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 1.82 by dl, Thu Dec 13 20:34:00 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.locks.AbstractQueuedSynchronizer;
25	+	import java.util.concurrent.atomic.AtomicInteger;
26	+	import java.util.concurrent.atomic.AtomicReference;
27		import java.io.Serializable;
28
29		/**
#	Line 33 | Line 38 | import java.io.Serializable;
38		* interoperable with {@code Hashtable} in programs that rely on its
39		* thread safety but not on its synchronization details.
40		*
41	<	* <p> Retrieval operations (including {@code get}) generally do not
41	>	* <p>Retrieval operations (including {@code get}) generally do not
42		* block, so may overlap with update operations (including {@code put}
43		* and {@code remove}). Retrievals reflect the results of the most
44		* recently <em>completed</em> update operations holding upon their
45	<	* onset.  For aggregate operations such as {@code putAll} and {@code
46	<	* clear}, concurrent retrievals may reflect insertion or removal of
47	<	* only some entries.  Similarly, Iterators and Enumerations return
48	<	* elements reflecting the state of the hash table at some point at or
49	<	* since the creation of the iterator/enumeration.  They do
50	<	* <em>not</em> throw {@link ConcurrentModificationException}.
51	<	* However, iterators are designed to be used by only one thread at a
52	<	* time.  Bear in mind that the results of aggregate status methods
53	<	* including {@code size}, {@code isEmpty}, and {@code containsValue}
54	<	* are typically useful only when a map is not undergoing concurrent
55	<	* updates in other threads.  Otherwise the results of these methods
56	<	* reflect transient states that may be adequate for monitoring
57	<	* purposes, but not for program control.
58	<	*
59	<	* <p> Resizing this or any other kind of hash table is a relatively
60	<	* slow operation, so, when possible, it is a good idea to provide
61	<	* estimates of expected table sizes in constructors. Also, for
62	<	* compatibility with previous versions of this class, constructors
63	<	* may optionally specify an expected {@code concurrencyLevel} as an
64	<	* additional hint for internal sizing.
45	>	* onset. (More formally, an update operation for a given key bears a
46	>	* <em>happens-before</em> relation with any (non-null) retrieval for
47	>	* that key reporting the updated value.)  For aggregate operations
48	>	* such as {@code putAll} and {@code clear}, concurrent retrievals may
49	>	* reflect insertion or removal of only some entries.  Similarly,
50	>	* Iterators and Enumerations return elements reflecting the state of
51	>	* the hash table at some point at or since the creation of the
52	>	* iterator/enumeration.  They do <em>not</em> throw {@link
53	>	* ConcurrentModificationException}.  However, iterators are designed
54	>	* to be used by only one thread at a time.  Bear in mind that the
55	>	* results of aggregate status methods including {@code size}, {@code
56	>	* isEmpty}, and {@code containsValue} are typically useful only when
57	>	* a map is not undergoing concurrent updates in other threads.
58	>	* Otherwise the results of these methods reflect transient states
59	>	* that may be adequate for monitoring or estimation purposes, but not
60	>	* for program control.
61	>	*
62	>	* <p>The table is dynamically expanded when there are too many
63	>	* collisions (i.e., keys that have distinct hash codes but fall into
64	>	* the same slot modulo the table size), with the expected average
65	>	* effect of maintaining roughly two bins per mapping (corresponding
66	>	* to a 0.75 load factor threshold for resizing). There may be much
67	>	* variance around this average as mappings are added and removed, but
68	>	* overall, this maintains a commonly accepted time/space tradeoff for
69	>	* hash tables.  However, resizing this or any other kind of hash
70	>	* table may be a relatively slow operation. When possible, it is a
71	>	* good idea to provide a size estimate as an optional {@code
72	>	* initialCapacity} constructor argument. An additional optional
73	>	* {@code loadFactor} constructor argument provides a further means of
74	>	* customizing initial table capacity by specifying the table density
75	>	* to be used in calculating the amount of space to allocate for the
76	>	* given number of elements.  Also, for compatibility with previous
77	>	* versions of this class, constructors may optionally specify an
78	>	* expected {@code concurrencyLevel} as an additional hint for
79	>	* internal sizing.  Note that using many keys with exactly the same
80	>	* {@code hashCode()} is a sure way to slow down performance of any
81	>	* hash table.
82	>	*
83	>	* <p>A {@link Set} projection of a ConcurrentHashMapV8 may be created
84	>	* (using {@link #newKeySet()} or {@link #newKeySet(int)}), or viewed
85	>	* (using {@link #keySet(Object)} when only keys are of interest, and the
86	>	* mapped values are (perhaps transiently) not used or all take the
87	>	* same mapping value.
88	>	*
89	>	* <p>A ConcurrentHashMapV8 can be used as scalable frequency map (a
90	>	* form of histogram or multiset) by using {@link LongAdder} values
91	>	* and initializing via {@link #computeIfAbsent}. For example, to add
92	>	* a count to a {@code ConcurrentHashMapV8<String,LongAdder> freqs}, you
93	>	* can use {@code freqs.computeIfAbsent(k -> new
94	>	* LongAdder()).increment();}
95		*
96		* <p>This class and its views and iterators implement all of the
97		* <em>optional</em> methods of the {@link Map} and {@link Iterator}
98		* interfaces.
99		*
100	<	* <p> Like {@link Hashtable} but unlike {@link HashMap}, this class
100	>	* <p>Like {@link Hashtable} but unlike {@link HashMap}, this class
101		* does <em>not</em> allow {@code null} to be used as a key or value.
102		*
103	+	* <p>ConcurrentHashMapV8s support parallel operations using the {@link
104	+	* ForkJoinPool#commonPool}. (Tasks that may be used in other contexts
105	+	* are available in class {@link ForkJoinTasks}). These operations are
106	+	* designed to be safely, and often sensibly, applied even with maps
107	+	* that are being concurrently updated by other threads; for example,
108	+	* when computing a snapshot summary of the values in a shared
109	+	* registry.  There are three kinds of operation, each with four
110	+	* forms, accepting functions with Keys, Values, Entries, and (Key,
111	+	* Value) arguments and/or return values. (The first three forms are
112	+	* also available via the {@link #keySet()}, {@link #values()} and
113	+	* {@link #entrySet()} views). Because the elements of a
114	+	* ConcurrentHashMapV8 are not ordered in any particular way, and may be
115	+	* processed in different orders in different parallel executions, the
116	+	* correctness of supplied functions should not depend on any
117	+	* ordering, or on any other objects or values that may transiently
118	+	* change while computation is in progress; and except for forEach
119	+	* actions, should ideally be side-effect-free.
120	+	*
121	+	* <ul>
122	+	* <li> forEach: Perform a given action on each element.
123	+	* A variant form applies a given transformation on each element
124	+	* before performing the action.</li>
125	+	*
126	+	* <li> search: Return the first available non-null result of
127	+	* applying a given function on each element; skipping further
128	+	* search when a result is found.</li>
129	+	*
130	+	* <li> reduce: Accumulate each element.  The supplied reduction
131	+	* function cannot rely on ordering (more formally, it should be
132	+	* both associative and commutative).  There are five variants:
133	+	*
134	+	* <ul>
135	+	*
136	+	* <li> Plain reductions. (There is not a form of this method for
137	+	* (key, value) function arguments since there is no corresponding
138	+	* return type.)</li>
139	+	*
140	+	* <li> Mapped reductions that accumulate the results of a given
141	+	* function applied to each element.</li>
142	+	*
143	+	* <li> Reductions to scalar doubles, longs, and ints, using a
144	+	* given basis value.</li>
145	+	*
146	+	* </li>
147	+	* </ul>
148	+	* </ul>
149	+	*
150	+	* <p>The concurrency properties of bulk operations follow
151	+	* from those of ConcurrentHashMapV8: Any non-null result returned
152	+	* from {@code get(key)} and related access methods bears a
153	+	* happens-before relation with the associated insertion or
154	+	* update.  The result of any bulk operation reflects the
155	+	* composition of these per-element relations (but is not
156	+	* necessarily atomic with respect to the map as a whole unless it
157	+	* is somehow known to be quiescent).  Conversely, because keys
158	+	* and values in the map are never null, null serves as a reliable
159	+	* atomic indicator of the current lack of any result.  To
160	+	* maintain this property, null serves as an implicit basis for
161	+	* all non-scalar reduction operations. For the double, long, and
162	+	* int versions, the basis should be one that, when combined with
163	+	* any other value, returns that other value (more formally, it
164	+	* should be the identity element for the reduction). Most common
165	+	* reductions have these properties; for example, computing a sum
166	+	* with basis 0 or a minimum with basis MAX_VALUE.
167	+	*
168	+	* <p>Search and transformation functions provided as arguments
169	+	* should similarly return null to indicate the lack of any result
170	+	* (in which case it is not used). In the case of mapped
171	+	* reductions, this also enables transformations to serve as
172	+	* filters, returning null (or, in the case of primitive
173	+	* specializations, the identity basis) if the element should not
174	+	* be combined. You can create compound transformations and
175	+	* filterings by composing them yourself under this "null means
176	+	* there is nothing there now" rule before using them in search or
177	+	* reduce operations.
178	+	*
179	+	* <p>Methods accepting and/or returning Entry arguments maintain
180	+	* key-value associations. They may be useful for example when
181	+	* finding the key for the greatest value. Note that "plain" Entry
182	+	* arguments can be supplied using {@code new
183	+	* AbstractMap.SimpleEntry(k,v)}.
184	+	*
185	+	* <p>Bulk operations may complete abruptly, throwing an
186	+	* exception encountered in the application of a supplied
187	+	* function. Bear in mind when handling such exceptions that other
188	+	* concurrently executing functions could also have thrown
189	+	* exceptions, or would have done so if the first exception had
190	+	* not occurred.
191	+	*
192	+	* <p>Parallel speedups for bulk operations compared to sequential
193	+	* processing are common but not guaranteed.  Operations involving
194	+	* brief functions on small maps may execute more slowly than
195	+	* sequential loops if the underlying work to parallelize the
196	+	* computation is more expensive than the computation itself.
197	+	* Similarly, parallelization may not lead to much actual parallelism
198	+	* if all processors are busy performing unrelated tasks.
199	+	*
200	+	* <p>All arguments to all task methods must be non-null.
201	+	*
202	+	* <p><em>jsr166e note: During transition, this class
203	+	* uses nested functional interfaces with different names but the
204	+	* same forms as those expected for JDK8.</em>
205	+	*
206		* <p>This class is a member of the
207		* <a href="{@docRoot}/../technotes/guides/collections/index.html">
208		* Java Collections Framework</a>.
209		*
72	–	* <p><em>jsr166e note: This class is a candidate replacement for
73	–	* java.util.concurrent.ConcurrentHashMap.<em>
74	–	*
210		* @since 1.5
211		* @author Doug Lea
212		* @param <K> the type of keys maintained by this map
213		* @param <V> the type of mapped values
214		*/
215		public class ConcurrentHashMapV8<K, V>
216	<	implements ConcurrentMap<K, V>, Serializable {
216	>	implements ConcurrentMap<K, V>, Serializable {
217		private static final long serialVersionUID = 7249069246763182397L;
218
219		/**
220	<	* A function computing a mapping from the given key to a value,
221	<	* or {@code null} if there is no mapping. This is a place-holder
222	<	* for an upcoming JDK8 interface.
223	<	*/
224	<	public static interface MappingFunction<K, V> {
225	<	/**
226	<	* Returns a value for the given key, or null if there is no
227	<	* mapping. If this function throws an (unchecked) exception,
228	<	* the exception is rethrown to its caller, and no mapping is
229	<	* recorded.  Because this function is invoked within
230	<	* atomicity control, the computation should be short and
231	<	* simple. The most common usage is to construct a new object
232	<	* serving as an initial mapped value.
220	>	* A partitionable iterator. A Spliterator can be traversed
221	>	* directly, but can also be partitioned (before traversal) by
222	>	* creating another Spliterator that covers a non-overlapping
223	>	* portion of the elements, and so may be amenable to parallel
224	>	* execution.
225	>	*
226	>	* <p>This interface exports a subset of expected JDK8
227	>	* functionality.
228	>	*
229	>	* <p>Sample usage: Here is one (of the several) ways to compute
230	>	* the sum of the values held in a map using the ForkJoin
231	>	* framework. As illustrated here, Spliterators are well suited to
232	>	* designs in which a task repeatedly splits off half its work
233	>	* into forked subtasks until small enough to process directly,
234	>	* and then joins these subtasks. Variants of this style can also
235	>	* be used in completion-based designs.
236	>	*
237	>	* <pre>
238	>	* {@code ConcurrentHashMapV8<String, Long> m = ...
239	>	* // split as if have 8 * parallelism, for load balance
240	>	* int n = m.size();
241	>	* int p = aForkJoinPool.getParallelism() * 8;
242	>	* int split = (n < p)? n : p;
243	>	* long sum = aForkJoinPool.invoke(new SumValues(m.valueSpliterator(), split, null));
244	>	* // ...
245	>	* static class SumValues extends RecursiveTask<Long> {
246	>	*   final Spliterator<Long> s;
247	>	*   final int split;             // split while > 1
248	>	*   final SumValues nextJoin;    // records forked subtasks to join
249	>	*   SumValues(Spliterator<Long> s, int depth, SumValues nextJoin) {
250	>	*     this.s = s; this.depth = depth; this.nextJoin = nextJoin;
251	>	*   }
252	>	*   public Long compute() {
253	>	*     long sum = 0;
254	>	*     SumValues subtasks = null; // fork subtasks
255	>	*     for (int s = split >>> 1; s > 0; s >>>= 1)
256	>	*       (subtasks = new SumValues(s.split(), s, subtasks)).fork();
257	>	*     while (s.hasNext())        // directly process remaining elements
258	>	*       sum += s.next();
259	>	*     for (SumValues t = subtasks; t != null; t = t.nextJoin)
260	>	*       sum += t.join();         // collect subtask results
261	>	*     return sum;
262	>	*   }
263	>	* }
264	>	* }</pre>
265	>	*/
266	>	public static interface Spliterator<T> extends Iterator<T> {
267	>	/**
268	>	* Returns a Spliterator covering approximately half of the
269	>	* elements, guaranteed not to overlap with those subsequently
270	>	* returned by this Spliterator.  After invoking this method,
271	>	* the current Spliterator will <em>not</em> produce any of
272	>	* the elements of the returned Spliterator, but the two
273	>	* Spliterators together will produce all of the elements that
274	>	* would have been produced by this Spliterator had this
275	>	* method not been called. The exact number of elements
276	>	* produced by the returned Spliterator is not guaranteed, and
277	>	* may be zero (i.e., with {@code hasNext()} reporting {@code
278	>	* false}) if this Spliterator cannot be further split.
279		*
280	<	* @param key the (non-null) key
281	<	* @return a value, or null if none
280	>	* @return a Spliterator covering approximately half of the
281	>	* elements
282	>	* @throws IllegalStateException if this Spliterator has
283	>	* already commenced traversing elements
284		*/
285	<	V map(K key);
285	>	Spliterator<T> split();
286		}
287
288		/*
#	Line 108 | Line 291 | public class ConcurrentHashMapV8<K, V>
291		* The primary design goal of this hash table is to maintain
292		* concurrent readability (typically method get(), but also
293		* iterators and related methods) while minimizing update
294	<	* contention.
294	>	* contention. Secondary goals are to keep space consumption about
295	>	* the same or better than java.util.HashMap, and to support high
296	>	* initial insertion rates on an empty table by many threads.
297		*
298		* Each key-value mapping is held in a Node.  Because Node fields
299		* can contain special values, they are defined using plain Object
300		* types. Similarly in turn, all internal methods that use them
301	<	* work off Object types. All public generic-typed methods relay
302	<	* in/out of these internal methods, supplying casts as needed.
301	>	* work off Object types. And similarly, so do the internal
302	>	* methods of auxiliary iterator and view classes. This also
303	>	* allows many of the public methods to be factored into a smaller
304	>	* number of internal methods (although sadly not so for the five
305	>	* variants of put-related operations). The validation-based
306	>	* approach explained below leads to a lot of code sprawl because
307	>	* retry-control precludes factoring into smaller methods.
308		*
309		* The table is lazily initialized to a power-of-two size upon the
310	<	* first insertion.  Each bin in the table contains a (typically
311	<	* short) list of Nodes.  Table accesses require volatile/atomic
312	<	* reads, writes, and CASes.  Because there is no other way to
313	<	* arrange this without adding further indirections, we use
314	<	* intrinsics (sun.misc.Unsafe) operations.  The lists of nodes
315	<	* within bins are always accurately traversable under volatile
316	<	* reads, so long as lookups check hash code and non-nullness of
317	<	* key and value before checking key equality. (All valid hash
318	<	* codes are nonnegative. Negative values are reserved for special
319	<	* forwarding nodes; see below.)
320	<	*
321	<	* A bin may be locked during update (insert, delete, and replace)
322	<	* operations.  We do not want to waste the space required to
323	<	* associate a distinct lock object with each bin, so instead use
324	<	* the first node of a bin list itself as a lock, using builtin
325	<	* "synchronized" locks. These save space and we can live with
326	<	* only plain block-structured lock/unlock operations. Using the
327	<	* first node of a list as a lock does not by itself suffice
328	<	* though: When a node is locked, any update must first validate
329	<	* that it is still the first node, and retry if not. (Because new
330	<	* nodes are always appended to lists, once a node is first in a
331	<	* bin, it remains first until deleted or the bin becomes
332	<	* invalidated.)  However, update operations can and sometimes do
333	<	* still traverse the bin until the point of update, which helps
334	<	* reduce cache misses on retries.  This is a converse of sorts to
335	<	* the lazy locking technique described by Herlihy & Shavit. If
336	<	* there is no existing node during a put operation, then one can
337	<	* be CAS'ed in (without need for lock except in computeIfAbsent);
338	<	* the CAS serves as validation. This is on average the most
339	<	* common case for put operations -- under random hash codes, the
340	<	* distribution of nodes in bins follows a Poisson distribution
341	<	* (see http://en.wikipedia.org/wiki/Poisson_distribution) with a
342	<	* parameter of 0.5 on average under the default loadFactor of
343	<	* 0.75.  The expected number of locks covering different elements
344	<	* (i.e., bins with 2 or more nodes) is approximately 10% at
345	<	* steady state under default settings.  Lock contention
346	<	* probability for two threads accessing arbitrary distinct
347	<	* elements is, roughly, 1 / (8 * #elements).
348	<	*
349	<	* The table is resized when occupancy exceeds a threshold.  Only
350	<	* a single thread performs the resize (using field "resizing", to
351	<	* arrange exclusion), but the table otherwise remains usable for
352	<	* both reads and updates. Resizing proceeds by transferring bins,
353	<	* one by one, from the table to the next table.  Upon transfer,
354	<	* the old table bin contains only a special forwarding node (with
355	<	* negative hash code ("MOVED")) that contains the next table as
356	<	* its key. On encountering a forwarding node, access and update
357	<	* operations restart, using the new table. To ensure concurrent
358	<	* readability of traversals, transfers must proceed from the last
359	<	* bin (table.length - 1) up towards the first.  Any traversal
360	<	* starting from the first bin can then arrange to move to the new
361	<	* table for the rest of the traversal without revisiting nodes.
362	<	* This constrains bin transfers to a particular order, and so can
363	<	* block indefinitely waiting for the next lock, and other threads
364	<	* cannot help with the transfer. However, expected stalls are
365	<	* infrequent enough to not warrant the additional overhead and
366	<	* complexity of access and iteration schemes that could admit
367	<	* out-of-order or concurrent bin transfers.
368	<	*
369	<	* A similar traversal scheme (not yet implemented) can apply to
370	<	* partial traversals during partitioned aggregate operations.
371	<	* Also, read-only operations give up if ever forwarded to a null
372	<	* table, which provides support for shutdown-style clearing,
373	<	* which is also not currently implemented.
374	<	*
375	<	* The element count is maintained using a LongAdder, which avoids
376	<	* contention on updates but can encounter cache thrashing if read
377	<	* too frequently during concurrent updates. To avoid reading so
378	<	* often, resizing is normally attempted only upon adding to a bin
379	<	* already holding two or more nodes. Under the default threshold
380	<	* (0.75), and uniform hash distributions, the probability of this
381	<	* occurring at threshold is around 13%, meaning that only about 1
382	<	* in 8 puts check threshold (and after resizing, many fewer do
383	<	* so). But this approximation has high variance for small table
384	<	* sizes, so we check on any collision for sizes <= 64.  Further,
385	<	* to increase the probability that a resize occurs soon enough, we
386	<	* offset the threshold (see THRESHOLD_OFFSET) by the expected
387	<	* number of puts between checks. This is currently set to 8, in
388	<	* accord with the default load factor. In practice, this is
389	<	* rarely overridden, and in any case is close enough to other
390	<	* plausible values not to waste dynamic probability computation
391	<	* for more precision.
310	>	* first insertion.  Each bin in the table normally contains a
311	>	* list of Nodes (most often, the list has only zero or one Node).
312	>	* Table accesses require volatile/atomic reads, writes, and
313	>	* CASes.  Because there is no other way to arrange this without
314	>	* adding further indirections, we use intrinsics
315	>	* (sun.misc.Unsafe) operations.  The lists of nodes within bins
316	>	* are always accurately traversable under volatile reads, so long
317	>	* as lookups check hash code and non-nullness of value before
318	>	* checking key equality.
319	>	*
320	>	* We use the top (sign) bit of Node hash fields for control
321	>	* purposes -- it is available anyway because of addressing
322	>	* constraints.  Nodes with negative hash fields are forwarding
323	>	* nodes to either TreeBins or resized tables.  The lower 31 bits
324	>	* of each normal Node's hash field contain a transformation of
325	>	* the key's hash code.
