>)c,
5821	>	s = t.rights;
5822	>	while (s != null) {
5823	>	V tr, sr;
5824	>	if ((sr = s.result) != null)
5825	>	t.result = (((tr = t.result) == null) ? sr :
5826	>	reducer.apply(tr, sr));
5827	>	s = t.rights = s.nextRight;
5828	>	}
5829	>	}
5830	>	}
5831	>	}
5832	>
5833	>	@SuppressWarnings("serial") static final class ReduceEntriesTask<K,V>
5834	>	extends Traverser<K,V,Map.Entry<K,V>> {
5835	>	final BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer;
5836	>	Map.Entry<K,V> result;
5837	>	ReduceEntriesTask<K,V> rights, nextRight;
5838	>	ReduceEntriesTask
5839	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5840	>	ReduceEntriesTask<K,V> nextRight,
5841	>	BiFun<Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer) {
5842	>	super(m, p, b); this.nextRight = nextRight;
5843	>	this.reducer = reducer;
5844	>	}
5845	>	public final Map.Entry<K,V> getRawResult() { return result; }
5846	>	@SuppressWarnings("unchecked") public final void compute() {
5847	>	final BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer =
5848	>	this.reducer;
5849	>	if (reducer == null)
5850	>	throw new NullPointerException();
5851	>	for (int b; (b = preSplit()) > 0;)
5852	>	(rights = new ReduceEntriesTask<K,V>
5853	>	(map, this, b, rights, reducer)).fork();
5854	>	Map.Entry<K,V> r = null;
5855	>	Object v;
5856	>	while ((v = advance()) != null) {
5857	>	Map.Entry<K,V> u = entryFor((K)nextKey, (V)v);
5858	>	r = (r == null) ? u : reducer.apply(r, u);
5859	>	}
5860	>	result = r;
5861	>	CountedCompleter<?> c;
5862	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5863	>	ReduceEntriesTask<K,V>
5864	>	t = (ReduceEntriesTask<K,V>)c,
5865	>	s = t.rights;
5866	>	while (s != null) {
5867	>	Map.Entry<K,V> tr, sr;
5868	>	if ((sr = s.result) != null)
5869	>	t.result = (((tr = t.result) == null) ? sr :
5870	>	reducer.apply(tr, sr));
5871	>	s = t.rights = s.nextRight;
5872	>	}
5873	>	}
5874	>	}
5875	>	}
5876	>
5877	>	@SuppressWarnings("serial") static final class MapReduceKeysTask<K,V,U>
5878	>	extends Traverser<K,V,U> {
5879	>	final Fun<? super K, ? extends U> transformer;
5880	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5881	>	U result;
5882	>	MapReduceKeysTask<K,V,U> rights, nextRight;
5883	>	MapReduceKeysTask
5884	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5885	>	MapReduceKeysTask<K,V,U> nextRight,
5886	>	Fun<? super K, ? extends U> transformer,
5887	>	BiFun<? super U, ? super U, ? extends U> reducer) {
5888	>	super(m, p, b); this.nextRight = nextRight;
5889	>	this.transformer = transformer;
5890	>	this.reducer = reducer;
5891	>	}
5892	>	public final U getRawResult() { return result; }
5893	>	@SuppressWarnings("unchecked") public final void compute() {
5894	>	final Fun<? super K, ? extends U> transformer =
5895	>	this.transformer;
5896	>	final BiFun<? super U, ? super U, ? extends U> reducer =
5897	>	this.reducer;
5898	>	if (transformer == null || reducer == null)
5899	>	throw new NullPointerException();
5900	>	for (int b; (b = preSplit()) > 0;)
5901	>	(rights = new MapReduceKeysTask<K,V,U>
5902	>	(map, this, b, rights, transformer, reducer)).fork();
5903	>	U r = null, u;
5904	>	while (advance() != null) {
5905	>	if ((u = transformer.apply((K)nextKey)) != null)
5906	>	r = (r == null) ? u : reducer.apply(r, u);
5907	>	}
5908	>	result = r;
5909	>	CountedCompleter<?> c;
5910	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5911	>	MapReduceKeysTask<K,V,U>
5912	>	t = (MapReduceKeysTask<K,V,U>)c,
5913	>	s = t.rights;
5914	>	while (s != null) {
5915	>	U tr, sr;
5916	>	if ((sr = s.result) != null)
5917	>	t.result = (((tr = t.result) == null) ? sr :
5918	>	reducer.apply(tr, sr));
5919	>	s = t.rights = s.nextRight;
5920	>	}
5921	>	}
5922	>	}
5923	>	}
5924	>
5925	>	@SuppressWarnings("serial") static final class MapReduceValuesTask<K,V,U>
5926	>	extends Traverser<K,V,U> {
5927	>	final Fun<? super V, ? extends U> transformer;