326	>	*
327	>	* Insertion (via put or its variants) of the first node in an
328	>	* empty bin is performed by just CASing it to the bin.  This is
329	>	* by far the most common case for put operations under most
330	>	* key/hash distributions.  Other update operations (insert,
331	>	* delete, and replace) require locks.  We do not want to waste
332	>	* the space required to associate a distinct lock object with
333	>	* each bin, so instead use the first node of a bin list itself as
334	>	* a lock. Locking support for these locks relies on builtin
335	>	* "synchronized" monitors.
336	>	*
337	>	* Using the first node of a list as a lock does not by itself
338	>	* suffice though: When a node is locked, any update must first
339	>	* validate that it is still the first node after locking it, and
340	>	* retry if not. Because new nodes are always appended to lists,
341	>	* once a node is first in a bin, it remains first until deleted
342	>	* or the bin becomes invalidated (upon resizing).  However,
343	>	* operations that only conditionally update may inspect nodes
344	>	* until the point of update. This is a converse of sorts to the
345	>	* lazy locking technique described by Herlihy & Shavit.
346	>	*
347	>	* The main disadvantage of per-bin locks is that other update
348	>	* operations on other nodes in a bin list protected by the same
349	>	* lock can stall, for example when user equals() or mapping
350	>	* functions take a long time.  However, statistically, under
351	>	* random hash codes, this is not a common problem.  Ideally, the
352	>	* frequency of nodes in bins follows a Poisson distribution
353	>	* (http://en.wikipedia.org/wiki/Poisson_distribution) with a
354	>	* parameter of about 0.5 on average, given the resizing threshold
355	>	* of 0.75, although with a large variance because of resizing
356	>	* granularity. Ignoring variance, the expected occurrences of
357	>	* list size k are (exp(-0.5) * pow(0.5, k) / factorial(k)). The
358	>	* first values are:
359	>	*
360	>	* 0:    0.60653066
361	>	* 1:    0.30326533
362	>	* 2:    0.07581633
363	>	* 3:    0.01263606
364	>	* 4:    0.00157952
365	>	* 5:    0.00015795
366	>	* 6:    0.00001316
367	>	* 7:    0.00000094
368	>	* 8:    0.00000006
369	>	* more: less than 1 in ten million
370	>	*
371	>	* Lock contention probability for two threads accessing distinct
372	>	* elements is roughly 1 / (8 * #elements) under random hashes.
373	>	*
374	>	* Actual hash code distributions encountered in practice
375	>	* sometimes deviate significantly from uniform randomness.  This
376	>	* includes the case when N > (1<<30), so some keys MUST collide.
377	>	* Similarly for dumb or hostile usages in which multiple keys are
378	>	* designed to have identical hash codes. Also, although we guard
379	>	* against the worst effects of this (see method spread), sets of
380	>	* hashes may differ only in bits that do not impact their bin
381	>	* index for a given power-of-two mask.  So we use a secondary
382	>	* strategy that applies when the number of nodes in a bin exceeds
383	>	* a threshold, and at least one of the keys implements
384	>	* Comparable.  These TreeBins use a balanced tree to hold nodes
385	>	* (a specialized form of red-black trees), bounding search time
386	>	* to O(log N).  Each search step in a TreeBin is around twice as
387	>	* slow as in a regular list, but given that N cannot exceed
388	>	* (1<<64) (before running out of addresses) this bounds search
389	>	* steps, lock hold times, etc, to reasonable constants (roughly
390	>	* 100 nodes inspected per operation worst case) so long as keys
391	>	* are Comparable (which is very common -- String, Long, etc).
392	>	* TreeBin nodes (TreeNodes) also maintain the same "next"
393	>	* traversal pointers as regular nodes, so can be traversed in
394	>	* iterators in the same way.
395	>	*
396	>	* The table is resized when occupancy exceeds a percentage
397	>	* threshold (nominally, 0.75, but see below).  Any thread
398	>	* noticing an overfull bin may assist in resizing after the
399	>	* initiating thread allocates and sets up the replacement
400	>	* array. However, rather than stalling, these other threads may
401	>	* proceed with insertions etc.  The use of TreeBins shields us
402	>	* from the worst case effects of overfilling while resizes are in
403	>	* progress.  Resizing proceeds by transferring bins, one by one,
404	>	* from the table to the next table. To enable concurrency, the
405	>	* next table must be (incrementally) prefilled with place-holders
406	>	* serving as reverse forwarders to the old table.  Because we are
407	>	* using power-of-two expansion, the elements from each bin must
408	>	* either stay at same index, or move with a power of two
409	>	* offset. We eliminate unnecessary node creation by catching
410	>	* cases where old nodes can be reused because their next fields
411	>	* won't change.  On average, only about one-sixth of them need
412	>	* cloning when a table doubles. The nodes they replace will be
413	>	* garbage collectable as soon as they are no longer referenced by
414	>	* any reader thread that may be in the midst of concurrently
415	>	* traversing table.  Upon transfer, the old table bin contains
416	>	* only a special forwarding node (with hash field "MOVED") that
417	>	* contains the next table as its key. On encountering a
418	>	* forwarding node, access and update operations restart, using
419	>	* the new table.
420	>	*
421	>	* Each bin transfer requires its bin lock, which can stall
422	>	* waiting for locks while resizing. However, because other
423	>	* threads can join in and help resize rather than contend for
424	>	* locks, average aggregate waits become shorter as resizing
425	>	* progresses.  The transfer operation must also ensure that all
426	>	* accessible bins in both the old and new table are usable by any
427	>	* traversal.  This is arranged by proceeding from the last bin
428	>	* (table.length - 1) up towards the first.  Upon seeing a
429	>	* forwarding node, traversals (see class Traverser) arrange to
430	>	* move to the new table without revisiting nodes.  However, to
431	>	* ensure that no intervening nodes are skipped, bin splitting can
432	>	* only begin after the associated reverse-forwarders are in
433	>	* place.
434	>	*
435	>	* The traversal scheme also applies to partial traversals of
436	>	* ranges of bins (via an alternate Traverser constructor)
437	>	* to support partitioned aggregate operations.  Also, read-only
438	>	* operations give up if ever forwarded to a null table, which
439	>	* provides support for shutdown-style clearing, which is also not
440	>	* currently implemented.
441	>	*
442	>	* Lazy table initialization minimizes footprint until first use,
443	>	* and also avoids resizings when the first operation is from a
444	>	* putAll, constructor with map argument, or deserialization.
445	>	* These cases attempt to override the initial capacity settings,
446	>	* but harmlessly fail to take effect in cases of races.
447	>	*
448	>	* The element count is maintained using a specialization of
449	>	* LongAdder. We need to incorporate a specialization rather than
450	>	* just use a LongAdder in order to access implicit
451	>	* contention-sensing that leads to creation of multiple
452	>	* CounterCells.  The counter mechanics avoid contention on
453	>	* updates but can encounter cache thrashing if read too
454	>	* frequently during concurrent access. To avoid reading so often,
455	>	* resizing under contention is attempted only upon adding to a
456	>	* bin already holding two or more nodes. Under uniform hash
457	>	* distributions, the probability of this occurring at threshold
458	>	* is around 13%, meaning that only about 1 in 8 puts check
459	>	* threshold (and after resizing, many fewer do so). The bulk
460	>	* putAll operation further reduces contention by only committing
461	>	* count updates upon these size checks.
462	>	*
463	>	* Maintaining API and serialization compatibility with previous
464	>	* versions of this class introduces several oddities. Mainly: We
465	>	* leave untouched but unused constructor arguments refering to
466	>	* concurrencyLevel. We accept a loadFactor constructor argument,
467	>	* but apply it only to initial table capacity (which is the only
468	>	* time that we can guarantee to honor it.) We also declare an
469	>	* unused "Segment" class that is instantiated in minimal form
470	>	* only when serializing.
471		*/
472
473		/* ---------------- Constants -------------- */
474
475		/**
476	<	* The smallest allowed table capacity.  Must be a power of 2, at
477	<	* least 2.
476	>	* The largest possible table capacity.  This value must be
477	>	* exactly 1<<30 to stay within Java array allocation and indexing
478	>	* bounds for power of two table sizes, and is further required
479	>	* because the top two bits of 32bit hash fields are used for
480	>	* control purposes.
481		*/
482	<	static final int MINIMUM_CAPACITY = 2;
482	>	private static final int MAXIMUM_CAPACITY = 1 << 30;
483
484		/**
485	<	* The largest allowed table capacity.  Must be a power of 2, at
486	<	* most 1<<30.
485	>	* The default initial table capacity.  Must be a power of 2
486	>	* (i.e., at least 1) and at most MAXIMUM_CAPACITY.
487		*/
488	<	static final int MAXIMUM_CAPACITY = 1 << 30;
488	>	private static final int DEFAULT_CAPACITY = 16;
489
490		/**
491	<	* The default initial table capacity.  Must be a power of 2, at
492	<	* least MINIMUM_CAPACITY and at most MAXIMUM_CAPACITY.
491	>	* The largest possible (non-power of two) array size.
492	>	* Needed by toArray and related methods.
493		*/
494	<	static final int DEFAULT_CAPACITY = 16;
494	>	static final int MAX_ARRAY_SIZE = Integer.MAX_VALUE - 8;
495
496		/**
497	<	* The default load factor for this table, used when not otherwise
498	<	* specified in a constructor.
497	>	* The default concurrency level for this table. Unused but
498	>	* defined for compatibility with previous versions of this class.
499		*/
500	<	static final float DEFAULT_LOAD_FACTOR = 0.75f;
500	>	private static final int DEFAULT_CONCURRENCY_LEVEL = 16;
501
502		/**
503	<	* The default concurrency level for this table. Unused, but
504	<	* defined for compatibility with previous versions of this class.
503	>	* The load factor for this table. Overrides of this value in
504	>	* constructors affect only the initial table capacity.  The
505	>	* actual floating point value isn't normally used -- it is
506	>	* simpler to use expressions such as {@code n - (n >>> 2)} for
507	>	* the associated resizing threshold.
508	>	*/
509	>	private static final float LOAD_FACTOR = 0.75f;
510	>
511	>	/**
512	>	* The bin count threshold for using a tree rather than list for a
513	>	* bin.  The value reflects the approximate break-even point for
514	>	* using tree-based operations.
515	>	*/
516	>	private static final int TREE_THRESHOLD = 8;
517	>
518	>	/**
519	>	* Minimum number of rebinnings per transfer step. Ranges are
520	>	* subdivided to allow multiple resizer threads.  This value
521	>	* serves as a lower bound to avoid resizers encountering
522	>	* excessive memory contention.  The value should be at least
523	>	* DEFAULT_CAPACITY.
524	>	*/
525	>	private static final int MIN_TRANSFER_STRIDE = 16;
526	>
527	>	/*
528	>	* Encodings for Node hash fields. See above for explanation.
529	>	*/
530	>	static final int MOVED     = 0x80000000; // hash field for forwarding nodes
531	>	static final int HASH_BITS = 0x7fffffff; // usable bits of normal node hash
532	>
533	>	/** Number of CPUS, to place bounds on some sizings */
534	>	static final int NCPU = Runtime.getRuntime().availableProcessors();
535	>
536	>	/* ---------------- Counters -------------- */
537	>
538	>	// Adapted from LongAdder and Striped64.
539	>	// See their internal docs for explanation.
540	>
541	>	// A padded cell for distributing counts
542	>	static final class CounterCell {
543	>	volatile long p0, p1, p2, p3, p4, p5, p6;
544	>	volatile long value;
545	>	volatile long q0, q1, q2, q3, q4, q5, q6;
546	>	CounterCell(long x) { value = x; }
547	>	}
548	>
549	>	/**
550	>	* Holder for the thread-local hash code determining which
551	>	* CounterCell to use. The code is initialized via the
552	>	* counterHashCodeGenerator, but may be moved upon collisions.
553	>	*/
554	>	static final class CounterHashCode {
555	>	int code;
556	>	}
557	>
558	>	/**
559	>	* Generates initial value for per-thread CounterHashCodes
560		*/
561	<	static final int DEFAULT_CONCURRENCY_LEVEL = 16;
561	>	static final AtomicInteger counterHashCodeGenerator = new AtomicInteger();
562
563		/**
564	<	* The count value to offset thresholds to compensate for checking
565	<	* for resizing only when inserting into bins with two or more
239	<	* elements. See above for explanation.
564	>	* Increment for counterHashCodeGenerator. See class ThreadLocal
565	>	* for explanation.
566		*/
567	<	static final int THRESHOLD_OFFSET = 8;
567	>	static final int SEED_INCREMENT = 0x61c88647;
568
569		/**
570	<	* Special node hash value indicating to use table in node.key
245	<	* Must be negative.
570	>	* Per-thread counter hash codes. Shared across all instances
571		*/
572	<	static final int MOVED = -1;
572	>	static final ThreadLocal<CounterHashCode> threadCounterHashCode =
573	>	new ThreadLocal<CounterHashCode>();
574
575		/* ---------------- Fields -------------- */
576
577		/**
578		* The array of bins. Lazily initialized upon first insertion.
579	<	* Size is always a power of two. Accessed directly by inner
254	<	* classes.
579	>	* Size is always a power of two. Accessed directly by iterators.
580		*/
581		transient volatile Node[] table;
582
583	<	/** The counter maintaining number of elements. */
584	<	private transient final LongAdder counter;
585	<	/** Nonzero when table is being initialized or resized. Updated via CAS. */
586	<	private transient volatile int resizing;
587	<	/** The target load factor for the table. */
588	<	private transient float loadFactor;
589	<	/** The next element count value upon which to resize the table. */
590	<	private transient int threshold;
591	<	/** The initial capacity of the table. */
592	<	private transient int initCap;
583	>	/**
584	>	* The next table to use; non-null only while resizing.
585	>	*/
586	>	private transient volatile Node[] nextTable;
587	>
588	>	/**
589	>	* Base counter value, used mainly when there is no contention,
590	>	* but also as a fallback during table initialization
591	>	* races. Updated via CAS.
592	>	*/
593	>	private transient volatile long baseCount;
594	>
595	>	/**
596	>	* Table initialization and resizing control.  When negative, the
597	>	* table is being initialized or resized: -1 for initialization,
598	>	* else -(1 + the number of active resizing threads).  Otherwise,
599	>	* when table is null, holds the initial table size to use upon
600	>	* creation, or 0 for default. After initialization, holds the
601	>	* next element count value upon which to resize the table.
602	>	*/
603	>	private transient volatile int sizeCtl;
604	>
605	>	/**
606	>	* The next table index (plus one) to split while resizing.
607	>	*/
608	>	private transient volatile int transferIndex;
609	>
610	>	/**
611	>	* The least available table index to split while resizing.
612	>	*/
613	>	private transient volatile int transferOrigin;
614	>
615	>	/**
616	>	* Spinlock (locked via CAS) used when resizing and/or creating Cells.
617	>	*/
618	>	private transient volatile int counterBusy;
619	>
620	>	/**
621	>	* Table of counter cells. When non-null, size is a power of 2.
622	>	*/
623	>	private transient volatile CounterCell[] counterCells;
624
625		// views
626	<	transient Set<K> keySet;
627	<	transient Set<Map.Entry<K,V>> entrySet;
628	<	transient Collection<V> values;
626	>	private transient KeySetView<K,V> keySet;
627	>	private transient ValuesView<K,V> values;
628	>	private transient EntrySetView<K,V> entrySet;
629
630		/** For serialization compatibility. Null unless serialized; see below */
631	<	Segment<K,V>[] segments;
631	>	private Segment<K,V>[] segments;
632
633	<	/**
634	<	* Applies a supplemental hash function to a given hashCode, which
635	<	* defends against poor quality hash functions.  The result must
636	<	* be non-negative, and for reasonable performance must have good
637	<	* avalanche properties; i.e., that each bit of the argument
638	<	* affects each bit (except sign bit) of the result.
633	>	/* ---------------- Table element access -------------- */
634	>
635	>	/*
636	>	* Volatile access methods are used for table elements as well as
637	>	* elements of in-progress next table while resizing.  Uses are
638	>	* null checked by callers, and implicitly bounds-checked, relying
639	>	* on the invariants that tab arrays have non-zero size, and all
640	>	* indices are masked with (tab.length - 1) which is never
641	>	* negative and always less than length. Note that, to be correct
642	>	* wrt arbitrary concurrency errors by users, bounds checks must
643	>	* operate on local variables, which accounts for some odd-looking
644	>	* inline assignments below.
645		*/
646	<	private static final int spread(int h) {
647	<	// Apply base step of MurmurHash; see http://code.google.com/p/smhasher/
648	<	h ^= h >>> 16;
649	<	h *= 0x85ebca6b;
650	<	h ^= h >>> 13;
651	<	h *= 0xc2b2ae35;
652	<	return (h >>> 16) ^ (h & 0x7fffffff); // mask out sign bit
646	>
647	>	static final Node tabAt(Node[] tab, int i) { // used by Traverser
648	>	return (Node)U.getObjectVolatile(tab, ((long)i << ASHIFT) + ABASE);
649	>	}
650	>
651	>	private static final boolean casTabAt(Node[] tab, int i, Node c, Node v) {
652	>	return U.compareAndSwapObject(tab, ((long)i << ASHIFT) + ABASE, c, v);
653	>	}
654	>
655	>	private static final void setTabAt(Node[] tab, int i, Node v) {
656	>	U.putObjectVolatile(tab, ((long)i << ASHIFT) + ABASE, v);
657		}
658
659	+	/* ---------------- Nodes -------------- */
660	+
661		/**
662		* Key-value entry. Note that this is never exported out as a
663	<	* user-visible Map.Entry.
663	>	* user-visible Map.Entry (see MapEntry below). Nodes with a hash
664	>	* field of MOVED are special, and do not contain user keys or
665	>	* values.  Otherwise, keys are never null, and null val fields
666	>	* indicate that a node is in the process of being deleted or
667	>	* created. For purposes of read-only access, a key may be read
668	>	* before a val, but can only be used after checking val to be
669	>	* non-null.
670		*/
671	<	static final class Node {
671	>	static class Node {
672		final int hash;
673		final Object key;
674		volatile Object val;
#	Line 308 | Line 682 | public class ConcurrentHashMapV8<K, V>
682		}
683		}
684
685	<	/*
312	<	* Volatile access methods are used for table elements as well as
313	<	* elements of in-progress next table while resizing.  Uses in
314	<	* 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	<	*/
322	<
323	<	static final Node tabAt(Node[] tab, int i) { // used in HashIterator
324	<	return (Node)UNSAFE.getObjectVolatile(tab, ((long)i<<ASHIFT)+ABASE);
325	<	}
685	>	/* ---------------- TreeBins -------------- */
686
687	<	private static final boolean casTabAt(Node[] tab, int i, Node c, Node v) {
688	<	return UNSAFE.compareAndSwapObject(tab, ((long)i<<ASHIFT)+ABASE, c, v);
687	>	/**
688	>	* Nodes for use in TreeBins
689	>	*/
690	>	static final class TreeNode extends Node {
691	>	TreeNode parent;  // red-black tree links
692	>	TreeNode left;
693	>	TreeNode right;
694	>	TreeNode prev;    // needed to unlink next upon deletion
695	>	boolean red;
696	>
697	>	TreeNode(int hash, Object key, Object val, Node next, TreeNode parent) {
698	>	super(hash, key, val, next);
699	>	this.parent = parent;
700	>	}
701		}
702
703	<	private static final void setTabAt(Node[] tab, int i, Node v) {
704	<	UNSAFE.putObjectVolatile(tab, ((long)i<<ASHIFT)+ABASE, v);
705	<	}
703	>	/**
704	>	* A specialized form of red-black tree for use in bins
705	>	* whose size exceeds a threshold.
706	>	*
707	>	* TreeBins use a special form of comparison for search and
708	>	* related operations (which is the main reason we cannot use
709	>	* existing collections such as TreeMaps). TreeBins contain
710	>	* Comparable elements, but may contain others, as well as
711	>	* elements that are Comparable but not necessarily Comparable<T>
712	>	* for the same T, so we cannot invoke compareTo among them. To
713	>	* handle this, the tree is ordered primarily by hash value, then
714	>	* by getClass().getName() order, and then by Comparator order
715	>	* among elements of the same class.  On lookup at a node, if
716	>	* elements are not comparable or compare as 0, both left and
717	>	* right children may need to be searched in the case of tied hash
718	>	* values. (This corresponds to the full list search that would be
719	>	* necessary if all elements were non-Comparable and had tied
720	>	* hashes.)  The red-black balancing code is updated from
721	>	* pre-jdk-collections
722	>	* (http://gee.cs.oswego.edu/dl/classes/collections/RBCell.java)
723	>	* based in turn on Cormen, Leiserson, and Rivest "Introduction to
724	>	* Algorithms" (CLR).
725	>	*
726	>	* TreeBins also maintain a separate locking discipline than
727	>	* regular bins. Because they are forwarded via special MOVED
728	>	* nodes at bin heads (which can never change once established),
729	>	* we cannot use those nodes as locks. Instead, TreeBin
730	>	* extends AbstractQueuedSynchronizer to support a simple form of
731	>	* read-write lock. For update operations and table validation,
732	>	* the exclusive form of lock behaves in the same way as bin-head
733	>	* locks. However, lookups use shared read-lock mechanics to allow
734	>	* multiple readers in the absence of writers.  Additionally,
735	>	* these lookups do not ever block: While the lock is not
736	>	* available, they proceed along the slow traversal path (via
737	>	* next-pointers) until the lock becomes available or the list is
738	>	* exhausted, whichever comes first. (These cases are not fast,
739	>	* but maximize aggregate expected throughput.)  The AQS mechanics
740	>	* for doing this are straightforward.  The lock state is held as
741	>	* AQS getState().  Read counts are negative; the write count (1)
742	>	* is positive.  There are no signalling preferences among readers
743	>	* and writers. Since we don't need to export full Lock API, we
744	>	* just override the minimal AQS methods and use them directly.