5928	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5929	>	U result;
5930	>	MapReduceValuesTask<K,V,U> rights, nextRight;
5931	>	MapReduceValuesTask
5932	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5933	>	MapReduceValuesTask<K,V,U> nextRight,
5934	>	Fun<? super V, ? extends U> transformer,
5935	>	BiFun<? super U, ? super U, ? extends U> reducer) {
5936	>	super(m, p, b); this.nextRight = nextRight;
5937	>	this.transformer = transformer;
5938	>	this.reducer = reducer;
5939	>	}
5940	>	public final U getRawResult() { return result; }
5941	>	@SuppressWarnings("unchecked") public final void compute() {
5942	>	final Fun<? super V, ? extends U> transformer =
5943	>	this.transformer;
5944	>	final BiFun<? super U, ? super U, ? extends U> reducer =
5945	>	this.reducer;
5946	>	if (transformer == null || reducer == null)
5947	>	throw new NullPointerException();
5948	>	for (int b; (b = preSplit()) > 0;)
5949	>	(rights = new MapReduceValuesTask<K,V,U>
5950	>	(map, this, b, rights, transformer, reducer)).fork();
5951	>	U r = null, u;
5952	>	Object v;
5953	>	while ((v = advance()) != null) {
5954	>	if ((u = transformer.apply((V)v)) != null)
5955	>	r = (r == null) ? u : reducer.apply(r, u);
5956	>	}
5957	>	result = r;
5958	>	CountedCompleter<?> c;
5959	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5960	>	MapReduceValuesTask<K,V,U>
5961	>	t = (MapReduceValuesTask<K,V,U>)c,
5962	>	s = t.rights;
5963	>	while (s != null) {
5964	>	U tr, sr;
5965	>	if ((sr = s.result) != null)
5966	>	t.result = (((tr = t.result) == null) ? sr :
5967	>	reducer.apply(tr, sr));
5968	>	s = t.rights = s.nextRight;
5969	>	}
5970	>	}
5971	>	}
5972	>	}
5973	>
5974	>	@SuppressWarnings("serial") static final class MapReduceEntriesTask<K,V,U>
5975	>	extends Traverser<K,V,U> {
5976	>	final Fun<Map.Entry<K,V>, ? extends U> transformer;
5977	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5978	>	U result;
5979	>	MapReduceEntriesTask<K,V,U> rights, nextRight;
5980	>	MapReduceEntriesTask
5981	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5982	>	MapReduceEntriesTask<K,V,U> nextRight,
5983	>	Fun<Map.Entry<K,V>, ? extends U> transformer,
5984	>	BiFun<? super U, ? super U, ? extends U> reducer) {
5985	>	super(m, p, b); this.nextRight = nextRight;
5986	>	this.transformer = transformer;
5987	>	this.reducer = reducer;
5988	>	}
5989	>	public final U getRawResult() { return result; }
5990	>	@SuppressWarnings("unchecked") public final void compute() {
5991	>	final Fun<Map.Entry<K,V>, ? extends U> transformer =
5992	>	this.transformer;
5993	>	final BiFun<? super U, ? super U, ? extends U> reducer =
5994	>	this.reducer;
5995	>	if (transformer == null || reducer == null)
5996	>	throw new NullPointerException();
5997	>	for (int b; (b = preSplit()) > 0;)
5998	>	(rights = new MapReduceEntriesTask<K,V,U>
5999	>	(map, this, b, rights, transformer, reducer)).fork();
6000	>	U r = null, u;
6001	>	Object v;
6002	>	while ((v = advance()) != null) {
6003	>	if ((u = transformer.apply(entryFor((K)nextKey, (V)v))) != null)
6004	>	r = (r == null) ? u : reducer.apply(r, u);
6005	>	}
6006	>	result = r;
6007	>	CountedCompleter<?> c;
6008	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6009	>	MapReduceEntriesTask<K,V,U>
6010	>	t = (MapReduceEntriesTask<K,V,U>)c,
6011	>	s = t.rights;
6012	>	while (s != null) {
6013	>	U tr, sr;
6014	>	if ((sr = s.result) != null)
6015	>	t.result = (((tr = t.result) == null) ? sr :
6016	>	reducer.apply(tr, sr));
6017	>	s = t.rights = s.nextRight;
6018	>	}
6019	>	}
6020	>	}
6021	>	}
6022	>
6023	>	@SuppressWarnings("serial") static final class MapReduceMappingsTask<K,V,U>
6024	>	extends Traverser<K,V,U> {
6025	>	final BiFun<? super K, ? super V, ? extends U> transformer;
6026	>	final BiFun<? super U, ? super U, ? extends U> reducer;
6027	>	U result;
6028	>	MapReduceMappingsTask<K,V,U> rights, nextRight;
6029	>	MapReduceMappingsTask
6030	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6031	>	MapReduceMappingsTask<K,V,U> nextRight,
6032	>	BiFun<? super K, ? super V, ? extends U> transformer,
6033	>	BiFun<? super U, ? super U, ? extends U> reducer) {
6034	>	super(m, p, b); this.nextRight = nextRight;
6035	>	this.transformer = transformer;
6036	>	this.reducer = reducer;
6037	>	}
6038	>	public final U getRawResult() { return result; }
6039	>	@SuppressWarnings("unchecked") public final void compute() {
6040	>	final BiFun<? super K, ? super V, ? extends U> transformer =
6041	>	this.transformer;
6042	>	final BiFun<? super U, ? super U, ? extends U> reducer =
6043	>	this.reducer;
6044	>	if (transformer == null || reducer == null)
6045	>	throw new NullPointerException();
6046	>	for (int b; (b = preSplit()) > 0;)
6047	>	(rights = new MapReduceMappingsTask<K,V,U>
6048	>	(map, this, b, rights, transformer, reducer)).fork();
6049	>	U r = null, u;
6050	>	Object v;
6051	>	while ((v = advance()) != null) {
6052	>	if ((u = transformer.apply((K)nextKey, (V)v)) != null)
6053	>	r = (r == null) ? u : reducer.apply(r, u);
6054	>	}
6055	>	result = r;
6056	>	CountedCompleter<?> c;
6057	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6058	>	MapReduceMappingsTask<K,V,U>
6059	>	t = (MapReduceMappingsTask<K,V,U>)c,
6060	>	s = t.rights;
6061	>	while (s != null) {
6062	>	U tr, sr;
6063	>	if ((sr = s.result) != null)
6064	>	t.result = (((tr = t.result) == null) ? sr :
6065	>	reducer.apply(tr, sr));
6066	>	s = t.rights = s.nextRight;
6067	>	}
6068	>	}
6069	>	}
6070	>	}
6071	>
6072	>	@SuppressWarnings("serial") static final class MapReduceKeysToDoubleTask<K,V>
6073	>	extends Traverser<K,V,Double> {
6074	>	final ObjectToDouble<? super K> transformer;
6075	>	final DoubleByDoubleToDouble reducer;
6076	>	final double basis;
6077	>	double result;
6078	>	MapReduceKeysToDoubleTask<K,V> rights, nextRight;
6079	>	MapReduceKeysToDoubleTask
6080	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6081	>	MapReduceKeysToDoubleTask<K,V> nextRight,
6082	>	ObjectToDouble<? super K> transformer,
6083	>	double basis,
6084	>	DoubleByDoubleToDouble reducer) {
6085	>	super(m, p, b); this.nextRight = nextRight;
6086	>	this.transformer = transformer;
6087	>	this.basis = basis; this.reducer = reducer;
6088	>	}
6089	>	public final Double getRawResult() { return result; }
6090	>	@SuppressWarnings("unchecked") public final void compute() {
6091	>	final ObjectToDouble<? super K> transformer =
6092	>	this.transformer;
6093	>	final DoubleByDoubleToDouble reducer = this.reducer;
6094	>	if (transformer == null || reducer == null)
6095	>	throw new NullPointerException();
6096	>	double r = this.basis;
6097	>	for (int b; (b = preSplit()) > 0;)
6098	>	(rights = new MapReduceKeysToDoubleTask<K,V>
6099	>	(map, this, b, rights, transformer, r, reducer)).fork();
6100	>	while (advance() != null)
6101	>	r = reducer.apply(r, transformer.apply((K)nextKey));
6102	>	result = r;
6103	>	CountedCompleter<?> c;
6104	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6105	>	MapReduceKeysToDoubleTask<K,V>
6106	>	t = (MapReduceKeysToDoubleTask<K,V>)c,
6107	>	s = t.rights;
6108	>	while (s != null) {
6109	>	t.result = reducer.apply(t.result, s.result);
6110	>	s = t.rights = s.nextRight;
6111	>	}
6112	>	}
6113	>	}
6114	>	}
6115	>
6116	>	@SuppressWarnings("serial") static final class MapReduceValuesToDoubleTask<K,V>
6117	>	extends Traverser<K,V,Double> {
6118	>	final ObjectToDouble<? super V> transformer;
6119	>	final DoubleByDoubleToDouble reducer;
6120	>	final double basis;
6121	>	double result;
6122	>	MapReduceValuesToDoubleTask<K,V> rights, nextRight;
6123	>	MapReduceValuesToDoubleTask
6124	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6125	>	MapReduceValuesToDoubleTask<K,V> nextRight,
6126	>	ObjectToDouble<? super V> transformer,
6127	>	double basis,
6128	>	DoubleByDoubleToDouble reducer) {
6129	>	super(m, p, b); this.nextRight = nextRight;
6130	>	this.transformer = transformer;