745	>	*/
746	>	static final class TreeBin extends AbstractQueuedSynchronizer {
747	>	private static final long serialVersionUID = 2249069246763182397L;
748	>	transient TreeNode root;  // root of tree
749	>	transient TreeNode first; // head of next-pointer list
750
751	<	private static final Node relaxedTabAt(Node[] tab, int i) {
752	<	return (Node)UNSAFE.getObject(tab, ((long)i<<ASHIFT)+ABASE);
753	<	}
751	>	/* AQS overrides */
752	>	public final boolean isHeldExclusively() { return getState() > 0; }
753	>	public final boolean tryAcquire(int ignore) {
754	>	if (compareAndSetState(0, 1)) {
755	>	setExclusiveOwnerThread(Thread.currentThread());
756	>	return true;
757	>	}
758	>	return false;
759	>	}
760	>	public final boolean tryRelease(int ignore) {
761	>	setExclusiveOwnerThread(null);
762	>	setState(0);
763	>	return true;
764	>	}
765	>	public final int tryAcquireShared(int ignore) {
766	>	for (int c;;) {
767	>	if ((c = getState()) > 0)
768	>	return -1;
769	>	if (compareAndSetState(c, c -1))
770	>	return 1;
771	>	}
772	>	}
773	>	public final boolean tryReleaseShared(int ignore) {
774	>	int c;
775	>	do {} while (!compareAndSetState(c = getState(), c + 1));
776	>	return c == -1;
777	>	}
778	>
779	>	/** From CLR */
780	>	private void rotateLeft(TreeNode p) {
781	>	if (p != null) {
782	>	TreeNode r = p.right, pp, rl;
783	>	if ((rl = p.right = r.left) != null)
784	>	rl.parent = p;
785	>	if ((pp = r.parent = p.parent) == null)
786	>	root = r;
787	>	else if (pp.left == p)
788	>	pp.left = r;
789	>	else
790	>	pp.right = r;
791	>	r.left = p;
792	>	p.parent = r;
793	>	}
794	>	}
795
796	<	private static final void relaxedSetTabAt(Node[] tab, int i, Node v) {
797	<	UNSAFE.putObject(tab, ((long)i<<ASHIFT)+ABASE, v);
798	<	}
796	>	/** From CLR */
797	>	private void rotateRight(TreeNode p) {
798	>	if (p != null) {
799	>	TreeNode l = p.left, pp, lr;
800	>	if ((lr = p.left = l.right) != null)
801	>	lr.parent = p;
802	>	if ((pp = l.parent = p.parent) == null)
803	>	root = l;
804	>	else if (pp.right == p)
805	>	pp.right = l;
806	>	else
807	>	pp.left = l;
808	>	l.right = p;
809	>	p.parent = l;
810	>	}
811	>	}
812
813	<	/* ---------------- Access and update operations -------------- */
813	>	/**
814	>	* Returns the TreeNode (or null if not found) for the given key
815	>	* starting at given root.
816	>	*/
817	>	@SuppressWarnings("unchecked") final TreeNode getTreeNode
818	>	(int h, Object k, TreeNode p) {
819	>	Class<?> c = k.getClass();
820	>	while (p != null) {
821	>	int dir, ph;  Object pk; Class<?> pc;
822	>	if ((ph = p.hash) == h) {
823	>	if ((pk = p.key) == k || k.equals(pk))
824	>	return p;
825	>	if (c != (pc = pk.getClass()) ||
826	>	!(k instanceof Comparable) ||
827	>	(dir = ((Comparable)k).compareTo((Comparable)pk)) == 0) {
828	>	if ((dir = (c == pc) ? 0 :
829	>	c.getName().compareTo(pc.getName())) == 0) {
830	>	TreeNode r = null, pl, pr; // check both sides
831	>	if ((pr = p.right) != null && h >= pr.hash &&
832	>	(r = getTreeNode(h, k, pr)) != null)
833	>	return r;
834	>	else if ((pl = p.left) != null && h <= pl.hash)
835	>	dir = -1;
836	>	else // nothing there
837	>	return null;
838	>	}
839	>	}
840	>	}
841	>	else
842	>	dir = (h < ph) ? -1 : 1;
843	>	p = (dir > 0) ? p.right : p.left;
844	>	}
845	>	return null;
846	>	}
847
848	<	/** Implementation for get and containsKey */
849	<	private final Object internalGet(Object k) {
850	<	int h = spread(k.hashCode());
851	<	Node[] tab = table;
852	<	retry: while (tab != null) {
853	<	Node e = tabAt(tab, (tab.length - 1) & h);
854	<	while (e != null) {
855	<	int eh = e.hash;
856	<	if (eh == h) {
857	<	Object ek = e.key, ev = e.val;
858	<	if (ev != null && ek != null && (k == ek || k.equals(ek)))
859	<	return ev;
860	<	}
861	<	else if (eh < 0) { // bin was moved during resize
862	<	tab = (Node[])e.key;
863	<	continue retry;
848	>	/**
849	>	* Wrapper for getTreeNode used by CHM.get. Tries to obtain
850	>	* read-lock to call getTreeNode, but during failure to get
851	>	* lock, searches along next links.
852	>	*/
853	>	final Object getValue(int h, Object k) {
854	>	Node r = null;
855	>	int c = getState(); // Must read lock state first
856	>	for (Node e = first; e != null; e = e.next) {
857	>	if (c <= 0 && compareAndSetState(c, c - 1)) {
858	>	try {
859	>	r = getTreeNode(h, k, root);
860	>	} finally {
861	>	releaseShared(0);
862	>	}
863	>	break;
864		}
865	<	e = e.next;
865	>	else if (e.hash == h && k.equals(e.key)) {
866	>	r = e;
867	>	break;
868	>	}
869	>	else
870	>	c = getState();
871		}
872	<	break;
872	>	return r == null ? null : r.val;
873		}
366	–	return null;
367	–	}
874
875	+	/**
876	+	* Finds or adds a node.
877	+	* @return null if added
878	+	*/
879	+	@SuppressWarnings("unchecked") final TreeNode putTreeNode
880	+	(int h, Object k, Object v) {
881	+	Class<?> c = k.getClass();
882	+	TreeNode pp = root, p = null;
883	+	int dir = 0;
884	+	while (pp != null) { // find existing node or leaf to insert at
885	+	int ph;  Object pk; Class<?> pc;
886	+	p = pp;
887	+	if ((ph = p.hash) == h) {
888	+	if ((pk = p.key) == k || k.equals(pk))
889	+	return p;
890	+	if (c != (pc = pk.getClass()) ||
891	+	!(k instanceof Comparable) ||
892	+	(dir = ((Comparable)k).compareTo((Comparable)pk)) == 0) {
893	+	TreeNode s = null, r = null, pr;
894	+	if ((dir = (c == pc) ? 0 :
895	+	c.getName().compareTo(pc.getName())) == 0) {
896	+	if ((pr = p.right) != null && h >= pr.hash &&
897	+	(r = getTreeNode(h, k, pr)) != null)
898	+	return r;
899	+	else // continue left
900	+	dir = -1;
901	+	}
902	+	else if ((pr = p.right) != null && h >= pr.hash)
903	+	s = pr;
904	+	if (s != null && (r = getTreeNode(h, k, s)) != null)
905	+	return r;
906	+	}
907	+	}
908	+	else
909	+	dir = (h < ph) ? -1 : 1;
910	+	pp = (dir > 0) ? p.right : p.left;
911	+	}
912
913	<	/** Implementation for put and putIfAbsent */
914	<	private final Object internalPut(Object k, Object v, boolean replace) {
915	<	int h = spread(k.hashCode());
916	<	Object oldVal = null;  // the previous value or null if none
917	<	Node[] tab = table;
918	<	for (;;) {
919	<	Node e; int i;
920	<	if (tab == null)
921	<	tab = grow(0);
922	<	else if ((e = tabAt(tab, i = (tab.length - 1) & h)) == null) {
923	<	if (casTabAt(tab, i, null, new Node(h, k, v, null)))
924	<	break;
913	>	TreeNode f = first;
914	>	TreeNode x = first = new TreeNode(h, k, v, f, p);
915	>	if (p == null)
916	>	root = x;
917	>	else { // attach and rebalance; adapted from CLR
918	>	TreeNode xp, xpp;
919	>	if (f != null)
920	>	f.prev = x;
921	>	if (dir <= 0)
922	>	p.left = x;
923	>	else
924	>	p.right = x;
925	>	x.red = true;
926	>	while (x != null && (xp = x.parent) != null && xp.red &&
927	>	(xpp = xp.parent) != null) {
928	>	TreeNode xppl = xpp.left;
929	>	if (xp == xppl) {
930	>	TreeNode y = xpp.right;
931	>	if (y != null && y.red) {
932	>	y.red = false;
933	>	xp.red = false;
934	>	xpp.red = true;
935	>	x = xpp;
936	>	}
937	>	else {
938	>	if (x == xp.right) {
939	>	rotateLeft(x = xp);
940	>	xpp = (xp = x.parent) == null ? null : xp.parent;
941	>	}
942	>	if (xp != null) {
943	>	xp.red = false;
944	>	if (xpp != null) {
945	>	xpp.red = true;
946	>	rotateRight(xpp);
947	>	}
948	>	}
949	>	}
950	>	}
951	>	else {
952	>	TreeNode y = xppl;
953	>	if (y != null && y.red) {
954	>	y.red = false;
955	>	xp.red = false;
956	>	xpp.red = true;
957	>	x = xpp;
958	>	}
959	>	else {
960	>	if (x == xp.left) {
961	>	rotateRight(x = xp);
962	>	xpp = (xp = x.parent) == null ? null : xp.parent;
963	>	}
964	>	if (xp != null) {
965	>	xp.red = false;
966	>	if (xpp != null) {
967	>	xpp.red = true;
968	>	rotateLeft(xpp);
969	>	}
970	>	}
971	>	}
972	>	}
973	>	}
974	>	TreeNode r = root;
975	>	if (r != null && r.red)
976	>	r.red = false;
977	>	}
978	>	return null;
979	>	}
980	>
981	>	/**
982	>	* Removes the given node, that must be present before this
983	>	* call.  This is messier than typical red-black deletion code
984	>	* because we cannot swap the contents of an interior node
985	>	* with a leaf successor that is pinned by "next" pointers
986	>	* that are accessible independently of lock. So instead we
987	>	* swap the tree linkages.
988	>	*/
989	>	final void deleteTreeNode(TreeNode p) {
990	>	TreeNode next = (TreeNode)p.next; // unlink traversal pointers
991	>	TreeNode pred = p.prev;
992	>	if (pred == null)
993	>	first = next;
994	>	else
995	>	pred.next = next;
996	>	if (next != null)
997	>	next.prev = pred;
998	>	TreeNode replacement;
999	>	TreeNode pl = p.left;
1000	>	TreeNode pr = p.right;
1001	>	if (pl != null && pr != null) {
1002	>	TreeNode s = pr, sl;
1003	>	while ((sl = s.left) != null) // find successor
1004	>	s = sl;
1005	>	boolean c = s.red; s.red = p.red; p.red = c; // swap colors
1006	>	TreeNode sr = s.right;
1007	>	TreeNode pp = p.parent;
1008	>	if (s == pr) { // p was s's direct parent
1009	>	p.parent = s;
1010	>	s.right = p;
1011	>	}
1012	>	else {
1013	>	TreeNode sp = s.parent;
1014	>	if ((p.parent = sp) != null) {
1015	>	if (s == sp.left)
1016	>	sp.left = p;
1017	>	else
1018	>	sp.right = p;
1019	>	}
1020	>	if ((s.right = pr) != null)
1021	>	pr.parent = s;
1022	>	}
1023	>	p.left = null;
1024	>	if ((p.right = sr) != null)
1025	>	sr.parent = p;
1026	>	if ((s.left = pl) != null)
1027	>	pl.parent = s;
1028	>	if ((s.parent = pp) == null)
1029	>	root = s;
1030	>	else if (p == pp.left)
1031	>	pp.left = s;
1032	>	else
1033	>	pp.right = s;
1034	>	replacement = sr;
1035	>	}
1036	>	else
1037	>	replacement = (pl != null) ? pl : pr;
1038	>	TreeNode pp = p.parent;
1039	>	if (replacement == null) {
1040	>	if (pp == null) {
1041	>	root = null;
1042	>	return;
1043	>	}
1044	>	replacement = p;
1045		}
383	–	else if (e.hash < 0)
384	–	tab = (Node[])e.key;
1046		else {
1047	<	boolean validated = false;
1048	<	boolean checkSize = false;
1049	<	synchronized (e) {
1050	<	if (tabAt(tab, i) == e) {
1051	<	validated = true;
1052	<	for (Node first = e;;) {
1053	<	Object ek, ev;
1054	<	if (e.hash == h &&
1055	<	(ek = e.key) != null &&
1056	<	(ev = e.val) != null &&
1057	<	(k == ek || k.equals(ek))) {
1058	<	oldVal = ev;
1059	<	if (replace)
1060	<	e.val = v;
1061	<	break;
1047	>	replacement.parent = pp;
1048	>	if (pp == null)
1049	>	root = replacement;
1050	>	else if (p == pp.left)
1051	>	pp.left = replacement;
1052	>	else
1053	>	pp.right = replacement;
1054	>	p.left = p.right = p.parent = null;
1055	>	}
1056	>	if (!p.red) { // rebalance, from CLR
1057	>	TreeNode x = replacement;
1058	>	while (x != null) {
1059	>	TreeNode xp, xpl;
1060	>	if (x.red || (xp = x.parent) == null) {
1061	>	x.red = false;
1062	>	break;
1063	>	}
1064	>	if (x == (xpl = xp.left)) {
1065	>	TreeNode sib = xp.right;
1066	>	if (sib != null && sib.red) {
1067	>	sib.red = false;
1068	>	xp.red = true;
1069	>	rotateLeft(xp);
1070	>	sib = (xp = x.parent) == null ? null : xp.right;
1071	>	}
1072	>	if (sib == null)
1073	>	x = xp;
1074	>	else {
1075	>	TreeNode sl = sib.left, sr = sib.right;
1076	>	if ((sr == null || !sr.red) &&
1077	>	(sl == null || !sl.red)) {
1078	>	sib.red = true;
1079	>	x = xp;
1080		}
1081	<	Node last = e;
1082	<	if ((e = e.next) == null) {
1083	<	last.next = new Node(h, k, v, null);
1084	<	if (last != first || tab.length <= 64)
1085	<	checkSize = true;
1086	<	break;
1081	>	else {
1082	>	if (sr == null || !sr.red) {
1083	>	if (sl != null)
1084	>	sl.red = false;
1085	>	sib.red = true;
1086	>	rotateRight(sib);
1087	>	sib = (xp = x.parent) == null ?
1088	>	null : xp.right;
1089	>	}
1090	>	if (sib != null) {
1091	>	sib.red = (xp == null) ? false : xp.red;
1092	>	if ((sr = sib.right) != null)
1093	>	sr.red = false;
1094	>	}
1095	>	if (xp != null) {
1096	>	xp.red = false;
1097	>	rotateLeft(xp);
1098	>	}
1099	>	x = root;
1100	>	}
1101	>	}
1102	>	}
1103	>	else { // symmetric
1104	>	TreeNode sib = xpl;
1105	>	if (sib != null && sib.red) {
1106	>	sib.red = false;
1107	>	xp.red = true;
1108	>	rotateRight(xp);
1109	>	sib = (xp = x.parent) == null ? null : xp.left;
1110	>	}
1111	>	if (sib == null)
1112	>	x = xp;
1113	>	else {
1114	>	TreeNode sl = sib.left, sr = sib.right;
1115	>	if ((sl == null || !sl.red) &&
1116	>	(sr == null || !sr.red)) {
1117	>	sib.red = true;
1118	>	x = xp;
1119	>	}
1120	>	else {
1121	>	if (sl == null || !sl.red) {
1122	>	if (sr != null)
1123	>	sr.red = false;
1124	>	sib.red = true;
1125	>	rotateLeft(sib);
1126	>	sib = (xp = x.parent) == null ?
1127	>	null : xp.left;
1128	>	}
1129	>	if (sib != null) {
1130	>	sib.red = (xp == null) ? false : xp.red;
1131	>	if ((sl = sib.left) != null)
1132	>	sl.red = false;
1133	>	}
1134	>	if (xp != null) {
1135	>	xp.red = false;
1136	>	rotateRight(xp);
1137	>	}
1138	>	x = root;
1139		}
1140		}
1141		}
1142		}
1143	<	if (validated) {
1144	<	if (checkSize && tab.length < MAXIMUM_CAPACITY &&
1145	<	resizing == 0 && counter.sum() >= threshold)
1146	<	grow(0);
1147	<	break;
1143	>	}
1144	>	if (p == replacement && (pp = p.parent) != null) {
1145	>	if (p == pp.left) // detach pointers
1146	>	pp.left = null;
1147	>	else if (p == pp.right)
1148	>	pp.right = null;
1149	>	p.parent = null;
1150	>	}
1151	>	}
1152	>	}
1153	>
1154	>	/* ---------------- Collision reduction methods -------------- */
1155	>
1156	>	/**
1157	>	* Spreads higher bits to lower, and also forces top bit to 0.
1158	>	* Because the table uses power-of-two masking, sets of hashes
1159	>	* that vary only in bits above the current mask will always
1160	>	* collide. (Among known examples are sets of Float keys holding
1161	>	* consecutive whole numbers in small tables.)  To counter this,
1162	>	* we apply a transform that spreads the impact of higher bits
1163	>	* downward. There is a tradeoff between speed, utility, and
1164	>	* quality of bit-spreading. Because many common sets of hashes
1165	>	* are already reasonably distributed across bits (so don't benefit
1166	>	* from spreading), and because we use trees to handle large sets
1167	>	* of collisions in bins, we don't need excessively high quality.
1168	>	*/
1169	>	private static final int spread(int h) {
1170	>	h ^= (h >>> 18) ^ (h >>> 12);
1171	>	return (h ^ (h >>> 10)) & HASH_BITS;
1172	>	}
1173	>
1174	>	/**
1175	>	* Replaces a list bin with a tree bin if key is comparable.  Call
1176	>	* only when locked.
1177	>	*/
1178	>	private final void replaceWithTreeBin(Node[] tab, int index, Object key) {
1179	>	if (key instanceof Comparable) {
1180	>	TreeBin t = new TreeBin();
1181	>	for (Node e = tabAt(tab, index); e != null; e = e.next)
1182	>	t.putTreeNode(e.hash, e.key, e.val);
1183	>	setTabAt(tab, index, new Node(MOVED, t, null, null));
1184	>	}
1185	>	}
1186	>
1187	>	/* ---------------- Internal access and update methods -------------- */
1188	>
1189	>	/** Implementation for get and containsKey */
1190	>	@SuppressWarnings("unchecked") private final V internalGet(Object k) {
1191	>	int h = spread(k.hashCode());
1192	>	retry: for (Node[] tab = table; tab != null;) {
1193	>	Node e; Object ek, ev; int eh;      // locals to read fields once
1194	>	for (e = tabAt(tab, (tab.length - 1) & h); e != null; e = e.next) {
1195	>	if ((eh = e.hash) < 0) {
1196	>	if ((ek = e.key) instanceof TreeBin)  // search TreeBin
1197	>	return (V)((TreeBin)ek).getValue(h, k);
1198	>	else {                        // restart with new table
1199	>	tab = (Node[])ek;
1200	>	continue retry;
1201	>	}
1202		}
1203	+	else if (eh == h && (ev = e.val) != null &&
1204	+	((ek = e.key) == k || k.equals(ek)))
1205	+	return (V)ev;
1206		}
1207	+	break;
1208		}
1209	<	if (oldVal == null)
421	<	counter.increment();
422	<	return oldVal;
1209	>	return null;
1210		}
1211
1212		/**
#	Line 427 | Line 1214 | public class ConcurrentHashMapV8<K, V>
1214		* Replaces node value with v, conditional upon match of cv if
1215		* non-null.  If resulting value is null, delete.
1216		*/
1217	<	private final Object internalReplace(Object k, Object v, Object cv) {
1217	>	@SuppressWarnings("unchecked") private final V internalReplace
1218	>	(Object k, V v, Object cv) {
1219		int h = spread(k.hashCode());
1220		Object oldVal = null;
1221	<	Node e; int i;
1222	<	Node[] tab = table;
1223	<	while (tab != null &&
1224	<	(e = tabAt(tab, i = (tab.length - 1) & h)) != null) {
1225	<	if (e.hash < 0)
1226	<	tab = (Node[])e.key;
1221	>	for (Node[] tab = table;;) {
1222	>	Node f; int i, fh; Object fk;
1223	>	if (tab == null ||
1224	>	(f = tabAt(tab, i = (tab.length - 1) & h)) == null)
1225	>	break;
1226	>	else if ((fh = f.hash) < 0) {
1227	>	if ((fk = f.key) instanceof TreeBin) {
1228	>	TreeBin t = (TreeBin)fk;
1229	>	boolean validated = false;
1230	>	boolean deleted = false;
1231	>	t.acquire(0);
1232	>	try {
1233	>	if (tabAt(tab, i) == f) {
1234	>	validated = true;
1235	>	TreeNode p = t.getTreeNode(h, k, t.root);
1236	>	if (p != null) {
1237	>	Object pv = p.val;
1238	>	if (cv == null || cv == pv || cv.equals(pv)) {
1239	>	oldVal = pv;
1240	>	if ((p.val = v) == null) {
1241	>	deleted = true;
1242	>	t.deleteTreeNode(p);
1243	>	}
1244	>	}
1245	>	}
1246	>	}
1247	>	} finally {
1248	>	t.release(0);
1249	>	}
1250	>	if (validated) {
1251	>	if (deleted)
1252	>	addCount(-1L, -1);
1253	>	break;
1254	>	}
1255	>	}
1256	>	else
1257	>	tab = (Node[])fk;
1258	>	}
1259	>	else if (fh != h && f.next == null) // precheck
1260	>	break;                          // rules out possible existence
1261		else {
1262		boolean validated = false;
1263		boolean deleted = false;
1264	<	synchronized (e) {
1265	<	if (tabAt(tab, i) == e) {
1264	>	synchronized(f) {
1265	>	if (tabAt(tab, i) == f) {
1266		validated = true;
1267	<	Node pred = null;
446	<	do {
1267	>	for (Node e = f, pred = null;;) {
1268		Object ek, ev;
1269		if (e.hash == h &&
1270	<	(ek = e.key) != null &&
1271	<	(ev = e.val) != null &&
451	<	(k == ek || k.equals(ek))) {
1270	>	((ev = e.val) != null) &&
1271	>	((ek = e.key) == k || k.equals(ek))) {
1272		if (cv == null || cv == ev || cv.equals(ev)) {
1273		oldVal = ev;
1274		if ((e.val = v) == null) {
#	Line 463 | Line 1283 | public class ConcurrentHashMapV8<K, V>
1283		break;
1284		}
1285		pred = e;
1286	<	} while ((e = e.next) != null);
1286	>	if ((e = e.next) == null)
1287	>	break;
1288	>	}
1289		}
1290		}
1291		if (validated) {
1292		if (deleted)
1293	<	counter.decrement();
1293	>	addCount(-1L, -1);
1294		break;
1295		}
1296		}
1297		}
1298	<	return oldVal;
1298	>	return (V)oldVal;
1299		}
1300
1301	<	/** Implementation for computeIfAbsent and compute */
1302	<	@SuppressWarnings("unchecked")
1303	<	private final V internalCompute(K k,
1304	<	MappingFunction<? super K, ? extends V> f,
1305	<	boolean replace) {
1301	>	/*
1302	>	* Internal versions of insertion methods
1303	>	* All have the same basic structure as the first (internalPut):
1304	>	*  1. If table uninitialized, create
1305	>	*  2. If bin empty, try to CAS new node
1306	>	*  3. If bin stale, use new table
1307	>	*  4. if bin converted to TreeBin, validate and relay to TreeBin methods
1308	>	*  5. Lock and validate; if valid, scan and add or update
1309	>	*
1310	>	* The putAll method differs mainly in attempting to pre-allocate
1311	>	* enough table space, and also more lazily performs count updates
1312	>	* and checks.