6131	>	this.basis = basis; this.reducer = reducer;
6132	>	}
6133	>	public final Double getRawResult() { return result; }
6134	>	@SuppressWarnings("unchecked") public final void compute() {
6135	>	final ObjectToDouble<? super V> transformer =
6136	>	this.transformer;
6137	>	final DoubleByDoubleToDouble reducer = this.reducer;
6138	>	if (transformer == null || reducer == null)
6139	>	throw new NullPointerException();
6140	>	double r = this.basis;
6141	>	for (int b; (b = preSplit()) > 0;)
6142	>	(rights = new MapReduceValuesToDoubleTask<K,V>
6143	>	(map, this, b, rights, transformer, r, reducer)).fork();
6144	>	Object v;
6145	>	while ((v = advance()) != null)
6146	>	r = reducer.apply(r, transformer.apply((V)v));
6147	>	result = r;
6148	>	CountedCompleter<?> c;
6149	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6150	>	MapReduceValuesToDoubleTask<K,V>
6151	>	t = (MapReduceValuesToDoubleTask<K,V>)c,
6152	>	s = t.rights;
6153	>	while (s != null) {
6154	>	t.result = reducer.apply(t.result, s.result);
6155	>	s = t.rights = s.nextRight;
6156	>	}
6157	>	}
6158	>	}
6159	>	}
6160	>
6161	>	@SuppressWarnings("serial") static final class MapReduceEntriesToDoubleTask<K,V>
6162	>	extends Traverser<K,V,Double> {
6163	>	final ObjectToDouble<Map.Entry<K,V>> transformer;
6164	>	final DoubleByDoubleToDouble reducer;
6165	>	final double basis;
6166	>	double result;
6167	>	MapReduceEntriesToDoubleTask<K,V> rights, nextRight;
6168	>	MapReduceEntriesToDoubleTask
6169	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6170	>	MapReduceEntriesToDoubleTask<K,V> nextRight,
6171	>	ObjectToDouble<Map.Entry<K,V>> transformer,
6172	>	double basis,
6173	>	DoubleByDoubleToDouble reducer) {
6174	>	super(m, p, b); this.nextRight = nextRight;
6175	>	this.transformer = transformer;
6176	>	this.basis = basis; this.reducer = reducer;
6177	>	}
6178	>	public final Double getRawResult() { return result; }
6179	>	@SuppressWarnings("unchecked") public final void compute() {
6180	>	final ObjectToDouble<Map.Entry<K,V>> transformer =
6181	>	this.transformer;
6182	>	final DoubleByDoubleToDouble reducer = this.reducer;
6183	>	if (transformer == null || reducer == null)
6184	>	throw new NullPointerException();
6185	>	double r = this.basis;
6186	>	for (int b; (b = preSplit()) > 0;)
6187	>	(rights = new MapReduceEntriesToDoubleTask<K,V>
6188	>	(map, this, b, rights, transformer, r, reducer)).fork();
6189	>	Object v;
6190	>	while ((v = advance()) != null)
6191	>	r = reducer.apply(r, transformer.apply(entryFor((K)nextKey, (V)v)));
6192	>	result = r;
6193	>	CountedCompleter<?> c;
6194	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6195	>	MapReduceEntriesToDoubleTask<K,V>
6196	>	t = (MapReduceEntriesToDoubleTask<K,V>)c,
6197	>	s = t.rights;
6198	>	while (s != null) {
6199	>	t.result = reducer.apply(t.result, s.result);
6200	>	s = t.rights = s.nextRight;
6201	>	}
6202	>	}
6203	>	}
6204	>	}
6205	>
6206	>	@SuppressWarnings("serial") static final class MapReduceMappingsToDoubleTask<K,V>
6207	>	extends Traverser<K,V,Double> {
6208	>	final ObjectByObjectToDouble<? super K, ? super V> transformer;
6209	>	final DoubleByDoubleToDouble reducer;
6210	>	final double basis;
6211	>	double result;
6212	>	MapReduceMappingsToDoubleTask<K,V> rights, nextRight;
6213	>	MapReduceMappingsToDoubleTask
6214	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6215	>	MapReduceMappingsToDoubleTask<K,V> nextRight,
6216	>	ObjectByObjectToDouble<? super K, ? super V> transformer,
6217	>	double basis,
6218	>	DoubleByDoubleToDouble reducer) {
6219	>	super(m, p, b); this.nextRight = nextRight;
6220	>	this.transformer = transformer;
6221	>	this.basis = basis; this.reducer = reducer;
6222	>	}
6223	>	public final Double getRawResult() { return result; }
6224	>	@SuppressWarnings("unchecked") public final void compute() {
6225	>	final ObjectByObjectToDouble<? super K, ? super V> transformer =
6226	>	this.transformer;
6227	>	final DoubleByDoubleToDouble reducer = this.reducer;
6228	>	if (transformer == null || reducer == null)
6229	>	throw new NullPointerException();
6230	>	double r = this.basis;
6231	>	for (int b; (b = preSplit()) > 0;)
6232	>	(rights = new MapReduceMappingsToDoubleTask<K,V>
6233	>	(map, this, b, rights, transformer, r, reducer)).fork();
6234	>	Object v;
6235	>	while ((v = advance()) != null)
6236	>	r = reducer.apply(r, transformer.apply((K)nextKey, (V)v));
6237	>	result = r;
6238	>	CountedCompleter<?> c;
6239	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6240	>	MapReduceMappingsToDoubleTask<K,V>
6241	>	t = (MapReduceMappingsToDoubleTask<K,V>)c,
6242	>	s = t.rights;
6243	>	while (s != null) {
6244	>	t.result = reducer.apply(t.result, s.result);
6245	>	s = t.rights = s.nextRight;
6246	>	}
6247	>	}
6248	>	}
6249	>	}
6250	>
6251	>	@SuppressWarnings("serial") static final class MapReduceKeysToLongTask<K,V>
6252	>	extends Traverser<K,V,Long> {
6253	>	final ObjectToLong<? super K> transformer;
6254	>	final LongByLongToLong reducer;
6255	>	final long basis;
6256	>	long result;
6257	>	MapReduceKeysToLongTask<K,V> rights, nextRight;
6258	>	MapReduceKeysToLongTask
6259	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6260	>	MapReduceKeysToLongTask<K,V> nextRight,
6261	>	ObjectToLong<? super K> transformer,
6262	>	long basis,
6263	>	LongByLongToLong reducer) {
6264	>	super(m, p, b); this.nextRight = nextRight;
6265	>	this.transformer = transformer;
6266	>	this.basis = basis; this.reducer = reducer;
6267	>	}
6268	>	public final Long getRawResult() { return result; }
6269	>	@SuppressWarnings("unchecked") public final void compute() {
6270	>	final ObjectToLong<? super K> transformer =
6271	>	this.transformer;
6272	>	final LongByLongToLong reducer = this.reducer;
6273	>	if (transformer == null || reducer == null)
6274	>	throw new NullPointerException();
6275	>	long r = this.basis;
6276	>	for (int b; (b = preSplit()) > 0;)
6277	>	(rights = new MapReduceKeysToLongTask<K,V>
6278	>	(map, this, b, rights, transformer, r, reducer)).fork();
6279	>	while (advance() != null)
6280	>	r = reducer.apply(r, transformer.apply((K)nextKey));
6281	>	result = r;
6282	>	CountedCompleter<?> c;
6283	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6284	>	MapReduceKeysToLongTask<K,V>
6285	>	t = (MapReduceKeysToLongTask<K,V>)c,
6286	>	s = t.rights;
6287	>	while (s != null) {
6288	>	t.result = reducer.apply(t.result, s.result);
6289	>	s = t.rights = s.nextRight;
6290	>	}
6291	>	}
6292	>	}
6293	>	}
6294	>
6295	>	@SuppressWarnings("serial") static final class MapReduceValuesToLongTask<K,V>
6296	>	extends Traverser<K,V,Long> {
6297	>	final ObjectToLong<? super V> transformer;
6298	>	final LongByLongToLong reducer;
6299	>	final long basis;
6300	>	long result;
6301	>	MapReduceValuesToLongTask<K,V> rights, nextRight;
6302	>	MapReduceValuesToLongTask
6303	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6304	>	MapReduceValuesToLongTask<K,V> nextRight,
6305	>	ObjectToLong<? super V> transformer,
6306	>	long basis,
6307	>	LongByLongToLong reducer) {
6308	>	super(m, p, b); this.nextRight = nextRight;
6309	>	this.transformer = transformer;
6310	>	this.basis = basis; this.reducer = reducer;
6311	>	}
6312	>	public final Long getRawResult() { return result; }
6313	>	@SuppressWarnings("unchecked") public final void compute() {
6314	>	final ObjectToLong<? super V> transformer =
6315	>	this.transformer;
6316	>	final LongByLongToLong reducer = this.reducer;
6317	>	if (transformer == null || reducer == null)
6318	>	throw new NullPointerException();
6319	>	long r = this.basis;
6320	>	for (int b; (b = preSplit()) > 0;)
6321	>	(rights = new MapReduceValuesToLongTask<K,V>
6322	>	(map, this, b, rights, transformer, r, reducer)).fork();
6323	>	Object v;
6324	>	while ((v = advance()) != null)
6325	>	r = reducer.apply(r, transformer.apply((V)v));
6326	>	result = r;
6327	>	CountedCompleter<?> c;