1313	>	*
1314	>	* Most of the function-accepting methods can't be factored nicely
1315	>	* because they require different functional forms, so instead
1316	>	* sprawl out similar mechanics.
1317	>	*/
1318	>
1319	>	/** Implementation for put and putIfAbsent */
1320	>	@SuppressWarnings("unchecked") private final V internalPut
1321	>	(K k, V v, boolean onlyIfAbsent) {
1322	>	if (k == null || v == null) throw new NullPointerException();
1323		int h = spread(k.hashCode());
1324	<	V val = null;
1325	<	boolean added = false;
1326	<	Node[] tab = table;
1327	<	for (;;) {
1328	<	Node e; int i;
1324	>	int len = 0;
1325	>	for (Node[] tab = table;;) {
1326	>	int i, fh; Node f; Object fk, fv;
1327	>	if (tab == null)
1328	>	tab = initTable();
1329	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1330	>	if (casTabAt(tab, i, null, new Node(h, k, v, null)))
1331	>	break;                   // no lock when adding to empty bin
1332	>	}
1333	>	else if ((fh = f.hash) < 0) {
1334	>	if ((fk = f.key) instanceof TreeBin) {
1335	>	TreeBin t = (TreeBin)fk;
1336	>	Object oldVal = null;
1337	>	t.acquire(0);
1338	>	try {
1339	>	if (tabAt(tab, i) == f) {
1340	>	len = 2;
1341	>	TreeNode p = t.putTreeNode(h, k, v);
1342	>	if (p != null) {
1343	>	oldVal = p.val;
1344	>	if (!onlyIfAbsent)
1345	>	p.val = v;
1346	>	}
1347	>	}
1348	>	} finally {
1349	>	t.release(0);
1350	>	}
1351	>	if (len != 0) {
1352	>	if (oldVal != null)
1353	>	return (V)oldVal;
1354	>	break;
1355	>	}
1356	>	}
1357	>	else
1358	>	tab = (Node[])fk;
1359	>	}
1360	>	else if (onlyIfAbsent && fh == h && (fv = f.val) != null &&
1361	>	((fk = f.key) == k || k.equals(fk))) // peek while nearby
1362	>	return (V)fv;
1363	>	else {
1364	>	Object oldVal = null;
1365	>	synchronized(f) {
1366	>	if (tabAt(tab, i) == f) {
1367	>	len = 1;
1368	>	for (Node e = f;; ++len) {
1369	>	Object ek, ev;
1370	>	if (e.hash == h &&
1371	>	(ev = e.val) != null &&
1372	>	((ek = e.key) == k || k.equals(ek))) {
1373	>	oldVal = ev;
1374	>	if (!onlyIfAbsent)
1375	>	e.val = v;
1376	>	break;
1377	>	}
1378	>	Node last = e;
1379	>	if ((e = e.next) == null) {
1380	>	last.next = new Node(h, k, v, null);
1381	>	if (len >= TREE_THRESHOLD)
1382	>	replaceWithTreeBin(tab, i, k);
1383	>	break;
1384	>	}
1385	>	}
1386	>	}
1387	>	}
1388	>	if (len != 0) {
1389	>	if (oldVal != null)
1390	>	return (V)oldVal;
1391	>	break;
1392	>	}
1393	>	}
1394	>	}
1395	>	addCount(1L, len);
1396	>	return null;
1397	>	}
1398	>
1399	>	/** Implementation for computeIfAbsent */
1400	>	@SuppressWarnings("unchecked") private final V internalComputeIfAbsent
1401	>	(K k, Fun<? super K, ?> mf) {
1402	>	if (k == null || mf == null)
1403	>	throw new NullPointerException();
1404	>	int h = spread(k.hashCode());
1405	>	Object val = null;
1406	>	int len = 0;
1407	>	for (Node[] tab = table;;) {
1408	>	Node f; int i; Object fk;
1409		if (tab == null)
1410	<	tab = grow(0);
1411	<	else if ((e = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1410	>	tab = initTable();
1411	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1412		Node node = new Node(h, k, null, null);
1413	<	boolean validated = false;
495	<	synchronized (node) {
1413	>	synchronized(node) {
1414		if (casTabAt(tab, i, null, node)) {
1415	<	validated = true;
1415	>	len = 1;
1416		try {
1417	<	val = f.map(k);
500	<	if (val != null) {
1417	>	if ((val = mf.apply(k)) != null)
1418		node.val = val;
502	–	added = true;
503	–	}
1419		} finally {
1420	<	if (!added)
1420	>	if (val == null)
1421		setTabAt(tab, i, null);
1422		}
1423		}
1424		}
1425	<	if (validated)
1425	>	if (len != 0)
1426		break;
1427		}
1428	<	else if (e.hash < 0)
1429	<	tab = (Node[])e.key;
1430	<	else if (Thread.holdsLock(e))
1431	<	throw new IllegalStateException("Recursive map computation");
1428	>	else if (f.hash < 0) {
1429	>	if ((fk = f.key) instanceof TreeBin) {
1430	>	TreeBin t = (TreeBin)fk;
1431	>	boolean added = false;
1432	>	t.acquire(0);
1433	>	try {
1434	>	if (tabAt(tab, i) == f) {
1435	>	len = 1;
1436	>	TreeNode p = t.getTreeNode(h, k, t.root);
1437	>	if (p != null)
1438	>	val = p.val;
1439	>	else if ((val = mf.apply(k)) != null) {
1440	>	added = true;
1441	>	len = 2;
1442	>	t.putTreeNode(h, k, val);
1443	>	}
1444	>	}
1445	>	} finally {
1446	>	t.release(0);
1447	>	}
1448	>	if (len != 0) {
1449	>	if (!added)
1450	>	return (V)val;
1451	>	break;
1452	>	}
1453	>	}
1454	>	else
1455	>	tab = (Node[])fk;
1456	>	}
1457		else {
1458	<	boolean validated = false;
1459	<	boolean checkSize = false;
1460	<	synchronized (e) {
1461	<	if (tabAt(tab, i) == e) {
1462	<	validated = true;
1463	<	for (Node first = e;;) {
1464	<	Object ek, ev, fv;
1458	>	for (Node e = f; e != null; e = e.next) { // prescan
1459	>	Object ek, ev;
1460	>	if (e.hash == h && (ev = e.val) != null &&
1461	>	((ek = e.key) == k || k.equals(ek)))
1462	>	return (V)ev;
1463	>	}
1464	>	boolean added = false;
1465	>	synchronized(f) {
1466	>	if (tabAt(tab, i) == f) {
1467	>	len = 1;
1468	>	for (Node e = f;; ++len) {
1469	>	Object ek, ev;
1470		if (e.hash == h &&
526	–	(ek = e.key) != null &&
1471		(ev = e.val) != null &&
1472	<	(k == ek || k.equals(ek))) {
1473	<	if (replace && (fv = f.map(k)) != null)
530	<	ev = e.val = fv;
531	<	val = (V)ev;
1472	>	((ek = e.key) == k || k.equals(ek))) {
1473	>	val = ev;
1474		break;
1475		}
1476		Node last = e;
1477		if ((e = e.next) == null) {
1478	<	if ((val = f.map(k)) != null) {
537	<	last.next = new Node(h, k, val, null);
1478	>	if ((val = mf.apply(k)) != null) {
1479		added = true;
1480	<	if (last != first || tab.length <= 64)
1481	<	checkSize = true;
1480	>	last.next = new Node(h, k, val, null);
1481	>	if (len >= TREE_THRESHOLD)
1482	>	replaceWithTreeBin(tab, i, k);
1483		}
1484		break;
1485		}
1486		}
1487		}
1488		}
1489	<	if (validated) {
1490	<	if (checkSize && tab.length < MAXIMUM_CAPACITY &&
1491	<	resizing == 0 && counter.sum() >= threshold)
550	<	grow(0);
1489	>	if (len != 0) {
1490	>	if (!added)
1491	>	return (V)val;
1492		break;
1493		}
1494		}
1495		}
1496	<	if (added)
1497	<	counter.increment();
1498	<	return val;
1496	>	if (val != null)
1497	>	addCount(1L, len);
1498	>	return (V)val;
1499		}
1500
1501	<	/*
1502	<	* Reclassifies nodes in each bin to new table.  Because we are
1503	<	* using power-of-two expansion, the elements from each bin must
1504	<	* either stay at same index, or move with a power of two
1505	<	* offset. We eliminate unnecessary node creation by catching
1506	<	* cases where old nodes can be reused because their next fields
1507	<	* won't change.  Statistically, at the default threshold, only
1508	<	* about one-sixth of them need cloning when a table doubles. The
1509	<	* nodes they replace will be garbage collectable as soon as they
1510	<	* are no longer referenced by any reader thread that may be in
1511	<	* the midst of concurrently traversing table.
1512	<	*
1513	<	* Transfers are done from the bottom up to preserve iterator
1514	<	* traversability. On each step, the old bin is locked,
1515	<	* moved/copied, and then replaced with a forwarding node.
1516	<	*/
1517	<	private static final void transfer(Node[] tab, Node[] nextTab) {
1518	<	int n = tab.length;
1519	<	int mask = nextTab.length - 1;
1520	<	Node fwd = new Node(MOVED, nextTab, null, null);
1521	<	for (int i = n - 1; i >= 0; --i) {
1522	<	for (Node e;;) {
1523	<	if ((e = tabAt(tab, i)) == null) {
1524	<	if (casTabAt(tab, i, e, fwd))
1525	<	break;
1501	>	/** Implementation for compute */
1502	>	@SuppressWarnings("unchecked") private final V internalCompute
1503	>	(K k, boolean onlyIfPresent,
1504	>	BiFun<? super K, ? super V, ? extends V> mf) {
1505	>	if (k == null || mf == null)
1506	>	throw new NullPointerException();
1507	>	int h = spread(k.hashCode());
1508	>	Object val = null;
1509	>	int delta = 0;
1510	>	int len = 0;
1511	>	for (Node[] tab = table;;) {
1512	>	Node f; int i, fh; Object fk;
1513	>	if (tab == null)
1514	>	tab = initTable();
1515	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1516	>	if (onlyIfPresent)
1517	>	break;
1518	>	Node node = new Node(h, k, null, null);
1519	>	synchronized(node) {
1520	>	if (casTabAt(tab, i, null, node)) {
1521	>	try {
1522	>	len = 1;
1523	>	if ((val = mf.apply(k, null)) != null) {
1524	>	node.val = val;
1525	>	delta = 1;
1526	>	}
1527	>	} finally {
1528	>	if (delta == 0)
1529	>	setTabAt(tab, i, null);
1530	>	}
1531	>	}
1532		}
1533	<	else {
1534	<	int idx = e.hash & mask;
1535	<	boolean validated = false;
1536	<	synchronized (e) {
1537	<	if (tabAt(tab, i) == e) {
1538	<	validated = true;
1539	<	Node lastRun = e;
1540	<	for (Node p = e.next; p != null; p = p.next) {
1541	<	int j = p.hash & mask;
1542	<	if (j != idx) {
1543	<	idx = j;
1544	<	lastRun = p;
1533	>	if (len != 0)
1534	>	break;
1535	>	}
1536	>	else if ((fh = f.hash) < 0) {
1537	>	if ((fk = f.key) instanceof TreeBin) {
1538	>	TreeBin t = (TreeBin)fk;
1539	>	t.acquire(0);
1540	>	try {
1541	>	if (tabAt(tab, i) == f) {
1542	>	len = 1;
1543	>	TreeNode p = t.getTreeNode(h, k, t.root);
1544	>	if (p == null && onlyIfPresent)
1545	>	break;
1546	>	Object pv = (p == null) ? null : p.val;
1547	>	if ((val = mf.apply(k, (V)pv)) != null) {
1548	>	if (p != null)
1549	>	p.val = val;
1550	>	else {
1551	>	len = 2;
1552	>	delta = 1;
1553	>	t.putTreeNode(h, k, val);
1554		}
1555		}
1556	<	relaxedSetTabAt(nextTab, idx, lastRun);
1557	<	for (Node p = e; p != lastRun; p = p.next) {
1558	<	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));
1556	>	else if (p != null) {
1557	>	delta = -1;
1558	>	t.deleteTreeNode(p);
1559		}
608	–	setTabAt(tab, i, fwd);
1560		}
1561	+	} finally {
1562	+	t.release(0);
1563		}
1564	<	if (validated)
1564	>	if (len != 0)
1565		break;
1566		}
1567	+	else
1568	+	tab = (Node[])fk;
1569	+	}
1570	+	else {
1571	+	synchronized(f) {
1572	+	if (tabAt(tab, i) == f) {
1573	+	len = 1;
1574	+	for (Node e = f, pred = null;; ++len) {
1575	+	Object ek, ev;
1576	+	if (e.hash == h &&
1577	+	(ev = e.val) != null &&
1578	+	((ek = e.key) == k || k.equals(ek))) {
1579	+	val = mf.apply(k, (V)ev);
1580	+	if (val != null)
1581	+	e.val = val;
1582	+	else {
1583	+	delta = -1;
1584	+	Node en = e.next;
1585	+	if (pred != null)
1586	+	pred.next = en;
1587	+	else
1588	+	setTabAt(tab, i, en);
1589	+	}
1590	+	break;
1591	+	}
1592	+	pred = e;
1593	+	if ((e = e.next) == null) {
1594	+	if (!onlyIfPresent &&
1595	+	(val = mf.apply(k, null)) != null) {
1596	+	pred.next = new Node(h, k, val, null);
1597	+	delta = 1;
1598	+	if (len >= TREE_THRESHOLD)
1599	+	replaceWithTreeBin(tab, i, k);
1600	+	}
1601	+	break;
1602	+	}
1603	+	}
1604	+	}
1605	+	}
1606	+	if (len != 0)
1607	+	break;
1608		}
1609		}
1610	+	if (delta != 0)
1611	+	addCount((long)delta, len);
1612	+	return (V)val;
1613		}
1614
1615	<	/**
1616	<	* If not already resizing, initializes or creates next table and
1617	<	* transfers bins. Rechecks occupancy after a transfer to see if
1618	<	* another resize is already needed because resizings are lagging
1619	<	* additions.
1620	<	*
1621	<	* @param sizeHint overridden capacity target (nonzero only from putAll)
1622	<	* @return current table
1623	<	*/
1624	<	private final Node[] grow(int sizeHint) {
1625	<	if (resizing == 0 &&
1626	<	UNSAFE.compareAndSwapInt(this, resizingOffset, 0, 1)) {
1627	<	try {
1628	<	for (;;) {
1629	<	int cap, n;
1630	<	Node[] tab = table;
1631	<	if (tab == null) {
1632	<	int c = initCap;
1633	<	if (c < sizeHint)
1634	<	c = sizeHint;
1635	<	if (c == DEFAULT_CAPACITY)
1636	<	cap = c;
1637	<	else if (c >= MAXIMUM_CAPACITY)
1638	<	cap = MAXIMUM_CAPACITY;
1639	<	else {
1640	<	cap = MINIMUM_CAPACITY;
1641	<	while (cap < c)
1642	<	cap <<= 1;
1615	>	/** Implementation for merge */
1616	>	@SuppressWarnings("unchecked") private final V internalMerge
1617	>	(K k, V v, BiFun<? super V, ? super V, ? extends V> mf) {
1618	>	if (k == null || v == null || mf == null)
1619	>	throw new NullPointerException();
1620	>	int h = spread(k.hashCode());
1621	>	Object val = null;
1622	>	int delta = 0;
1623	>	int len = 0;
1624	>	for (Node[] tab = table;;) {
1625	>	int i; Node f; Object fk, fv;
1626	>	if (tab == null)
1627	>	tab = initTable();
1628	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1629	>	if (casTabAt(tab, i, null, new Node(h, k, v, null))) {
1630	>	delta = 1;
1631	>	val = v;
1632	>	break;
1633	>	}
1634	>	}
1635	>	else if (f.hash < 0) {
1636	>	if ((fk = f.key) instanceof TreeBin) {
1637	>	TreeBin t = (TreeBin)fk;
1638	>	t.acquire(0);
1639	>	try {
1640	>	if (tabAt(tab, i) == f) {
1641	>	len = 1;
1642	>	TreeNode p = t.getTreeNode(h, k, t.root);
1643	>	val = (p == null) ? v : mf.apply((V)p.val, v);
1644	>	if (val != null) {
1645	>	if (p != null)
1646	>	p.val = val;
1647	>	else {
1648	>	len = 2;
1649	>	delta = 1;
1650	>	t.putTreeNode(h, k, val);
1651	>	}
1652	>	}
1653	>	else if (p != null) {
1654	>	delta = -1;
1655	>	t.deleteTreeNode(p);
1656	>	}
1657		}
1658	+	} finally {
1659	+	t.release(0);
1660		}
1661	<	else if ((n = tab.length) < MAXIMUM_CAPACITY &&
649	<	(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))
1661	>	if (len != 0)
1662		break;
1663		}
1664	<	} finally {
1665	<	resizing = 0;
1664	>	else
1665	>	tab = (Node[])fk;
1666	>	}
1667	>	else {
1668	>	synchronized(f) {
1669	>	if (tabAt(tab, i) == f) {
1670	>	len = 1;
1671	>	for (Node e = f, pred = null;; ++len) {
1672	>	Object ek, ev;
1673	>	if (e.hash == h &&
1674	>	(ev = e.val) != null &&
1675	>	((ek = e.key) == k || k.equals(ek))) {
1676	>	val = mf.apply((V)ev, v);
1677	>	if (val != null)
1678	>	e.val = val;
1679	>	else {
1680	>	delta = -1;
1681	>	Node en = e.next;
1682	>	if (pred != null)
1683	>	pred.next = en;
1684	>	else
1685	>	setTabAt(tab, i, en);
1686	>	}
1687	>	break;
1688	>	}
1689	>	pred = e;
1690	>	if ((e = e.next) == null) {
1691	>	val = v;
1692	>	pred.next = new Node(h, k, val, null);
1693	>	delta = 1;
1694	>	if (len >= TREE_THRESHOLD)
1695	>	replaceWithTreeBin(tab, i, k);
1696	>	break;
1697	>	}
1698	>	}
1699	>	}
1700	>	}
1701	>	if (len != 0)
1702	>	break;
1703		}
1704		}
1705	<	else if (table == null)
1706	<	Thread.yield(); // lost initialization race; just spin
1707	<	return table;
1708	<	}
1709	<
1710	<	/**
1711	<	* Implementation for putAll and constructor with Map
1712	<	* argument. Tries to first override initial capacity or grow
1713	<	* based on map size to pre-allocate table space.