6328	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6329	>	MapReduceValuesToLongTask<K,V>
6330	>	t = (MapReduceValuesToLongTask<K,V>)c,
6331	>	s = t.rights;
6332	>	while (s != null) {
6333	>	t.result = reducer.apply(t.result, s.result);
6334	>	s = t.rights = s.nextRight;
6335	>	}
6336	>	}
6337	>	}
6338	>	}
6339	>
6340	>	@SuppressWarnings("serial") static final class MapReduceEntriesToLongTask<K,V>
6341	>	extends Traverser<K,V,Long> {
6342	>	final ObjectToLong<Map.Entry<K,V>> transformer;
6343	>	final LongByLongToLong reducer;
6344	>	final long basis;
6345	>	long result;
6346	>	MapReduceEntriesToLongTask<K,V> rights, nextRight;
6347	>	MapReduceEntriesToLongTask
6348	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6349	>	MapReduceEntriesToLongTask<K,V> nextRight,
6350	>	ObjectToLong<Map.Entry<K,V>> transformer,
6351	>	long basis,
6352	>	LongByLongToLong reducer) {
6353	>	super(m, p, b); this.nextRight = nextRight;
6354	>	this.transformer = transformer;
6355	>	this.basis = basis; this.reducer = reducer;
6356	>	}
6357	>	public final Long getRawResult() { return result; }
6358	>	@SuppressWarnings("unchecked") public final void compute() {
6359	>	final ObjectToLong<Map.Entry<K,V>> transformer =
6360	>	this.transformer;
6361	>	final LongByLongToLong reducer = this.reducer;
6362	>	if (transformer == null || reducer == null)
6363	>	throw new NullPointerException();
6364	>	long r = this.basis;
6365	>	for (int b; (b = preSplit()) > 0;)
6366	>	(rights = new MapReduceEntriesToLongTask<K,V>
6367	>	(map, this, b, rights, transformer, r, reducer)).fork();
6368	>	Object v;
6369	>	while ((v = advance()) != null)
6370	>	r = reducer.apply(r, transformer.apply(entryFor((K)nextKey, (V)v)));
6371	>	result = r;
6372	>	CountedCompleter<?> c;
6373	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6374	>	MapReduceEntriesToLongTask<K,V>
6375	>	t = (MapReduceEntriesToLongTask<K,V>)c,
6376	>	s = t.rights;
6377	>	while (s != null) {
6378	>	t.result = reducer.apply(t.result, s.result);
6379	>	s = t.rights = s.nextRight;
6380	>	}
6381	>	}
6382	>	}
6383	>	}
6384	>
6385	>	@SuppressWarnings("serial") static final class MapReduceMappingsToLongTask<K,V>
6386	>	extends Traverser<K,V,Long> {
6387	>	final ObjectByObjectToLong<? super K, ? super V> transformer;
6388	>	final LongByLongToLong reducer;
6389	>	final long basis;
6390	>	long result;
6391	>	MapReduceMappingsToLongTask<K,V> rights, nextRight;
6392	>	MapReduceMappingsToLongTask
6393	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6394	>	MapReduceMappingsToLongTask<K,V> nextRight,
6395	>	ObjectByObjectToLong<? super K, ? super V> transformer,
6396	>	long basis,
6397	>	LongByLongToLong reducer) {
6398	>	super(m, p, b); this.nextRight = nextRight;
6399	>	this.transformer = transformer;
6400	>	this.basis = basis; this.reducer = reducer;
6401	>	}
6402	>	public final Long getRawResult() { return result; }
6403	>	@SuppressWarnings("unchecked") public final void compute() {
6404	>	final ObjectByObjectToLong<? super K, ? super V> transformer =
6405	>	this.transformer;
6406	>	final LongByLongToLong reducer = this.reducer;
6407	>	if (transformer == null || reducer == null)
6408	>	throw new NullPointerException();
6409	>	long r = this.basis;
6410	>	for (int b; (b = preSplit()) > 0;)
6411	>	(rights = new MapReduceMappingsToLongTask<K,V>
6412	>	(map, this, b, rights, transformer, r, reducer)).fork();
6413	>	Object v;
6414	>	while ((v = advance()) != null)
6415	>	r = reducer.apply(r, transformer.apply((K)nextKey, (V)v));
6416	>	result = r;
6417	>	CountedCompleter<?> c;
6418	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6419	>	MapReduceMappingsToLongTask<K,V>
6420	>	t = (MapReduceMappingsToLongTask<K,V>)c,
6421	>	s = t.rights;
6422	>	while (s != null) {
6423	>	t.result = reducer.apply(t.result, s.result);
6424	>	s = t.rights = s.nextRight;
6425	>	}
6426	>	}
6427	>	}
6428	>	}
6429	>
6430	>	@SuppressWarnings("serial") static final class MapReduceKeysToIntTask<K,V>
6431	>	extends Traverser<K,V,Integer> {
6432	>	final ObjectToInt<? super K> transformer;
6433	>	final IntByIntToInt reducer;
6434	>	final int basis;
6435	>	int result;
6436	>	MapReduceKeysToIntTask<K,V> rights, nextRight;
6437	>	MapReduceKeysToIntTask
6438	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6439	>	MapReduceKeysToIntTask<K,V> nextRight,
6440	>	ObjectToInt<? super K> transformer,
6441	>	int basis,
6442	>	IntByIntToInt reducer) {
6443	>	super(m, p, b); this.nextRight = nextRight;
6444	>	this.transformer = transformer;
6445	>	this.basis = basis; this.reducer = reducer;
6446	>	}
6447	>	public final Integer getRawResult() { return result; }
6448	>	@SuppressWarnings("unchecked") public final void compute() {
6449	>	final ObjectToInt<? super K> transformer =
6450	>	this.transformer;
6451	>	final IntByIntToInt reducer = this.reducer;
6452	>	if (transformer == null || reducer == null)
6453	>	throw new NullPointerException();
6454	>	int r = this.basis;
6455	>	for (int b; (b = preSplit()) > 0;)
6456	>	(rights = new MapReduceKeysToIntTask<K,V>
6457	>	(map, this, b, rights, transformer, r, reducer)).fork();
6458	>	while (advance() != null)
6459	>	r = reducer.apply(r, transformer.apply((K)nextKey));
6460	>	result = r;
6461	>	CountedCompleter<?> c;
6462	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6463	>	MapReduceKeysToIntTask<K,V>
6464	>	t = (MapReduceKeysToIntTask<K,V>)c,
6465	>	s = t.rights;
6466	>	while (s != null) {
6467	>	t.result = reducer.apply(t.result, s.result);
6468	>	s = t.rights = s.nextRight;
6469	>	}
6470	>	}
6471	>	}
6472	>	}
6473	>
6474	>	@SuppressWarnings("serial") static final class MapReduceValuesToIntTask<K,V>
6475	>	extends Traverser<K,V,Integer> {
6476	>	final ObjectToInt<? super V> transformer;
6477	>	final IntByIntToInt reducer;
6478	>	final int basis;
6479	>	int result;
6480	>	MapReduceValuesToIntTask<K,V> rights, nextRight;
6481	>	MapReduceValuesToIntTask
6482	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6483	>	MapReduceValuesToIntTask<K,V> nextRight,
6484	>	ObjectToInt<? super V> transformer,
6485	>	int basis,
6486	>	IntByIntToInt reducer) {
6487	>	super(m, p, b); this.nextRight = nextRight;
6488	>	this.transformer = transformer;
6489	>	this.basis = basis; this.reducer = reducer;
6490	>	}
6491	>	public final Integer getRawResult() { return result; }
6492	>	@SuppressWarnings("unchecked") public final void compute() {
6493	>	final ObjectToInt<? super V> transformer =
6494	>	this.transformer;
6495	>	final IntByIntToInt reducer = this.reducer;
6496	>	if (transformer == null || reducer == null)
6497	>	throw new NullPointerException();
6498	>	int r = this.basis;
6499	>	for (int b; (b = preSplit()) > 0;)
6500	>	(rights = new MapReduceValuesToIntTask<K,V>
6501	>	(map, this, b, rights, transformer, r, reducer)).fork();
6502	>	Object v;
6503	>	while ((v = advance()) != null)
6504	>	r = reducer.apply(r, transformer.apply((V)v));
6505	>	result = r;
6506	>	CountedCompleter<?> c;
6507	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6508	>	MapReduceValuesToIntTask<K,V>
6509	>	t = (MapReduceValuesToIntTask<K,V>)c,
6510	>	s = t.rights;
6511	>	while (s != null) {
6512	>	t.result = reducer.apply(t.result, s.result);
6513	>	s = t.rights = s.nextRight;
6514	>	}
6515	>	}
6516	>	}
6517	>	}
6518	>
6519	>	@SuppressWarnings("serial") static final class MapReduceEntriesToIntTask<K,V>
6520	>	extends Traverser<K,V,Integer> {
6521	>	final ObjectToInt<Map.Entry<K,V>> transformer;
6522	>	final IntByIntToInt reducer;
6523	>	final int basis;
6524	>	int result;
6525	>	MapReduceEntriesToIntTask<K,V> rights, nextRight;
6526	>	MapReduceEntriesToIntTask
6527	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6528	>	MapReduceEntriesToIntTask<K,V> nextRight,
6529	>	ObjectToInt<Map.Entry<K,V>> transformer,
6530	>	int basis,
6531	>	IntByIntToInt reducer) {