1714	<	*/
1715	<	private final void internalPutAll(Map<? extends K, ? extends V> m) {
1716	<	int s = m.size();
1717	<	grow((s >= (MAXIMUM_CAPACITY >>> 1)) ? s : s + (s >>> 1));
1718	<	for (Map.Entry<? extends K, ? extends V> e : m.entrySet()) {
1719	<	Object k = e.getKey();
1720	<	Object v = e.getValue();
1721	<	if (k == null || v == null)
1722	<	throw new NullPointerException();
1723	<	internalPut(k, v, true);
1705	>	if (delta != 0)
1706	>	addCount((long)delta, len);
1707	>	return (V)val;
1708	>	}
1709	>
1710	>	/** Implementation for putAll */
1711	>	private final void internalPutAll(Map<?, ?> m) {
1712	>	tryPresize(m.size());
1713	>	long delta = 0L;     // number of uncommitted additions
1714	>	boolean npe = false; // to throw exception on exit for nulls
1715	>	try {                // to clean up counts on other exceptions
1716	>	for (Map.Entry<?, ?> entry : m.entrySet()) {
1717	>	Object k, v;
1718	>	if (entry == null || (k = entry.getKey()) == null ||
1719	>	(v = entry.getValue()) == null) {
1720	>	npe = true;
1721	>	break;
1722	>	}
1723	>	int h = spread(k.hashCode());
1724	>	for (Node[] tab = table;;) {
1725	>	int i; Node f; int fh; Object fk;
1726	>	if (tab == null)
1727	>	tab = initTable();
1728	>	else if ((f = tabAt(tab, i = (tab.length - 1) & h)) == null){
1729	>	if (casTabAt(tab, i, null, new Node(h, k, v, null))) {
1730	>	++delta;
1731	>	break;
1732	>	}
1733	>	}
1734	>	else if ((fh = f.hash) < 0) {
1735	>	if ((fk = f.key) instanceof TreeBin) {
1736	>	TreeBin t = (TreeBin)fk;
1737	>	boolean validated = false;
1738	>	t.acquire(0);
1739	>	try {
1740	>	if (tabAt(tab, i) == f) {
1741	>	validated = true;
1742	>	TreeNode p = t.getTreeNode(h, k, t.root);
1743	>	if (p != null)
1744	>	p.val = v;
1745	>	else {
1746	>	t.putTreeNode(h, k, v);
1747	>	++delta;
1748	>	}
1749	>	}
1750	>	} finally {
1751	>	t.release(0);
1752	>	}
1753	>	if (validated)
1754	>	break;
1755	>	}
1756	>	else
1757	>	tab = (Node[])fk;
1758	>	}
1759	>	else {
1760	>	int len = 0;
1761	>	synchronized(f) {
1762	>	if (tabAt(tab, i) == f) {
1763	>	len = 1;
1764	>	for (Node e = f;; ++len) {
1765	>	Object ek, ev;
1766	>	if (e.hash == h &&
1767	>	(ev = e.val) != null &&
1768	>	((ek = e.key) == k || k.equals(ek))) {
1769	>	e.val = v;
1770	>	break;
1771	>	}
1772	>	Node last = e;
1773	>	if ((e = e.next) == null) {
1774	>	++delta;
1775	>	last.next = new Node(h, k, v, null);
1776	>	if (len >= TREE_THRESHOLD)
1777	>	replaceWithTreeBin(tab, i, k);
1778	>	break;
1779	>	}
1780	>	}
1781	>	}
1782	>	}
1783	>	if (len != 0) {
1784	>	if (len > 1)
1785	>	addCount(delta, len);
1786	>	break;
1787	>	}
1788	>	}
1789	>	}
1790	>	}
1791	>	} finally {
1792	>	if (delta != 0L)
1793	>	addCount(delta, 2);
1794		}
1795	+	if (npe)
1796	+	throw new NullPointerException();
1797		}
1798
1799		/**
1800	<	* Implementation for clear. Steps through each bin, removing all nodes.
1800	>	* Implementation for clear. Steps through each bin, removing all
1801	>	* nodes.
1802		*/
1803		private final void internalClear() {
1804	<	long deletions = 0L;
1804	>	long delta = 0L; // negative number of deletions
1805		int i = 0;
1806		Node[] tab = table;
1807		while (tab != null && i < tab.length) {
1808	<	Node e = tabAt(tab, i);
1809	<	if (e == null)
1808	>	Node f = tabAt(tab, i);
1809	>	if (f == null)
1810		++i;
1811	<	else if (e.hash < 0)
1812	<	tab = (Node[])e.key;
1811	>	else if (f.hash < 0) {
1812	>	Object fk;
1813	>	if ((fk = f.key) instanceof TreeBin) {
1814	>	TreeBin t = (TreeBin)fk;
1815	>	t.acquire(0);
1816	>	try {
1817	>	if (tabAt(tab, i) == f) {
1818	>	for (Node p = t.first; p != null; p = p.next) {
1819	>	if (p.val != null) { // (currently always true)
1820	>	p.val = null;
1821	>	--delta;
1822	>	}
1823	>	}
1824	>	t.first = null;
1825	>	t.root = null;
1826	>	++i;
1827	>	}
1828	>	} finally {
1829	>	t.release(0);
1830	>	}
1831	>	}
1832	>	else
1833	>	tab = (Node[])fk;
1834	>	}
1835		else {
1836	<	boolean validated = false;
1837	<	synchronized (e) {
1838	<	if (tabAt(tab, i) == e) {
1839	<	validated = true;
707	<	do {
708	<	if (e.val != null) {
1836	>	synchronized(f) {
1837	>	if (tabAt(tab, i) == f) {
1838	>	for (Node e = f; e != null; e = e.next) {
1839	>	if (e.val != null) {  // (currently always true)
1840		e.val = null;
1841	<	++deletions;
1841	>	--delta;
1842		}
1843	<	} while ((e = e.next) != null);
1843	>	}
1844		setTabAt(tab, i, null);
1845	+	++i;
1846		}
1847		}
1848	<	if (validated) {
1849	<	++i;
1850	<	if (deletions > THRESHOLD_OFFSET) { // bound lag in counts
1851	<	counter.add(-deletions);
1852	<	deletions = 0L;
1848	>	}
1849	>	}
1850	>	if (delta != 0L)
1851	>	addCount(delta, -1);
1852	>	}
1853	>
1854	>	/* ---------------- Table Initialization and Resizing -------------- */
1855	>
1856	>	/**
1857	>	* Returns a power of two table size for the given desired capacity.
1858	>	* See Hackers Delight, sec 3.2
1859	>	*/
1860	>	private static final int tableSizeFor(int c) {
1861	>	int n = c - 1;
1862	>	n |= n >>> 1;
1863	>	n |= n >>> 2;
1864	>	n |= n >>> 4;
1865	>	n |= n >>> 8;
1866	>	n |= n >>> 16;
1867	>	return (n < 0) ? 1 : (n >= MAXIMUM_CAPACITY) ? MAXIMUM_CAPACITY : n + 1;
1868	>	}
1869	>
1870	>	/**
1871	>	* Initializes table, using the size recorded in sizeCtl.
1872	>	*/
1873	>	private final Node[] initTable() {
1874	>	Node[] tab; int sc;
1875	>	while ((tab = table) == null) {
1876	>	if ((sc = sizeCtl) < 0)
1877	>	Thread.yield(); // lost initialization race; just spin
1878	>	else if (U.compareAndSwapInt(this, SIZECTL, sc, -1)) {
1879	>	try {
1880	>	if ((tab = table) == null) {
1881	>	int n = (sc > 0) ? sc : DEFAULT_CAPACITY;
1882	>	tab = table = new Node[n];
1883	>	sc = n - (n >>> 2);
1884		}
1885	+	} finally {
1886	+	sizeCtl = sc;
1887	+	}
1888	+	break;
1889	+	}
1890	+	}
1891	+	return tab;
1892	+	}
1893	+
1894	+	/**
1895	+	* Adds to count, and if table is too small and not already
1896	+	* resizing, initiates transfer. If already resizing, helps
1897	+	* perform transfer if work is available.  Rechecks occupancy
1898	+	* after a transfer to see if another resize is already needed
1899	+	* because resizings are lagging additions.
1900	+	*
1901	+	* @param x the count to add
1902	+	* @param check if <0, don't check resize, if <= 1 only check if uncontended
1903	+	*/
1904	+	private final void addCount(long x, int check) {
1905	+	CounterCell[] as; long b, s;
1906	+	if ((as = counterCells) != null ||
1907	+	!U.compareAndSwapLong(this, BASECOUNT, b = baseCount, s = b + x)) {
1908	+	CounterHashCode hc; CounterCell a; long v; int m;
1909	+	boolean uncontended = true;
1910	+	if ((hc = threadCounterHashCode.get()) == null ||
1911	+	as == null || (m = as.length - 1) < 0 ||
1912	+	(a = as[m & hc.code]) == null ||
1913	+	!(uncontended =
1914	+	U.compareAndSwapLong(a, CELLVALUE, v = a.value, v + x))) {
1915	+	fullAddCount(x, hc, uncontended);
1916	+	return;
1917	+	}
1918	+	if (check <= 1)
1919	+	return;
1920	+	s = sumCount();
1921	+	}
1922	+	if (check >= 0) {
1923	+	Node[] tab, nt; int sc;
1924	+	while (s >= (long)(sc = sizeCtl) && (tab = table) != null &&
1925	+	tab.length < MAXIMUM_CAPACITY) {
1926	+	if (sc < 0) {
1927	+	if (sc == -1 || transferIndex <= transferOrigin ||
1928	+	(nt = nextTable) == null)
1929	+	break;
1930	+	if (U.compareAndSwapInt(this, SIZECTL, sc, sc - 1))
1931	+	transfer(tab, nt);
1932		}
1933	+	else if (U.compareAndSwapInt(this, SIZECTL, sc, -2))
1934	+	transfer(tab, null);
1935	+	s = sumCount();
1936		}
1937		}
725	–	if (deletions != 0L)
726	–	counter.add(-deletions);
1938		}
1939
1940		/**
1941	<	* Base class for key, value, and entry iterators, plus internal
1942	<	* implementations of public traversal-based methods, to avoid
1943	<	* duplicating traversal code.
1941	>	* Tries to presize table to accommodate the given number of elements.
1942	>	*
1943	>	* @param size number of elements (doesn't need to be perfectly accurate)
1944		*/
1945	<	class HashIterator {
1946	<	private Node next;          // the next entry to return
1947	<	private Node[] tab;         // current table; updated if resized
1948	<	private Node lastReturned;  // the last entry returned, for remove
1949	<	private Object nextVal;     // cached value of next
1950	<	private int index;          // index of bin to use next
1951	<	private int baseIndex;      // current index of initial table
1952	<	private final int baseSize; // initial table size
1945	>	private final void tryPresize(int size) {
1946	>	int c = (size >= (MAXIMUM_CAPACITY >>> 1)) ? MAXIMUM_CAPACITY :
1947	>	tableSizeFor(size + (size >>> 1) + 1);
1948	>	int sc;
1949	>	while ((sc = sizeCtl) >= 0) {
1950	>	Node[] tab = table; int n;
1951	>	if (tab == null || (n = tab.length) == 0) {
1952	>	n = (sc > c) ? sc : c;
1953	>	if (U.compareAndSwapInt(this, SIZECTL, sc, -1)) {
1954	>	try {
1955	>	if (table == tab) {
1956	>	table = new Node[n];
1957	>	sc = n - (n >>> 2);
1958	>	}
1959	>	} finally {
1960	>	sizeCtl = sc;
1961	>	}
1962	>	}
1963	>	}
1964	>	else if (c <= sc || n >= MAXIMUM_CAPACITY)
1965	>	break;
1966	>	else if (tab == table &&
1967	>	U.compareAndSwapInt(this, SIZECTL, sc, -2))
1968	>	transfer(tab, null);
1969	>	}
1970	>	}
1971
1972	<	HashIterator() {
1973	<	Node[] t = tab = table;
1974	<	if (t == null)
1975	<	baseSize = 0;
1976	<	else {
1977	<	baseSize = t.length;
1978	<	advance(null);
1972	>	/*
1973	>	* Moves and/or copies the nodes in each bin to new table. See
1974	>	* above for explanation.
1975	>	*/
1976	>	private final void transfer(Node[] tab, Node[] nextTab) {
1977	>	int n = tab.length, stride;
1978	>	if ((stride = (NCPU > 1) ? (n >>> 3) / NCPU : n) < MIN_TRANSFER_STRIDE)
1979	>	stride = MIN_TRANSFER_STRIDE; // subdivide range
1980	>	if (nextTab == null) {            // initiating
1981	>	try {
1982	>	nextTab = new Node[n << 1];
1983	>	} catch(Throwable ex) {       // try to cope with OOME
1984	>	sizeCtl = Integer.MAX_VALUE;
1985	>	return;
1986	>	}
1987	>	nextTable = nextTab;
1988	>	transferOrigin = n;
1989	>	transferIndex = n;
1990	>	Node rev = new Node(MOVED, tab, null, null);
1991	>	for (int k = n; k > 0;) {    // progressively reveal ready slots
1992	>	int nextk = k > stride? k - stride : 0;
1993	>	for (int m = nextk; m < k; ++m)
1994	>	nextTab[m] = rev;
1995	>	for (int m = n + nextk; m < n + k; ++m)
1996	>	nextTab[m] = rev;
1997	>	U.putOrderedInt(this, TRANSFERORIGIN, k = nextk);
1998	>	}
1999	>	}
2000	>	int nextn = nextTab.length;
2001	>	Node fwd = new Node(MOVED, nextTab, null, null);
2002	>	boolean advance = true;
2003	>	for (int i = 0, bound = 0;;) {
2004	>	int nextIndex, nextBound; Node f; Object fk;
2005	>	while (advance) {
2006	>	if (--i >= bound)
2007	>	advance = false;
2008	>	else if ((nextIndex = transferIndex) <= transferOrigin) {
2009	>	i = -1;
2010	>	advance = false;
2011	>	}
2012	>	else if (U.compareAndSwapInt
2013	>	(this, TRANSFERINDEX, nextIndex,
2014	>	nextBound = (nextIndex > stride?
2015	>	nextIndex - stride : 0))) {
2016	>	bound = nextBound;
2017	>	i = nextIndex - 1;
2018	>	advance = false;
2019	>	}
2020	>	}
2021	>	if (i < 0 || i >= n || i + n >= nextn) {
2022	>	for (int sc;;) {
2023	>	if (U.compareAndSwapInt(this, SIZECTL, sc = sizeCtl, ++sc)) {
2024	>	if (sc == -1) {
2025	>	nextTable = null;
2026	>	table = nextTab;
2027	>	sizeCtl = (n << 1) - (n >>> 1);
2028	>	}
2029	>	return;
2030	>	}
2031	>	}
2032	>	}
2033	>	else if ((f = tabAt(tab, i)) == null) {
2034	>	if (casTabAt(tab, i, null, fwd)) {
2035	>	setTabAt(nextTab, i, null);
2036	>	setTabAt(nextTab, i + n, null);
2037	>	advance = true;
2038	>	}
2039	>	}
2040	>	else if (f.hash >= 0) {
2041	>	synchronized(f) {
2042	>	if (tabAt(tab, i) == f) {
2043	>	int runBit = f.hash & n;
2044	>	Node lastRun = f, lo = null, hi = null;
2045	>	for (Node p = f.next; p != null; p = p.next) {
2046	>	int b = p.hash & n;
2047	>	if (b != runBit) {
2048	>	runBit = b;
2049	>	lastRun = p;
2050	>	}
2051	>	}
2052	>	if (runBit == 0)
2053	>	lo = lastRun;
2054	>	else
2055	>	hi = lastRun;
2056	>	for (Node p = f; p != lastRun; p = p.next) {
2057	>	int ph = p.hash;
2058	>	Object pk = p.key, pv = p.val;
2059	>	if ((ph & n) == 0)
2060	>	lo = new Node(ph, pk, pv, lo);
2061	>	else
2062	>	hi = new Node(ph, pk, pv, hi);
2063	>	}
2064	>	setTabAt(nextTab, i, lo);
2065	>	setTabAt(nextTab, i + n, hi);
2066	>	setTabAt(tab, i, fwd);
2067	>	advance = true;
2068	>	}
2069	>	}
2070	>	}
2071	>	else if ((fk = f.key) instanceof TreeBin) {
2072	>	TreeBin t = (TreeBin)fk;
2073	>	t.acquire(0);
2074	>	try {
2075	>	if (tabAt(tab, i) == f) {
2076	>	TreeBin lt = new TreeBin();
2077	>	TreeBin ht = new TreeBin();
2078	>	int lc = 0, hc = 0;
2079	>	for (Node e = t.first; e != null; e = e.next) {
2080	>	int h = e.hash;
2081	>	Object k = e.key, v = e.val;
2082	>	if ((h & n) == 0) {
2083	>	++lc;
2084	>	lt.putTreeNode(h, k, v);
2085	>	}
2086	>	else {
2087	>	++hc;
2088	>	ht.putTreeNode(h, k, v);
2089	>	}
2090	>	}
2091	>	Node ln, hn; // throw away trees if too small
2092	>	if (lc < TREE_THRESHOLD) {
2093	>	ln = null;
2094	>	for (Node p = lt.first; p != null; p = p.next)
2095	>	ln = new Node(p.hash, p.key, p.val, ln);
2096	>	}
2097	>	else
2098	>	ln = new Node(MOVED, lt, null, null);
2099	>	setTabAt(nextTab, i, ln);
2100	>	if (hc < TREE_THRESHOLD) {
2101	>	hn = null;
2102	>	for (Node p = ht.first; p != null; p = p.next)
2103	>	hn = new Node(p.hash, p.key, p.val, hn);
2104	>	}
2105	>	else
2106	>	hn = new Node(MOVED, ht, null, null);
2107	>	setTabAt(nextTab, i + n, hn);
2108	>	setTabAt(tab, i, fwd);
2109	>	advance = true;
2110	>	}
2111	>	} finally {
2112	>	t.release(0);
2113	>	}
2114		}
2115	+	else
2116	+	advance = true; // already processed
2117		}
2118	+	}
2119
2120	<	public final boolean hasNext()         { return next != null; }
754	<	public final boolean hasMoreElements() { return next != null; }
2120	>	/* ---------------- Counter support -------------- */
2121
2122	<	/**
2123	<	* Advances next.  Normally, iteration proceeds bin-by-bin
2124	<	* traversing lists.  However, if the table has been resized,
2125	<	* then all future steps must traverse both the bin at the
2126	<	* current index as well as at (index + baseSize); and so on
2127	<	* for further resizings. To paranoically cope with potential
2128	<	* (improper) sharing of iterators across threads, table reads
2129	<	* are bounds-checked.
2130	<	*/
2131	<	final void advance(Node e) {
2132	<	for (;;) {
2133	<	Node[] t; int i; // for bounds checks
2134	<	if (e != null) {
2135	<	Object ek = e.key, ev = e.val;
2136	<	if (ev != null && ek != null) {
2137	<	nextVal = ev;
2138	<	next = e;
2139	<	break;
2122	>	final long sumCount() {
2123	>	CounterCell[] as = counterCells; CounterCell a;
2124	>	long sum = baseCount;
2125	>	if (as != null) {
2126	>	for (int i = 0; i < as.length; ++i) {
2127	>	if ((a = as[i]) != null)
2128	>	sum += a.value;
2129	>	}
2130	>	}
2131	>	return sum;
2132	>	}
2133	>
2134	>	// See LongAdder version for explanation
2135	>	private final void fullAddCount(long x, CounterHashCode hc,
2136	>	boolean wasUncontended) {
2137	>	int h;
2138	>	if (hc == null) {
2139	>	hc = new CounterHashCode();
2140	>	int s = counterHashCodeGenerator.addAndGet(SEED_INCREMENT);
2141	>	h = hc.code = (s == 0) ? 1 : s; // Avoid zero
2142	>	threadCounterHashCode.set(hc);
2143	>	}
2144	>	else
2145	>	h = hc.code;
2146	>	boolean collide = false;                // True if last slot nonempty
2147	>	for (;;) {
2148	>	CounterCell[] as; CounterCell a; int n; long v;
2149	>	if ((as = counterCells) != null && (n = as.length) > 0) {
2150	>	if ((a = as[(n - 1) & h]) == null) {
2151	>	if (counterBusy == 0) {            // Try to attach new Cell
2152	>	CounterCell r = new CounterCell(x); // Optimistic create
2153	>	if (counterBusy == 0 &&
2154	>	U.compareAndSwapInt(this, COUNTERBUSY, 0, 1)) {
2155	>	boolean created = false;
2156	>	try {               // Recheck under lock
2157	>	CounterCell[] rs; int m, j;
2158	>	if ((rs = counterCells) != null &&
2159	>	(m = rs.length) > 0 &&
2160	>	rs[j = (m - 1) & h] == null) {
2161	>	rs[j] = r;
2162	>	created = true;
2163	>	}
2164	>	} finally {
2165	>	counterBusy = 0;
2166	>	}
2167	>	if (created)
2168	>	break;
2169	>	continue;           // Slot is now non-empty
2170	>	}
2171		}
2172	<	e = e.next;
2172	>	collide = false;
2173		}
2174	<	else if (baseIndex < baseSize && (t = tab) != null &&
2175	<	t.length > (i = index) && i >= 0) {
2176	<	if ((e = tabAt(t, i)) != null && e.hash < 0) {
2177	<	tab = (Node[])e.key;
2178	<	e = null;
2174	>	else if (!wasUncontended)       // CAS already known to fail
2175	>	wasUncontended = true;      // Continue after rehash
2176	>	else if (U.compareAndSwapLong(a, CELLVALUE, v = a.value, v + x))
2177	>	break;
2178	>	else if (counterCells != as || n >= NCPU)
2179	>	collide = false;            // At max size or stale
2180	>	else if (!collide)
2181	>	collide = true;
2182	>	else if (counterBusy == 0 &&
2183	>	U.compareAndSwapInt(this, COUNTERBUSY, 0, 1)) {
2184	>	try {
2185	>	if (counterCells == as) {// Expand table unless stale
2186	>	CounterCell[] rs = new CounterCell[n << 1];
2187	>	for (int i = 0; i < n; ++i)
2188	>	rs[i] = as[i];
2189	>	counterCells = rs;
2190	>	}
2191	>	} finally {
2192	>	counterBusy = 0;
2193		}
2194	<	else if (i + baseSize < t.length)
2195	<	index += baseSize;    // visit forwarded upper slots
785	<	else
786	<	index = ++baseIndex;
2194	>	collide = false;
2195	>	continue;                   // Retry with expanded table
2196		}
2197	<	else {
2198	<	next = null;
2199	<	break;
2197	>	h ^= h << 13;                   // Rehash
2198	>	h ^= h >>> 17;
2199	>	h ^= h << 5;
2200	>	}
2201	>	else if (counterBusy == 0 && counterCells == as &&
2202	>	U.compareAndSwapInt(this, COUNTERBUSY, 0, 1)) {
2203	>	boolean init = false;
2204	>	try {                           // Initialize table
2205	>	if (counterCells == as) {
2206	>	CounterCell[] rs = new CounterCell[2];
2207	>	rs[h & 1] = new CounterCell(x);
2208	>	counterCells = rs;
2209	>	init = true;
2210	>	}
2211	>	} finally {
2212	>	counterBusy = 0;
2213		}
2214	+	if (init)
2215	+	break;
2216		}
2217	+	else if (U.compareAndSwapLong(this, BASECOUNT, v = baseCount, v + x))
2218	+	break;                          // Fall back on using base
2219		}
2220	+	hc.code = h;                            // Record index for next time
2221	+	}
2222
2223	<	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	<	}
2223	>	/* ----------------Table Traversal -------------- */
2224
2225	<	final Object nextValue() {
2226	<	Node e = next;
2227	<	if (e == null)
2228	<	throw new NoSuchElementException();
2229	<	Object v = nextVal;
2230	<	advance((lastReturned = e).next);
2231	<	return v;
2225	>	/**
2226	>	* Encapsulates traversal for methods such as containsValue; also
2227	>	* serves as a base class for other iterators and bulk tasks.
2228	>	*
2229	>	* At each step, the iterator snapshots the key ("nextKey") and
2230	>	* value ("nextVal") of a valid node (i.e., one that, at point of
2231	>	* snapshot, has a non-null user value). Because val fields can
2232	>	* change (including to null, indicating deletion), field nextVal
2233	>	* might not be accurate at point of use, but still maintains the
2234	>	* weak consistency property of holding a value that was once
2235	>	* valid. To support iterator.remove, the nextKey field is not
2236	>	* updated (nulled out) when the iterator cannot advance.
2237	>	*
2238	>	* Internal traversals directly access these fields, as in:
2239	>	* {@code while (it.advance() != null) { process(it.nextKey); }}
2240	>	*
2241	>	* Exported iterators must track whether the iterator has advanced
2242	>	* (in hasNext vs next) (by setting/checking/nulling field
2243	>	* nextVal), and then extract key, value, or key-value pairs as
2244	>	* return values of next().
2245	>	*
2246	>	* The iterator visits once each still-valid node that was
2247	>	* reachable upon iterator construction. It might miss some that
2248	>	* were added to a bin after the bin was visited, which is OK wrt
2249	>	* consistency guarantees. Maintaining this property in the face
2250	>	* of possible ongoing resizes requires a fair amount of
2251	>	* bookkeeping state that is difficult to optimize away amidst
2252	>	* volatile accesses.  Even so, traversal maintains reasonable
2253	>	* throughput.
2254	>	*
2255	>	* Normally, iteration proceeds bin-by-bin traversing lists.
2256	>	* However, if the table has been resized, then all future steps
2257	>	* must traverse both the bin at the current index as well as at
2258	>	* (index + baseSize); and so on for further resizings. To
2259	>	* paranoically cope with potential sharing by users of iterators
2260	>	* across threads, iteration terminates if a bounds checks fails
2261	>	* for a table read.
2262	>	*
2263	>	* This class extends CountedCompleter to streamline parallel
2264	>	* iteration in bulk operations. This adds only a few fields of
2265	>	* space overhead, which is small enough in cases where it is not
2266	>	* needed to not worry about it.  Because CountedCompleter is
2267	>	* Serializable, but iterators need not be, we need to add warning
2268	>	* suppressions.
2269	>	*/
2270	>	@SuppressWarnings("serial") static class Traverser<K,V,R>
2271	>	extends CountedCompleter<R> {
2272	>	final ConcurrentHashMapV8<K, V> map;
2273	>	Node next;           // the next entry to use
2274	>	Object nextKey;      // cached key field of next
2275	>	Object nextVal;      // cached val field of next
2276	>	Node[] tab;          // current table; updated if resized
2277	>	int index;           // index of bin to use next
2278	>	int baseIndex;       // current index of initial table
2279	>	int baseLimit;       // index bound for initial table
2280	>	int baseSize;        // initial table size
2281	>	int batch;           // split control
2282	>
2283	>	/** Creates iterator for all entries in the table. */
2284	>	Traverser(ConcurrentHashMapV8<K, V> map) {
2285	>	this.map = map;
2286	>	}
2287	>
2288	>	/** Creates iterator for split() methods and task constructors */
2289	>	Traverser(ConcurrentHashMapV8<K,V> map, Traverser<K,V,?> it, int batch) {
2290	>	super(it);
2291	>	this.batch = batch;
2292	>	if ((this.map = map) != null && it != null) { // split parent
2293	>	Node[] t;
2294	>	if ((t = it.tab) == null &&
2295	>	(t = it.tab = map.table) != null)
2296	>	it.baseLimit = it.baseSize = t.length;
2297	>	this.tab = t;
2298	>	this.baseSize = it.baseSize;
2299	>	int hi = this.baseLimit = it.baseLimit;
2300	>	it.baseLimit = this.index = this.baseIndex =
2301	>	(hi + it.baseIndex + 1) >>> 1;
2302	>	}
2303		}
2304
2305	<	final WriteThroughEntry nextEntry() {
2305	>	/**
2306	>	* Advances next; returns nextVal or null if terminated.
2307	>	* See above for explanation.
2308	>	*/
2309	>	final Object advance() {
2310		Node e = next;
2311	<	if (e == null)
2312	<	throw new NoSuchElementException();
2313	<	WriteThroughEntry entry =
2314	<	new WriteThroughEntry(e.key, nextVal);
2315	<	advance((lastReturned = e).next);
2316	<	return entry;
2311	>	Object ev = null;
2312	>	outer: do {
2313	>	if (e != null)                  // advance past used/skipped node
2314	>	e = e.next;
2315	>	while (e == null) {             // get to next non-null bin
2316	>	ConcurrentHashMapV8<K, V> m;
2317	>	Node[] t; int b, i, n; Object ek; // checks must use locals
2318	>	if ((t = tab) != null)
2319	>	n = t.length;
2320	>	else if ((m = map) != null && (t = tab = m.table) != null)
2321	>	n = baseLimit = baseSize = t.length;
2322	>	else
2323	>	break outer;
2324	>	if ((b = baseIndex) >= baseLimit ||
2325	>	(i = index) < 0 || i >= n)
2326	>	break outer;
2327	>	if ((e = tabAt(t, i)) != null && e.hash < 0) {
2328	>	if ((ek = e.key) instanceof TreeBin)
2329	>	e = ((TreeBin)ek).first;
2330	>	else {
2331	>	tab = (Node[])ek;
2332	>	continue;           // restarts due to null val
2333	>	}
2334	>	}                           // visit upper slots if present
2335	>	index = (i += baseSize) < n ? i : (baseIndex = b + 1);
2336	>	}
2337	>	nextKey = e.key;
2338	>	} while ((ev = e.val) == null);    // skip deleted or special nodes
2339	>	next = e;
2340	>	return nextVal = ev;
2341		}
2342
2343		public final void remove() {
2344	<	if (lastReturned == null)
2344	>	Object k = nextKey;
2345	>	if (k == null && (advance() == null || (k = nextKey) == null))
2346		throw new IllegalStateException();
2347	<	ConcurrentHashMapV8.this.remove(lastReturned.key);
827	<	lastReturned = null;
2347	>	map.internalReplace(k, null, null);
2348		}
2349
2350	<	/** Helper for serialization */
2351	<	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	<	}
2350	>	public final boolean hasNext() {
2351	>	return nextVal != null || advance() != null;
2352		}
2353
2354	<	/** 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	<	}
2354	>	public final boolean hasMoreElements() { return hasNext(); }
2355
2356	<	/** 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	<	}
2356	>	public void compute() { } // default no-op CountedCompleter body
2357
2358	<	/** Helper for Map.toString */
2359	<	final String mapToString() {
2360	<	Node e = next;
2361	<	if (e == null)
2362	<	return "{}";
2363	<	StringBuilder sb = new StringBuilder();
2364	<	sb.append('{');
2365	<	for (;;) {
2366	<	sb.append(e.key   == this ? "(this Map)" : e.key);
2367	<	sb.append('=');
2368	<	sb.append(nextVal == this ? "(this Map)" : nextVal);
2369	<	advance(e.next);
2370	<	if ((e = next) != null)
2371	<	sb.append(',').append(' ');
2372	<	else
2373	<	return sb.append('}').toString();
2358	>	/**
2359	>	* Returns a batch value > 0 if this task should (and must) be
2360	>	* split, if so, adding to pending count, and in any case
2361	>	* updating batch value. The initial batch value is approx
2362	>	* exp2 of the number of times (minus one) to split task by
2363	>	* two before executing leaf action. This value is faster to
2364	>	* compute and more convenient to use as a guide to splitting
2365	>	* than is the depth, since it is used while dividing by two
2366	>	* anyway.
2367	>	*/
2368	>	final int preSplit() {
2369	>	ConcurrentHashMapV8<K, V> m; int b; Node[] t;  ForkJoinPool pool;
2370	>	if ((b = batch) < 0 && (m = map) != null) { // force initialization
2371	>	if ((t = tab) == null && (t = tab = m.table) != null)
2372	>	baseLimit = baseSize = t.length;
2373	>	if (t != null) {
2374	>	long n = m.sumCount();
2375	>	int par = ((pool = getPool()) == null) ?
2376	>	ForkJoinPool.getCommonPoolParallelism() :
2377	>	pool.getParallelism();
2378	>	int sp = par << 3; // slack of 8
2379	>	b = (n <= 0L) ? 0 : (n < (long)sp) ? (int)n : sp;
2380	>	}
2381		}
2382	+	b = (b <= 1 || baseIndex == baseLimit) ? 0 : (b >>> 1);
2383	+	if ((batch = b) > 0)
2384	+	addToPendingCount(1);
2385	+	return b;
2386		}
2387	+
2388		}
2389
2390		/* ---------------- Public operations -------------- */
2391
2392		/**
2393	<	* 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.
2393	>	* Creates a new, empty map with the default initial table size (16).
2394		*/
2395	<	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;
910	<	this.counter = new LongAdder();
2395	>	public ConcurrentHashMapV8() {
2396		}
2397
2398		/**
2399	<	* Creates a new, empty map with the specified initial capacity
2400	<	* and load factor and with the default concurrencyLevel (16).
2399	>	* Creates a new, empty map with an initial table size
2400	>	* accommodating the specified number of elements without the need
2401	>	* to dynamically resize.
2402		*
2403		* @param initialCapacity The implementation performs internal
2404		* sizing to accommodate this many elements.
2405	<	* @param loadFactor  the load factor threshold, used to control resizing.
2406	<	* Resizing may be performed when the average number of elements per
2407	<	* bin exceeds this threshold.
2405	>	* @throws IllegalArgumentException if the initial capacity of
2406	>	* elements is negative
2407	>	*/
2408	>	public ConcurrentHashMapV8(int initialCapacity) {
2409	>	if (initialCapacity < 0)
2410	>	throw new IllegalArgumentException();
2411	>	int cap = ((initialCapacity >= (MAXIMUM_CAPACITY >>> 1)) ?
2412	>	MAXIMUM_CAPACITY :
2413	>	tableSizeFor(initialCapacity + (initialCapacity >>> 1) + 1));
2414	>	this.sizeCtl = cap;
2415	>	}
2416	>
2417	>	/**
2418	>	* Creates a new map with the same mappings as the given map.
2419	>	*
2420	>	* @param m the map
2421	>	*/
2422	>	public ConcurrentHashMapV8(Map<? extends K, ? extends V> m) {
2423	>	this.sizeCtl = DEFAULT_CAPACITY;
2424	>	internalPutAll(m);
2425	>	}
2426	>
2427	>	/**
2428	>	* Creates a new, empty map with an initial table size based on
2429	>	* the given number of elements ({@code initialCapacity}) and
2430	>	* initial table density ({@code loadFactor}).
2431	>	*
2432	>	* @param initialCapacity the initial capacity. The implementation
2433	>	* performs internal sizing to accommodate this many elements,
2434	>	* given the specified load factor.
2435	>	* @param loadFactor the load factor (table density) for
2436	>	* establishing the initial table size
2437		* @throws IllegalArgumentException if the initial capacity of
2438		* elements is negative or the load factor is nonpositive
2439		*
2440		* @since 1.6
2441		*/
2442		public ConcurrentHashMapV8(int initialCapacity, float loadFactor) {
2443	<	this(initialCapacity, loadFactor, DEFAULT_CONCURRENCY_LEVEL);
2443	>	this(initialCapacity, loadFactor, 1);
2444		}
2445
2446		/**
2447	<	* Creates a new, empty map with the specified initial capacity,
2448	<	* and with default load factor (0.75) and concurrencyLevel (16).
2447	>	* Creates a new, empty map with an initial table size based on
2448	>	* the given number of elements ({@code initialCapacity}), table
2449	>	* density ({@code loadFactor}), and number of concurrently
2450	>	* updating threads ({@code concurrencyLevel}).
2451		*
2452		* @param initialCapacity the initial capacity. The implementation
2453	<	* performs internal sizing to accommodate this many elements.
2454	<	* @throws IllegalArgumentException if the initial capacity of
2455	<	* elements is negative.
2453	>	* performs internal sizing to accommodate this many elements,
2454	>	* given the specified load factor.
2455	>	* @param loadFactor the load factor (table density) for
2456	>	* establishing the initial table size
2457	>	* @param concurrencyLevel the estimated number of concurrently
2458	>	* updating threads. The implementation may use this value as
2459	>	* a sizing hint.
2460	>	* @throws IllegalArgumentException if the initial capacity is
2461	>	* negative or the load factor or concurrencyLevel are
2462	>	* nonpositive
2463		*/
2464	<	public ConcurrentHashMapV8(int initialCapacity) {
2465	<	this(initialCapacity, DEFAULT_LOAD_FACTOR, DEFAULT_CONCURRENCY_LEVEL);
2464	>	public ConcurrentHashMapV8(int initialCapacity,
2465	>	float loadFactor, int concurrencyLevel) {
2466	>	if (!(loadFactor > 0.0f) || initialCapacity < 0 || concurrencyLevel <= 0)
2467	>	throw new IllegalArgumentException();
2468	>	if (initialCapacity < concurrencyLevel)   // Use at least as many bins
2469	>	initialCapacity = concurrencyLevel;   // as estimated threads
2470	>	long size = (long)(1.0 + (long)initialCapacity / loadFactor);
2471	>	int cap = (size >= (long)MAXIMUM_CAPACITY) ?
2472	>	MAXIMUM_CAPACITY : tableSizeFor((int)size);
2473	>	this.sizeCtl = cap;
2474		}
2475
2476		/**
2477	<	* Creates a new, empty map with a default initial capacity (16),
2478	<	* load factor (0.75) and concurrencyLevel (16).
2477	>	* Creates a new {@link Set} backed by a ConcurrentHashMapV8
2478	>	* from the given type to {@code Boolean.TRUE}.
2479	>	*
2480	>	* @return the new set
2481		*/
2482	<	public ConcurrentHashMapV8() {
2483	<	this(DEFAULT_CAPACITY, DEFAULT_LOAD_FACTOR, DEFAULT_CONCURRENCY_LEVEL);
2482	>	public static <K> KeySetView<K,Boolean> newKeySet() {
2483	>	return new KeySetView<K,Boolean>(new ConcurrentHashMapV8<K,Boolean>(),
2484	>	Boolean.TRUE);
2485		}
2486
2487		/**
2488	<	* Creates a new map with the same mappings as the given map.
2489	<	* 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).
2488	>	* Creates a new {@link Set} backed by a ConcurrentHashMapV8
2489	>	* from the given type to {@code Boolean.TRUE}.
2490		*
2491	<	* @param m the map
2491	>	* @param initialCapacity The implementation performs internal
2492	>	* sizing to accommodate this many elements.
2493	>	* @throws IllegalArgumentException if the initial capacity of
2494	>	* elements is negative
2495	>	* @return the new set
2496		*/
2497	<	public ConcurrentHashMapV8(Map<? extends K, ? extends V> m) {
2498	<	this(DEFAULT_CAPACITY, DEFAULT_LOAD_FACTOR, DEFAULT_CONCURRENCY_LEVEL);
2499	<	if (m == null)
963	<	throw new NullPointerException();
964	<	internalPutAll(m);
2497	>	public static <K> KeySetView<K,Boolean> newKeySet(int initialCapacity) {
2498	>	return new KeySetView<K,Boolean>
2499	>	(new ConcurrentHashMapV8<K,Boolean>(initialCapacity), Boolean.TRUE);
2500		}
2501
2502		/**
2503	<	* Returns {@code true} if this map contains no key-value mappings.
969	<	*
970	<	* @return {@code true} if this map contains no key-value mappings
2503	>	* {@inheritDoc}
2504		*/
2505		public boolean isEmpty() {
2506	<	return counter.sum() <= 0L; // ignore transient negative values
2506	>	return sumCount() <= 0L; // ignore transient negative values
2507		}
2508
2509		/**
2510	<	* 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
2510	>	* {@inheritDoc}
2511		*/
2512		public int size() {
2513	<	long n = counter.sum();
2514	<	return ((n >>> 31) == 0) ? (int)n : (n < 0L) ? 0 : Integer.MAX_VALUE;
2513	>	long n = sumCount();
2514	>	return ((n < 0L) ? 0 :
2515	>	(n > (long)Integer.MAX_VALUE) ? Integer.MAX_VALUE :
2516	>	(int)n);
2517	>	}
2518	>
2519	>	/**
2520	>	* Returns the number of mappings. This method should be used
2521	>	* instead of {@link #size} because a ConcurrentHashMapV8 may
2522	>	* contain more mappings than can be represented as an int. The
2523	>	* value returned is an estimate; the actual count may differ if
2524	>	* there are concurrent insertions or removals.
2525	>	*
2526	>	* @return the number of mappings
2527	>	*/
2528	>	public long mappingCount() {
2529	>	long n = sumCount();
2530	>	return (n < 0L) ? 0L : n; // ignore transient negative values
2531		}
2532
2533		/**
#	Line 996 | Line 2541 | public class ConcurrentHashMapV8<K, V>
2541		*
2542		* @throws NullPointerException if the specified key is null
2543		*/
999	–	@SuppressWarnings("unchecked")
2544		public V get(Object key) {
2545	<	if (key == null)
2546	<	throw new NullPointerException();
2547	<	return (V)internalGet(key);
2545	>	return internalGet(key);
2546	>	}
2547	>
2548	>	/**
2549	>	* Returns the value to which the specified key is mapped,
2550	>	* or the given defaultValue if this map contains no mapping for the key.
2551	>	*
2552	>	* @param key the key
2553	>	* @param defaultValue the value to return if this map contains
2554	>	* no mapping for the given key
2555	>	* @return the mapping for the key, if present; else the defaultValue
2556	>	* @throws NullPointerException if the specified key is null
2557	>	*/
2558	>	public V getValueOrDefault(Object key, V defaultValue) {
2559	>	V v;
2560	>	return (v = internalGet(key)) == null ? defaultValue : v;
2561		}
2562
2563		/**
#	Line 1009 | Line 2566 | public class ConcurrentHashMapV8<K, V>
2566		* @param  key   possible key
2567		* @return {@code true} if and only if the specified object
2568		*         is a key in this table, as determined by the
2569	<	*         {@code equals} method; {@code false} otherwise.
2569	>	*         {@code equals} method; {@code false} otherwise
2570		* @throws NullPointerException if the specified key is null
2571		*/
2572		public boolean containsKey(Object key) {
1016	–	if (key == null)
1017	–	throw new NullPointerException();
2573		return internalGet(key) != null;
2574		}
2575
2576		/**
2577		* Returns {@code true} if this map maps one or more keys to the
2578	<	* specified value. Note: This method requires a full internal
2579	<	* traversal of the hash table, and so is much slower than
1025	<	* method {@code containsKey}.
2578	>	* specified value. Note: This method may require a full traversal
2579	>	* of the map, and is much slower than method {@code containsKey}.
2580		*
2581		* @param value value whose presence in this map is to be tested
2582		* @return {@code true} if this map maps one or more keys to the
#	Line 1032 | Line 2586 | public class ConcurrentHashMapV8<K, V>
2586		public boolean containsValue(Object value) {
2587		if (value == null)
2588		throw new NullPointerException();
2589	<	return new HashIterator().containsVal(value);
2589	>	Object v;
2590	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
2591	>	while ((v = it.advance()) != null) {
2592	>	if (v == value || value.equals(v))
2593	>	return true;
2594	>	}
2595	>	return false;
2596		}
2597
2598		/**
#	Line 1058 | Line 2618 | public class ConcurrentHashMapV8<K, V>
2618		* Maps the specified key to the specified value in this table.
2619		* Neither the key nor the value can be null.
2620		*
2621	<	* <p> The value can be retrieved by calling the {@code get} method
2621	>	* <p>The value can be retrieved by calling the {@code get} method
2622		* with a key that is equal to the original key.