6532	>	super(m, p, b); this.nextRight = nextRight;
6533	>	this.transformer = transformer;
6534	>	this.basis = basis; this.reducer = reducer;
6535	>	}
6536	>	public final Integer getRawResult() { return result; }
6537	>	@SuppressWarnings("unchecked") public final void compute() {
6538	>	final ObjectToInt<Map.Entry<K,V>> transformer =
6539	>	this.transformer;
6540	>	final IntByIntToInt reducer = this.reducer;
6541	>	if (transformer == null || reducer == null)
6542	>	throw new NullPointerException();
6543	>	int r = this.basis;
6544	>	for (int b; (b = preSplit()) > 0;)
6545	>	(rights = new MapReduceEntriesToIntTask<K,V>
6546	>	(map, this, b, rights, transformer, r, reducer)).fork();
6547	>	Object v;
6548	>	while ((v = advance()) != null)
6549	>	r = reducer.apply(r, transformer.apply(entryFor((K)nextKey, (V)v)));
6550	>	result = r;
6551	>	CountedCompleter<?> c;
6552	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6553	>	MapReduceEntriesToIntTask<K,V>
6554	>	t = (MapReduceEntriesToIntTask<K,V>)c,
6555	>	s = t.rights;
6556	>	while (s != null) {
6557	>	t.result = reducer.apply(t.result, s.result);
6558	>	s = t.rights = s.nextRight;
6559	>	}
6560	>	}
6561	>	}
6562	>	}
6563	>
6564	>	@SuppressWarnings("serial") static final class MapReduceMappingsToIntTask<K,V>
6565	>	extends Traverser<K,V,Integer> {
6566	>	final ObjectByObjectToInt<? super K, ? super V> transformer;
6567	>	final IntByIntToInt reducer;
6568	>	final int basis;
6569	>	int result;
6570	>	MapReduceMappingsToIntTask<K,V> rights, nextRight;
6571	>	MapReduceMappingsToIntTask
6572	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6573	>	MapReduceMappingsToIntTask<K,V> nextRight,
6574	>	ObjectByObjectToInt<? super K, ? super V> transformer,
6575	>	int basis,
6576	>	IntByIntToInt reducer) {
6577	>	super(m, p, b); this.nextRight = nextRight;
6578	>	this.transformer = transformer;
6579	>	this.basis = basis; this.reducer = reducer;
6580	>	}
6581	>	public final Integer getRawResult() { return result; }
6582	>	@SuppressWarnings("unchecked") public final void compute() {
6583	>	final ObjectByObjectToInt<? super K, ? super V> transformer =
6584	>	this.transformer;
6585	>	final IntByIntToInt reducer = this.reducer;
6586	>	if (transformer == null || reducer == null)
6587	>	throw new NullPointerException();
6588	>	int r = this.basis;
6589	>	for (int b; (b = preSplit()) > 0;)
6590	>	(rights = new MapReduceMappingsToIntTask<K,V>
6591	>	(map, this, b, rights, transformer, r, reducer)).fork();
6592	>	Object v;
6593	>	while ((v = advance()) != null)
6594	>	r = reducer.apply(r, transformer.apply((K)nextKey, (V)v));
6595	>	result = r;
6596	>	CountedCompleter<?> c;
6597	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6598	>	MapReduceMappingsToIntTask<K,V>
6599	>	t = (MapReduceMappingsToIntTask<K,V>)c,
6600	>	s = t.rights;
6601	>	while (s != null) {
6602	>	t.result = reducer.apply(t.result, s.result);
6603	>	s = t.rights = s.nextRight;
6604	>	}
6605	>	}
6606		}
6607		}
6608
6609		// Unsafe mechanics
6610		private static final sun.misc.Unsafe UNSAFE;
6611		private static final long counterOffset;
6612	<	private static final long resizingOffset;
6612	>	private static final long sizeCtlOffset;
6613		private static final long ABASE;
6614		private static final int ASHIFT;
6615
#	Line 1579 | Line 6620 | public class ConcurrentHashMapV8<K, V>
6620		Class<?> k = ConcurrentHashMapV8.class;
6621		counterOffset = UNSAFE.objectFieldOffset
6622		(k.getDeclaredField("counter"));
6623	<	resizingOffset = UNSAFE.objectFieldOffset
6624	<	(k.getDeclaredField("resizing"));
6623	>	sizeCtlOffset = UNSAFE.objectFieldOffset
6624	>	(k.getDeclaredField("sizeCtl"));
6625		Class<?> sc = Node[].class;
6626		ABASE = UNSAFE.arrayBaseOffset(sc);
6627		ss = UNSAFE.arrayIndexScale(sc);
#	Line 1619 | Line 6660 | public class ConcurrentHashMapV8<K, V>
6660		}
6661		}
6662		}
1622	–
6663		}

Diff Legend

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