2623		*
2624		* @param key key with which the specified value is to be associated
#	Line 1067 | Line 2627 | public class ConcurrentHashMapV8<K, V>
2627		*         {@code null} if there was no mapping for {@code key}
2628		* @throws NullPointerException if the specified key or value is null
2629		*/
1070	–	@SuppressWarnings("unchecked")
2630		public V put(K key, V value) {
2631	<	if (key == null || value == null)
1073	<	throw new NullPointerException();
1074	<	return (V)internalPut(key, value, true);
2631	>	return internalPut(key, value, false);
2632		}
2633
2634		/**
#	Line 1081 | Line 2638 | public class ConcurrentHashMapV8<K, V>
2638		*         or {@code null} if there was no mapping for the key
2639		* @throws NullPointerException if the specified key or value is null
2640		*/
1084	–	@SuppressWarnings("unchecked")
2641		public V putIfAbsent(K key, V value) {
2642	<	if (key == null || value == null)
1087	<	throw new NullPointerException();
1088	<	return (V)internalPut(key, value, false);
2642	>	return internalPut(key, value, true);
2643		}
2644
2645		/**
#	Line 1096 | Line 2650 | public class ConcurrentHashMapV8<K, V>
2650		* @param m mappings to be stored in this map
2651		*/
2652		public void putAll(Map<? extends K, ? extends V> m) {
1099	–	if (m == null)
1100	–	throw new NullPointerException();
2653		internalPutAll(m);
2654		}
2655
2656		/**
2657		* If the specified key is not already associated with a value,
2658	<	* computes its value using the given mappingFunction, and if
2659	<	* non-null, enters it into the map.  This is equivalent to
2660	<	*  <pre> {@code
2658	>	* computes its value using the given mappingFunction and enters
2659	>	* it into the map unless null.  This is equivalent to
2660	>	* <pre> {@code
2661		* if (map.containsKey(key))
2662		*   return map.get(key);
2663	<	* value = mappingFunction.map(key);
2663	>	* value = mappingFunction.apply(key);
2664		* if (value != null)
2665		*   map.put(key, value);
2666		* return value;}</pre>
2667		*
2668	<	* except that the action is performed atomically.  Some attempted
2669	<	* update operations on this map by other threads may be blocked
2670	<	* while computation is in progress, so the computation should be
2671	<	* short and simple, and must not attempt to update any other
2672	<	* mappings of this Map. The most appropriate usage is to
2668	>	* except that the action is performed atomically.  If the
2669	>	* function returns {@code null} no mapping is recorded. If the
2670	>	* function itself throws an (unchecked) exception, the exception
2671	>	* is rethrown to its caller, and no mapping is recorded.  Some
2672	>	* attempted update operations on this map by other threads may be
2673	>	* blocked while computation is in progress, so the computation
2674	>	* should be short and simple, and must not attempt to update any
2675	>	* other mappings of this Map. The most appropriate usage is to
2676		* construct a new object serving as an initial mapped value, or
2677		* memoized result, as in:
2678	+	*
2679		*  <pre> {@code
2680	<	* map.computeIfAbsent(key, new MappingFunction<K, V>() {
2680	>	* map.computeIfAbsent(key, new Fun<K, V>() {
2681		*   public V map(K k) { return new Value(f(k)); }});}</pre>
2682		*
2683		* @param key key with which the specified value is to be associated
2684		* @param mappingFunction the function to compute a value
2685		* @return the current (existing or computed) value associated with
2686	<	*         the specified key, or {@code null} if the computation
1131	<	*         returned {@code null}.
2686	>	*         the specified key, or null if the computed value is null
2687		* @throws NullPointerException if the specified key or mappingFunction
2688	<	*         is null,
2688	>	*         is null
2689		* @throws IllegalStateException if the computation detectably
2690		*         attempts a recursive update to this map that would
2691	<	*         otherwise never complete.
2691	>	*         otherwise never complete
2692		* @throws RuntimeException or Error if the mappingFunction does so,
2693	<	*         in which case the mapping is left unestablished.
2693	>	*         in which case the mapping is left unestablished
2694		*/
2695	<	public V computeIfAbsent(K key, MappingFunction<? super K, ? extends V> mappingFunction) {
2696	<	if (key == null || mappingFunction == null)
2697	<	throw new NullPointerException();
1143	<	return internalCompute(key, mappingFunction, false);
2695	>	public V computeIfAbsent
2696	>	(K key, Fun<? super K, ? extends V> mappingFunction) {
2697	>	return internalComputeIfAbsent(key, mappingFunction);
2698		}
2699
2700		/**
2701	<	* Computes the value associated with the given key using the given
2702	<	* mappingFunction, and if non-null, enters it into the map.  This
1149	<	* is equivalent to
2701	>	* If the given key is present, computes a new mapping value given a key and
2702	>	* its current mapped value. This is equivalent to
2703		*  <pre> {@code
2704	<	* value = mappingFunction.map(key);
2705	<	* if (value != null)
2706	<	*   map.put(key, value);
2707	<	* else
2708	<	*   value = map.get(key);
2709	<	* return value;}</pre>
2704	>	*   if (map.containsKey(key)) {
2705	>	*     value = remappingFunction.apply(key, map.get(key));
2706	>	*     if (value != null)
2707	>	*       map.put(key, value);
2708	>	*     else
2709	>	*       map.remove(key);
2710	>	*   }
2711	>	* }</pre>
2712	>	*
2713	>	* except that the action is performed atomically.  If the
2714	>	* function returns {@code null}, the mapping is removed.  If the
2715	>	* function itself throws an (unchecked) exception, the exception
2716	>	* is rethrown to its caller, and the current mapping is left
2717	>	* unchanged.  Some attempted update operations on this map by
2718	>	* other threads may be blocked while computation is in progress,
2719	>	* so the computation should be short and simple, and must not
2720	>	* attempt to update any other mappings of this Map. For example,
2721	>	* to either create or append new messages to a value mapping:
2722		*
2723	<	* except that the action is performed atomically.  Some attempted
2724	<	* update operations on this map by other threads may be blocked
2725	<	* while computation is in progress, so the computation should be
2726	<	* short and simple, and must not attempt to update any other
2727	<	* mappings of this Map.
2723	>	* @param key key with which the specified value is to be associated
2724	>	* @param remappingFunction the function to compute a value
2725	>	* @return the new value associated with the specified key, or null if none
2726	>	* @throws NullPointerException if the specified key or remappingFunction
2727	>	*         is null
2728	>	* @throws IllegalStateException if the computation detectably
2729	>	*         attempts a recursive update to this map that would
2730	>	*         otherwise never complete
2731	>	* @throws RuntimeException or Error if the remappingFunction does so,
2732	>	*         in which case the mapping is unchanged
2733	>	*/
2734	>	public V computeIfPresent
2735	>	(K key, BiFun<? super K, ? super V, ? extends V> remappingFunction) {
2736	>	return internalCompute(key, true, remappingFunction);
2737	>	}
2738	>
2739	>	/**
2740	>	* Computes a new mapping value given a key and
2741	>	* its current mapped value (or {@code null} if there is no current
2742	>	* mapping). This is equivalent to
2743	>	*  <pre> {@code
2744	>	*   value = remappingFunction.apply(key, map.get(key));
2745	>	*   if (value != null)
2746	>	*     map.put(key, value);
2747	>	*   else
2748	>	*     map.remove(key);
2749	>	* }</pre>
2750	>	*
2751	>	* except that the action is performed atomically.  If the
2752	>	* function returns {@code null}, the mapping is removed.  If the
2753	>	* function itself throws an (unchecked) exception, the exception
2754	>	* is rethrown to its caller, and the current mapping is left
2755	>	* unchanged.  Some attempted update operations on this map by
2756	>	* other threads may be blocked while computation is in progress,
2757	>	* so the computation should be short and simple, and must not
2758	>	* attempt to update any other mappings of this Map. For example,
2759	>	* to either create or append new messages to a value mapping:
2760	>	*
2761	>	* <pre> {@code
2762	>	* Map<Key, String> map = ...;
2763	>	* final String msg = ...;
2764	>	* map.compute(key, new BiFun<Key, String, String>() {
2765	>	*   public String apply(Key k, String v) {
2766	>	*    return (v == null) ? msg : v + msg;});}}</pre>
2767		*
2768		* @param key key with which the specified value is to be associated
2769	<	* @param mappingFunction the function to compute a value
2770	<	* @return the current value associated with
2771	<	*         the specified key, or {@code null} if the computation
2772	<	*         returned {@code null} and the value was not otherwise present.
1169	<	* @throws NullPointerException if the specified key or mappingFunction
1170	<	*         is null,
2769	>	* @param remappingFunction the function to compute a value
2770	>	* @return the new value associated with the specified key, or null if none
2771	>	* @throws NullPointerException if the specified key or remappingFunction
2772	>	*         is null
2773		* @throws IllegalStateException if the computation detectably
2774		*         attempts a recursive update to this map that would
2775	<	*         otherwise never complete.
2776	<	* @throws RuntimeException or Error if the mappingFunction does so,
2777	<	*         in which case the mapping is unchanged.
2775	>	*         otherwise never complete
2776	>	* @throws RuntimeException or Error if the remappingFunction does so,
2777	>	*         in which case the mapping is unchanged
2778		*/
2779	<	public V compute(K key, MappingFunction<? super K, ? extends V> mappingFunction) {
2780	<	if (key == null || mappingFunction == null)
2781	<	throw new NullPointerException();
2782	<	return internalCompute(key, mappingFunction, true);
2779	>	public V compute
2780	>	(K key, BiFun<? super K, ? super V, ? extends V> remappingFunction) {
2781	>	return internalCompute(key, false, remappingFunction);
2782	>	}
2783	>
2784	>	/**
2785	>	* If the specified key is not already associated
2786	>	* with a value, associate it with the given value.
2787	>	* Otherwise, replace the value with the results of
2788	>	* the given remapping function. This is equivalent to:
2789	>	*  <pre> {@code
2790	>	*   if (!map.containsKey(key))
2791	>	*     map.put(value);
2792	>	*   else {
2793	>	*     newValue = remappingFunction.apply(map.get(key), value);
2794	>	*     if (value != null)
2795	>	*       map.put(key, value);
2796	>	*     else
2797	>	*       map.remove(key);
2798	>	*   }
2799	>	* }</pre>
2800	>	* except that the action is performed atomically.  If the
2801	>	* function returns {@code null}, the mapping is removed.  If the
2802	>	* function itself throws an (unchecked) exception, the exception
2803	>	* is rethrown to its caller, and the current mapping is left
2804	>	* unchanged.  Some attempted update operations on this map by
2805	>	* other threads may be blocked while computation is in progress,
2806	>	* so the computation should be short and simple, and must not
2807	>	* attempt to update any other mappings of this Map.
2808	>	*/
2809	>	public V merge
2810	>	(K key, V value,
2811	>	BiFun<? super V, ? super V, ? extends V> remappingFunction) {
2812	>	return internalMerge(key, value, remappingFunction);
2813		}
2814
2815		/**
#	Line 1189 | Line 2821 | public class ConcurrentHashMapV8<K, V>
2821		*         {@code null} if there was no mapping for {@code key}
2822		* @throws NullPointerException if the specified key is null
2823		*/
1192	–	@SuppressWarnings("unchecked")
2824		public V remove(Object key) {
2825	<	if (key == null)
1195	<	throw new NullPointerException();
1196	<	return (V)internalReplace(key, null, null);
2825	>	return internalReplace(key, null, null);
2826		}
2827
2828		/**
#	Line 1202 | Line 2831 | public class ConcurrentHashMapV8<K, V>
2831		* @throws NullPointerException if the specified key is null
2832		*/
2833		public boolean remove(Object key, Object value) {
2834	<	if (key == null)
1206	<	throw new NullPointerException();
1207	<	if (value == null)
1208	<	return false;
1209	<	return internalReplace(key, null, value) != null;
2834	>	return value != null && internalReplace(key, null, value) != null;
2835		}
2836
2837		/**
#	Line 1227 | Line 2852 | public class ConcurrentHashMapV8<K, V>
2852		*         or {@code null} if there was no mapping for the key
2853		* @throws NullPointerException if the specified key or value is null
2854		*/
1230	–	@SuppressWarnings("unchecked")
2855		public V replace(K key, V value) {
2856		if (key == null || value == null)
2857		throw new NullPointerException();
2858	<	return (V)internalReplace(key, value, null);
2858	>	return internalReplace(key, value, null);
2859		}
2860
2861		/**
#	Line 1244 | Line 2868 | public class ConcurrentHashMapV8<K, V>
2868		/**
2869		* Returns a {@link Set} view of the keys contained in this map.
2870		* The set is backed by the map, so changes to the map are
2871	<	* 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.
2871	>	* reflected in the set, and vice-versa.
2872		*
2873	<	* <p>The view's {@code iterator} is a "weakly consistent" iterator
2874	<	* that will never throw {@link ConcurrentModificationException},
2875	<	* and guarantees to traverse elements as they existed upon
2876	<	* construction of the iterator, and may (but is not guaranteed to)
2877	<	* reflect any modifications subsequent to construction.
2873	>	* @return the set view
2874	>	*/
2875	>	public KeySetView<K,V> keySet() {
2876	>	KeySetView<K,V> ks = keySet;
2877	>	return (ks != null) ? ks : (keySet = new KeySetView<K,V>(this, null));
2878	>	}
2879	>
2880	>	/**
2881	>	* Returns a {@link Set} view of the keys in this map, using the
2882	>	* given common mapped value for any additions (i.e., {@link
2883	>	* Collection#add} and {@link Collection#addAll}). This is of
2884	>	* course only appropriate if it is acceptable to use the same
2885	>	* value for all additions from this view.
2886	>	*
2887	>	* @param mappedValue the mapped value to use for any
2888	>	* additions.
2889	>	* @return the set view
2890	>	* @throws NullPointerException if the mappedValue is null
2891		*/
2892	<	public Set<K> keySet() {
2893	<	Set<K> ks = keySet;
2894	<	return (ks != null) ? ks : (keySet = new KeySet());
2892	>	public KeySetView<K,V> keySet(V mappedValue) {
2893	>	if (mappedValue == null)
2894	>	throw new NullPointerException();
2895	>	return new KeySetView<K,V>(this, mappedValue);
2896		}
2897
2898		/**
2899		* Returns a {@link Collection} view of the values contained in this map.
2900		* The collection is backed by the map, so changes to the map are
2901	<	* 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.
2901	>	* reflected in the collection, and vice-versa.
2902		*/
2903	<	public Collection<V> values() {
2904	<	Collection<V> vs = values;
2905	<	return (vs != null) ? vs : (values = new Values());
2903	>	public ValuesView<K,V> values() {
2904	>	ValuesView<K,V> vs = values;
2905	>	return (vs != null) ? vs : (values = new ValuesView<K,V>(this));
2906		}
2907
2908		/**
#	Line 1300 | Line 2922 | public class ConcurrentHashMapV8<K, V>
2922		* reflect any modifications subsequent to construction.
2923		*/
2924		public Set<Map.Entry<K,V>> entrySet() {
2925	<	Set<Map.Entry<K,V>> es = entrySet;
2926	<	return (es != null) ? es : (entrySet = new EntrySet());
2925	>	EntrySetView<K,V> es = entrySet;
2926	>	return (es != null) ? es : (entrySet = new EntrySetView<K,V>(this));
2927		}
2928
2929		/**
#	Line 1311 | Line 2933 | public class ConcurrentHashMapV8<K, V>
2933		* @see #keySet()
2934		*/
2935		public Enumeration<K> keys() {
2936	<	return new KeyIterator();
2936	>	return new KeyIterator<K,V>(this);
2937		}
2938
2939		/**
#	Line 1321 | Line 2943 | public class ConcurrentHashMapV8<K, V>
2943		* @see #values()
2944		*/
2945		public Enumeration<V> elements() {
2946	<	return new ValueIterator();
2946	>	return new ValueIterator<K,V>(this);
2947	>	}
2948	>
2949	>	/**
2950	>	* Returns a partitionable iterator of the keys in this map.
2951	>	*
2952	>	* @return a partitionable iterator of the keys in this map
2953	>	*/
2954	>	public Spliterator<K> keySpliterator() {
2955	>	return new KeyIterator<K,V>(this);
2956	>	}
2957	>
2958	>	/**
2959	>	* Returns a partitionable iterator of the values in this map.
2960	>	*
2961	>	* @return a partitionable iterator of the values in this map
2962	>	*/
2963	>	public Spliterator<V> valueSpliterator() {
2964	>	return new ValueIterator<K,V>(this);
2965	>	}
2966	>
2967	>	/**
2968	>	* Returns a partitionable iterator of the entries in this map.
2969	>	*
2970	>	* @return a partitionable iterator of the entries in this map
2971	>	*/
2972	>	public Spliterator<Map.Entry<K,V>> entrySpliterator() {
2973	>	return new EntryIterator<K,V>(this);
2974		}
2975
2976		/**
#	Line 1332 | Line 2981 | public class ConcurrentHashMapV8<K, V>
2981		* @return the hash code value for this map
2982		*/
2983		public int hashCode() {
2984	<	return new HashIterator().mapHashCode();
2984	>	int h = 0;
2985	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
2986	>	Object v;
2987	>	while ((v = it.advance()) != null) {
2988	>	h += it.nextKey.hashCode() ^ v.hashCode();
2989	>	}
2990	>	return h;
2991		}
2992
2993		/**
#	Line 1347 | Line 3002 | public class ConcurrentHashMapV8<K, V>
3002		* @return a string representation of this map
3003		*/
3004		public String toString() {
3005	<	return new HashIterator().mapToString();
3005	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
3006	>	StringBuilder sb = new StringBuilder();
3007	>	sb.append('{');
3008	>	Object v;
3009	>	if ((v = it.advance()) != null) {
3010	>	for (;;) {
3011	>	Object k = it.nextKey;
3012	>	sb.append(k == this ? "(this Map)" : k);
3013	>	sb.append('=');
3014	>	sb.append(v == this ? "(this Map)" : v);
3015	>	if ((v = it.advance()) == null)
3016	>	break;
3017	>	sb.append(',').append(' ');
3018	>	}
3019	>	}
3020	>	return sb.append('}').toString();
3021		}
3022
3023		/**
#	Line 1361 | Line 3031 | public class ConcurrentHashMapV8<K, V>
3031		* @return {@code true} if the specified object is equal to this map
3032		*/
3033		public boolean equals(Object o) {
3034	<	if (o == this)
3035	<	return true;
3036	<	if (!(o instanceof Map))
3037	<	return false;
3038	<	Map<?,?> m = (Map<?,?>) o;
3039	<	try {
3040	<	for (Map.Entry<K,V> e : this.entrySet())
3041	<	if (! e.getValue().equals(m.get(e.getKey())))
3034	>	if (o != this) {
3035	>	if (!(o instanceof Map))
3036	>	return false;
3037	>	Map<?,?> m = (Map<?,?>) o;
3038	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
3039	>	Object val;
3040	>	while ((val = it.advance()) != null) {
3041	>	Object v = m.get(it.nextKey);
3042	>	if (v == null || (v != val && !v.equals(val)))
3043		return false;
3044	+	}
3045		for (Map.Entry<?,?> e : m.entrySet()) {
3046	<	Object k = e.getKey();
3047	<	Object v = e.getValue();
3048	<	if (k == null || v == null || !v.equals(get(k)))
3046	>	Object mk, mv, v;
3047	>	if ((mk = e.getKey()) == null ||
3048	>	(mv = e.getValue()) == null ||
3049	>	(v = internalGet(mk)) == null ||
3050	>	(mv != v && !mv.equals(v)))
3051		return false;
3052		}
1379	–	return true;
1380	–	} catch (ClassCastException unused) {
1381	–	return false;
1382	–	} catch (NullPointerException unused) {
1383	–	return false;
3053		}
3054	+	return true;
3055		}
3056
3057	<	/**
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	<	}
3057	>	/* ----------------Iterators -------------- */
3058
3059	<	/**
3060	<	* Sets our entry's value and writes through to the map. The
3061	<	* value to return is somewhat arbitrary here. Since a
3062	<	* WriteThroughEntry does not necessarily track asynchronous
3063	<	* changes, the most recent "previous" value could be
3064	<	* different from what we return (or could even have been
3065	<	* removed in which case the put will re-establish). We do not
3066	<	* and cannot guarantee more.
3067	<	*/
3068	<	public V setValue(V value) {
3069	<	if (value == null) throw new NullPointerException();
3070	<	V v = super.setValue(value);
3071	<	ConcurrentHashMapV8.this.put(getKey(), value);
3072	<	return v;
3059	>	@SuppressWarnings("serial") static final class KeyIterator<K,V>
3060	>	extends Traverser<K,V,Object>
3061	>	implements Spliterator<K>, Enumeration<K> {
3062	>	KeyIterator(ConcurrentHashMapV8<K, V> map) { super(map); }
3063	>	KeyIterator(ConcurrentHashMapV8<K, V> map, Traverser<K,V,Object> it) {
3064	>	super(map, it, -1);
3065	>	}
3066	>	public KeyIterator<K,V> split() {
3067	>	if (nextKey != null)
3068	>	throw new IllegalStateException();
3069	>	return new KeyIterator<K,V>(map, this);
3070	>	}
3071	>	@SuppressWarnings("unchecked") public final K next() {
3072	>	if (nextVal == null && advance() == null)
3073	>	throw new NoSuchElementException();
3074	>	Object k = nextKey;
3075	>	nextVal = null;
3076	>	return (K) k;
3077		}
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	–	}
3078
3079	<	final class EntryIterator extends HashIterator
1431	<	implements Iterator<Entry<K,V>> {
1432	<	public final Map.Entry<K,V> next() { return super.nextEntry(); }
3079	>	public final K nextElement() { return next(); }
3080		}
3081
3082	<	final class KeySet extends AbstractSet<K> {
3083	<	public int size() {
3084	<	return ConcurrentHashMapV8.this.size();
3082	>	@SuppressWarnings("serial") static final class ValueIterator<K,V>
3083	>	extends Traverser<K,V,Object>
3084	>	implements Spliterator<V>, Enumeration<V> {
3085	>	ValueIterator(ConcurrentHashMapV8<K, V> map) { super(map); }
3086	>	ValueIterator(ConcurrentHashMapV8<K, V> map, Traverser<K,V,Object> it) {
3087	>	super(map, it, -1);
3088		}
3089	<	public boolean isEmpty() {
3090	<	return ConcurrentHashMapV8.this.isEmpty();
3091	<	}
3092	<	public void clear() {
1443	<	ConcurrentHashMapV8.this.clear();
1444	<	}
1445	<	public Iterator<K> iterator() {
1446	<	return new KeyIterator();
1447	<	}
1448	<	public boolean contains(Object o) {
1449	<	return ConcurrentHashMapV8.this.containsKey(o);
3089	>	public ValueIterator<K,V> split() {
3090	>	if (nextKey != null)
3091	>	throw new IllegalStateException();
3092	>	return new ValueIterator<K,V>(map, this);
3093		}
3094	<	public boolean remove(Object o) {
3095	<	return ConcurrentHashMapV8.this.remove(o) != null;
3094	>
3095	>	@SuppressWarnings("unchecked") public final V next() {
3096	>	Object v;
3097	>	if ((v = nextVal) == null && (v = advance()) == null)
3098	>	throw new NoSuchElementException();
3099	>	nextVal = null;
3100	>	return (V) v;
3101		}
3102	+
3103	+	public final V nextElement() { return next(); }
3104		}
3105
3106	<	final class Values extends AbstractCollection<V> {
3107	<	public int size() {
3108	<	return ConcurrentHashMapV8.this.size();
3109	<	}
3110	<	public boolean isEmpty() {
3111	<	return ConcurrentHashMapV8.this.isEmpty();
1462	<	}
1463	<	public void clear() {
1464	<	ConcurrentHashMapV8.this.clear();
3106	>	@SuppressWarnings("serial") static final class EntryIterator<K,V>
3107	>	extends Traverser<K,V,Object>
3108	>	implements Spliterator<Map.Entry<K,V>> {
3109	>	EntryIterator(ConcurrentHashMapV8<K, V> map) { super(map); }
3110	>	EntryIterator(ConcurrentHashMapV8<K, V> map, Traverser<K,V,Object> it) {
3111	>	super(map, it, -1);
3112		}
3113	<	public Iterator<V> iterator() {
3114	<	return new ValueIterator();
3113	>	public EntryIterator<K,V> split() {
3114	>	if (nextKey != null)
3115	>	throw new IllegalStateException();
3116	>	return new EntryIterator<K,V>(map, this);
3117		}
3118	<	public boolean contains(Object o) {
3119	<	return ConcurrentHashMapV8.this.containsValue(o);
3118	>
3119	>	@SuppressWarnings("unchecked") public final Map.Entry<K,V> next() {
3120	>	Object v;
3121	>	if ((v = nextVal) == null && (v = advance()) == null)
3122	>	throw new NoSuchElementException();
3123	>	Object k = nextKey;
3124	>	nextVal = null;
3125	>	return new MapEntry<K,V>((K)k, (V)v, map);
3126		}
3127		}
3128
3129	<	final class EntrySet extends AbstractSet<Map.Entry<K,V>> {
3130	<	public int size() {
3131	<	return ConcurrentHashMapV8.this.size();
3132	<	}
3133	<	public boolean isEmpty() {
3134	<	return ConcurrentHashMapV8.this.isEmpty();
3135	<	}
3136	<	public void clear() {
3137	<	ConcurrentHashMapV8.this.clear();
3138	<	}
3139	<	public Iterator<Map.Entry<K,V>> iterator() {
1485	<	return new EntryIterator();
3129	>	/**
3130	>	* Exported Entry for iterators
3131	>	*/
3132	>	static final class MapEntry<K,V> implements Map.Entry<K, V> {
3133	>	final K key; // non-null
3134	>	V val;       // non-null
3135	>	final ConcurrentHashMapV8<K, V> map;
3136	>	MapEntry(K key, V val, ConcurrentHashMapV8<K, V> map) {
3137	>	this.key = key;
3138	>	this.val = val;
3139	>	this.map = map;
3140		}
3141	<	public boolean contains(Object o) {
3142	<	if (!(o instanceof Map.Entry))
3143	<	return false;
3144	<	Map.Entry<?,?> e = (Map.Entry<?,?>)o;
3145	<	V v = ConcurrentHashMapV8.this.get(e.getKey());
3146	<	return v != null && v.equals(e.getValue());
3141	>	public final K getKey()       { return key; }
3142	>	public final V getValue()     { return val; }
3143	>	public final int hashCode()   { return key.hashCode() ^ val.hashCode(); }
3144	>	public final String toString(){ return key + "=" + val; }
3145	>
3146	>	public final boolean equals(Object o) {
3147	>	Object k, v; Map.Entry<?,?> e;
3148	>	return ((o instanceof Map.Entry) &&
3149	>	(k = (e = (Map.Entry<?,?>)o).getKey()) != null &&
3150	>	(v = e.getValue()) != null &&
3151	>	(k == key || k.equals(key)) &&
3152	>	(v == val || v.equals(val)));
3153		}
3154	<	public boolean remove(Object o) {
3155	<	if (!(o instanceof Map.Entry))
3156	<	return false;
3157	<	Map.Entry<?,?> e = (Map.Entry<?,?>)o;
3158	<	return ConcurrentHashMapV8.this.remove(e.getKey(), e.getValue());
3154	>
3155	>	/**
3156	>	* Sets our entry's value and writes through to the map. The
3157	>	* value to return is somewhat arbitrary here. Since we do not
3158	>	* necessarily track asynchronous changes, the most recent
3159	>	* "previous" value could be different from what we return (or
3160	>	* could even have been removed in which case the put will
3161	>	* re-establish). We do not and cannot guarantee more.
3162	>	*/
3163	>	public final V setValue(V value) {
3164	>	if (value == null) throw new NullPointerException();
3165	>	V v = val;
3166	>	val = value;
3167	>	map.put(key, value);
3168	>	return v;
3169		}
3170		}
3171
3172	+	/**
3173	+	* Returns exportable snapshot entry for the given key and value
3174	+	* when write-through can't or shouldn't be used.
3175	+	*/
3176	+	static <K,V> AbstractMap.SimpleEntry<K,V> entryFor(K k, V v) {
3177	+	return new AbstractMap.SimpleEntry<K,V>(k, v);
3178	+	}
3179	+
3180		/* ---------------- Serialization Support -------------- */
3181
3182		/**
3183	<	* Helper class used in previous version, declared for the sake of
3184	<	* serialization compatibility
3183	>	* Stripped-down version of helper class used in previous version,
3184	>	* declared for the sake of serialization compatibility
3185		*/
3186	<	static class Segment<K,V> extends java.util.concurrent.locks.ReentrantLock
1509	<	implements Serializable {
3186	>	static class Segment<K,V> implements Serializable {
3187		private static final long serialVersionUID = 2249069246763182397L;
3188		final float loadFactor;
3189		Segment(float lf) { this.loadFactor = lf; }
#	Line 1521 | Line 3198 | public class ConcurrentHashMapV8<K, V>
3198		* for each key-value mapping, followed by a null pair.
3199		* The key-value mappings are emitted in no particular order.
3200		*/
3201	<	@SuppressWarnings("unchecked")
3202	<	private void writeObject(java.io.ObjectOutputStream s)
3203	<	throws java.io.IOException {
3201	>	@SuppressWarnings("unchecked") private void writeObject
3202	>	(java.io.ObjectOutputStream s)
3203	>	throws java.io.IOException {
3204		if (segments == null) { // for serialization compatibility
3205		segments = (Segment<K,V>[])
3206		new Segment<?,?>[DEFAULT_CONCURRENCY_LEVEL];
3207		for (int i = 0; i < segments.length; ++i)
3208	<	segments[i] = new Segment<K,V>(loadFactor);
3208	>	segments[i] = new Segment<K,V>(LOAD_FACTOR);
3209		}
3210		s.defaultWriteObject();
3211	<	new HashIterator().writeEntries(s);
3211	>	Traverser<K,V,Object> it = new Traverser<K,V,Object>(this);
3212	>	Object v;
3213	>	while ((v = it.advance()) != null) {
3214	>	s.writeObject(it.nextKey);
3215	>	s.writeObject(v);
3216	>	}
3217		s.writeObject(null);
3218		s.writeObject(null);
3219		segments = null; // throw away
#	Line 1541 | Line 3223 | public class ConcurrentHashMapV8<K, V>
3223		* Reconstitutes the instance from a stream (that is, deserializes it).
3224		* @param s the stream
3225		*/
3226	<	@SuppressWarnings("unchecked")
3227	<	private void readObject(java.io.ObjectInputStream s)
3228	<	throws java.io.IOException, ClassNotFoundException {
3226	>	@SuppressWarnings("unchecked") private void readObject
3227	>	(java.io.ObjectInputStream s)
3228	>	throws java.io.IOException, ClassNotFoundException {
3229		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);
3230		this.segments = null; // unneeded
3231
3232	<	// Read the keys and values, and put the mappings in the table
3232	>	// Create all nodes, then place in table once size is known
3233	>	long size = 0L;
3234	>	Node p = null;
3235		for (;;) {
3236	<	K key = (K) s.readObject();
3237	<	V value = (V) s.readObject();
3238	<	if (key == null)
3236	>	K k = (K) s.readObject();
3237	>	V v = (V) s.readObject();
3238	>	if (k != null && v != null) {
3239	>	int h = spread(k.hashCode());
3240	>	p = new Node(h, k, v, p);
3241	>	++size;
3242	>	}
3243	>	else
3244		break;
3245	<	put(key, value);
3245	>	}
3246	>	if (p != null) {
3247	>	boolean init = false;
3248	>	int n;
3249	>	if (size >= (long)(MAXIMUM_CAPACITY >>> 1))
3250	>	n = MAXIMUM_CAPACITY;
3251	>	else {
3252	>	int sz = (int)size;
3253	>	n = tableSizeFor(sz + (sz >>> 1) + 1);
3254	>	}
3255	>	int sc = sizeCtl;
3256	>	boolean collide = false;
3257	>	if (n > sc &&
3258	>	U.compareAndSwapInt(this, SIZECTL, sc, -1)) {
3259	>	try {
3260	>	if (table == null) {
3261	>	init = true;
3262	>	Node[] tab = new Node[n];
3263	>	int mask = n - 1;
3264	>	while (p != null) {
3265	>	int j = p.hash & mask;
3266	>	Node next = p.next;
3267	>	Node q = p.next = tabAt(tab, j);
3268	>	setTabAt(tab, j, p);
3269	>	if (!collide && q != null && q.hash == p.hash)
3270	>	collide = true;
3271	>	p = next;
3272	>	}
3273	>	table = tab;
3274	>	addCount(size, -1);
3275	>	sc = n - (n >>> 2);
3276	>	}
3277	>	} finally {
3278	>	sizeCtl = sc;
3279	>	}
3280	>	if (collide) { // rescan and convert to TreeBins
3281	>	Node[] tab = table;
3282	>	for (int i = 0; i < tab.length; ++i) {
3283	>	int c = 0;
3284	>	for (Node e = tabAt(tab, i); e != null; e = e.next) {
3285	>	if (++c > TREE_THRESHOLD &&
3286	>	(e.key instanceof Comparable)) {
3287	>	replaceWithTreeBin(tab, i, e.key);
3288	>	break;
3289	>	}
3290	>	}
3291	>	}
3292	>	}
3293	>	}
3294	>	if (!init) { // Can only happen if unsafely published.
3295	>	while (p != null) {
3296	>	internalPut((K)p.key, (V)p.val, false);
3297	>	p = p.next;
3298	>	}
3299	>	}
3300	>	}
3301	>	}
3302	>
3303	>	// -------------------------------------------------------
3304	>
3305	>	// Sams
3306	>	/** Interface describing a void action of one argument */
3307	>	public interface Action<A> { void apply(A a); }
3308	>	/** Interface describing a void action of two arguments */
3309	>	public interface BiAction<A,B> { void apply(A a, B b); }
3310	>	/** Interface describing a function of one argument */
3311	>	public interface Fun<A,T> { T apply(A a); }
3312	>	/** Interface describing a function of two arguments */
3313	>	public interface BiFun<A,B,T> { T apply(A a, B b); }
3314	>	/** Interface describing a function of no arguments */
3315	>	public interface Generator<T> { T apply(); }
3316	>	/** Interface describing a function mapping its argument to a double */
3317	>	public interface ObjectToDouble<A> { double apply(A a); }
3318	>	/** Interface describing a function mapping its argument to a long */
3319	>	public interface ObjectToLong<A> { long apply(A a); }
3320	>	/** Interface describing a function mapping its argument to an int */
3321	>	public interface ObjectToInt<A> {int apply(A a); }
3322	>	/** Interface describing a function mapping two arguments to a double */
3323	>	public interface ObjectByObjectToDouble<A,B> { double apply(A a, B b); }
3324	>	/** Interface describing a function mapping two arguments to a long */
3325	>	public interface ObjectByObjectToLong<A,B> { long apply(A a, B b); }
3326	>	/** Interface describing a function mapping two arguments to an int */
3327	>	public interface ObjectByObjectToInt<A,B> {int apply(A a, B b); }
3328	>	/** Interface describing a function mapping a double to a double */
3329	>	public interface DoubleToDouble { double apply(double a); }
3330	>	/** Interface describing a function mapping a long to a long */
3331	>	public interface LongToLong { long apply(long a); }
3332	>	/** Interface describing a function mapping an int to an int */
3333	>	public interface IntToInt { int apply(int a); }
3334	>	/** Interface describing a function mapping two doubles to a double */
3335	>	public interface DoubleByDoubleToDouble { double apply(double a, double b); }
3336	>	/** Interface describing a function mapping two longs to a long */
3337	>	public interface LongByLongToLong { long apply(long a, long b); }
3338	>	/** Interface describing a function mapping two ints to an int */
3339	>	public interface IntByIntToInt { int apply(int a, int b); }
3340	>
3341	>
3342	>	// -------------------------------------------------------
3343	>
3344	>	/**
3345	>	* Performs the given action for each (key, value).
3346	>	*
3347	>	* @param action the action
3348	>	*/
3349	>	public void forEach(BiAction<K,V> action) {
3350	>	ForkJoinTasks.forEach
3351	>	(this, action).invoke();
3352	>	}
3353	>
3354	>	/**
3355	>	* Performs the given action for each non-null transformation
3356	>	* of each (key, value).
3357	>	*
3358	>	* @param transformer a function returning the transformation
3359	>	* for an element, or null of there is no transformation (in
3360	>	* which case the action is not applied).
3361	>	* @param action the action
3362	>	*/
3363	>	public <U> void forEach(BiFun<? super K, ? super V, ? extends U> transformer,
3364	>	Action<U> action) {
3365	>	ForkJoinTasks.forEach
3366	>	(this, transformer, action).invoke();
3367	>	}
3368	>
3369	>	/**
3370	>	* Returns a non-null result from applying the given search
3371	>	* function on each (key, value), or null if none.  Upon
3372	>	* success, further element processing is suppressed and the
3373	>	* results of any other parallel invocations of the search
3374	>	* function are ignored.
3375	>	*
3376	>	* @param searchFunction a function returning a non-null
3377	>	* result on success, else null
3378	>	* @return a non-null result from applying the given search
3379	>	* function on each (key, value), or null if none
3380	>	*/
3381	>	public <U> U search(BiFun<? super K, ? super V, ? extends U> searchFunction) {
3382	>	return ForkJoinTasks.search
3383	>	(this, searchFunction).invoke();
3384	>	}
3385	>
3386	>	/**
3387	>	* Returns the result of accumulating the given transformation
3388	>	* of all (key, value) pairs using the given reducer to
3389	>	* combine values, or null if none.
3390	>	*
3391	>	* @param transformer a function returning the transformation
3392	>	* for an element, or null of there is no transformation (in
3393	>	* which case it is not combined).
3394	>	* @param reducer a commutative associative combining function
3395	>	* @return the result of accumulating the given transformation
3396	>	* of all (key, value) pairs
3397	>	*/
3398	>	public <U> U reduce(BiFun<? super K, ? super V, ? extends U> transformer,
3399	>	BiFun<? super U, ? super U, ? extends U> reducer) {
3400	>	return ForkJoinTasks.reduce
3401	>	(this, transformer, reducer).invoke();
3402	>	}
3403	>
3404	>	/**
3405	>	* Returns the result of accumulating the given transformation
3406	>	* of all (key, value) pairs using the given reducer to
3407	>	* combine values, and the given basis as an identity value.
3408	>	*
3409	>	* @param transformer a function returning the transformation
3410	>	* for an element
3411	>	* @param basis the identity (initial default value) for the reduction
3412	>	* @param reducer a commutative associative combining function
3413	>	* @return the result of accumulating the given transformation
3414	>	* of all (key, value) pairs
3415	>	*/
3416	>	public double reduceToDouble(ObjectByObjectToDouble<? super K, ? super V> transformer,
3417	>	double basis,
3418	>	DoubleByDoubleToDouble reducer) {
3419	>	return ForkJoinTasks.reduceToDouble
3420	>	(this, transformer, basis, reducer).invoke();
3421	>	}
3422	>
3423	>	/**
3424	>	* Returns the result of accumulating the given transformation
3425	>	* of all (key, value) pairs using the given reducer to
3426	>	* combine values, and the given basis as an identity value.
3427	>	*
3428	>	* @param transformer a function returning the transformation
3429	>	* for an element
3430	>	* @param basis the identity (initial default value) for the reduction
3431	>	* @param reducer a commutative associative combining function
3432	>	* @return the result of accumulating the given transformation
3433	>	* of all (key, value) pairs
3434	>	*/
3435	>	public long reduceToLong(ObjectByObjectToLong<? super K, ? super V> transformer,
3436	>	long basis,
3437	>	LongByLongToLong reducer) {
3438	>	return ForkJoinTasks.reduceToLong
3439	>	(this, transformer, basis, reducer).invoke();
3440	>	}
3441	>
3442	>	/**
3443	>	* Returns the result of accumulating the given transformation
3444	>	* of all (key, value) pairs using the given reducer to
3445	>	* combine values, and the given basis as an identity value.
3446	>	*
3447	>	* @param transformer a function returning the transformation
3448	>	* for an element
3449	>	* @param basis the identity (initial default value) for the reduction
3450	>	* @param reducer a commutative associative combining function
3451	>	* @return the result of accumulating the given transformation
3452	>	* of all (key, value) pairs
3453	>	*/
3454	>	public int reduceToInt(ObjectByObjectToInt<? super K, ? super V> transformer,
3455	>	int basis,
3456	>	IntByIntToInt reducer) {
3457	>	return ForkJoinTasks.reduceToInt
3458	>	(this, transformer, basis, reducer).invoke();
3459	>	}
3460	>
3461	>	/**
3462	>	* Performs the given action for each key.
3463	>	*
3464	>	* @param action the action
3465	>	*/
3466	>	public void forEachKey(Action<K> action) {
3467	>	ForkJoinTasks.forEachKey
3468	>	(this, action).invoke();
3469	>	}
3470	>
3471	>	/**
3472	>	* Performs the given action for each non-null transformation
3473	>	* of each key.
3474	>	*
3475	>	* @param transformer a function returning the transformation
3476	>	* for an element, or null of there is no transformation (in
3477	>	* which case the action is not applied).
3478	>	* @param action the action
3479	>	*/
3480	>	public <U> void forEachKey(Fun<? super K, ? extends U> transformer,
3481	>	Action<U> action) {
3482	>	ForkJoinTasks.forEachKey
3483	>	(this, transformer, action).invoke();
3484	>	}
3485	>
3486	>	/**
3487	>	* Returns a non-null result from applying the given search
3488	>	* function on each key, or null if none. Upon success,
3489	>	* further element processing is suppressed and the results of
3490	>	* any other parallel invocations of the search function are
3491	>	* ignored.
3492	>	*
3493	>	* @param searchFunction a function returning a non-null
3494	>	* result on success, else null
3495	>	* @return a non-null result from applying the given search
3496	>	* function on each key, or null if none
3497	>	*/
3498	>	public <U> U searchKeys(Fun<? super K, ? extends U> searchFunction) {
3499	>	return ForkJoinTasks.searchKeys
3500	>	(this, searchFunction).invoke();
3501	>	}
3502	>
3503	>	/**
3504	>	* Returns the result of accumulating all keys using the given
3505	>	* reducer to combine values, or null if none.
3506	>	*
3507	>	* @param reducer a commutative associative combining function
3508	>	* @return the result of accumulating all keys using the given
3509	>	* reducer to combine values, or null if none
3510	>	*/
3511	>	public K reduceKeys(BiFun<? super K, ? super K, ? extends K> reducer) {
3512	>	return ForkJoinTasks.reduceKeys
3513	>	(this, reducer).invoke();
3514	>	}
3515	>
3516	>	/**
3517	>	* Returns the result of accumulating the given transformation
3518	>	* of all keys using the given reducer to combine values, or
3519	>	* null if none.
3520	>	*
3521	>	* @param transformer a function returning the transformation
3522	>	* for an element, or null of there is no transformation (in
3523	>	* which case it is not combined).
3524	>	* @param reducer a commutative associative combining function
3525	>	* @return the result of accumulating the given transformation
3526	>	* of all keys
3527	>	*/
3528	>	public <U> U reduceKeys(Fun<? super K, ? extends U> transformer,
3529	>	BiFun<? super U, ? super U, ? extends U> reducer) {
3530	>	return ForkJoinTasks.reduceKeys
3531	>	(this, transformer, reducer).invoke();
3532	>	}
3533	>
3534	>	/**
3535	>	* Returns the result of accumulating the given transformation
3536	>	* of all keys using the given reducer to combine values, and
3537	>	* the given basis as an identity value.
3538	>	*
3539	>	* @param transformer a function returning the transformation
3540	>	* for an element
3541	>	* @param basis the identity (initial default value) for the reduction
3542	>	* @param reducer a commutative associative combining function
3543	>	* @return  the result of accumulating the given transformation
3544	>	* of all keys
3545	>	*/
3546	>	public double reduceKeysToDouble(ObjectToDouble<? super K> transformer,
3547	>	double basis,
3548	>	DoubleByDoubleToDouble reducer) {
3549	>	return ForkJoinTasks.reduceKeysToDouble
3550	>	(this, transformer, basis, reducer).invoke();
3551	>	}
3552	>
3553	>	/**
3554	>	* Returns the result of accumulating the given transformation
3555	>	* of all keys using the given reducer to combine values, and
3556	>	* the given basis as an identity value.
3557	>	*
3558	>	* @param transformer a function returning the transformation
3559	>	* for an element
3560	>	* @param basis the identity (initial default value) for the reduction
3561	>	* @param reducer a commutative associative combining function
3562	>	* @return the result of accumulating the given transformation
3563	>	* of all keys
3564	>	*/
3565	>	public long reduceKeysToLong(ObjectToLong<? super K> transformer,
3566	>	long basis,
3567	>	LongByLongToLong reducer) {
3568	>	return ForkJoinTasks.reduceKeysToLong
3569	>	(this, transformer, basis, reducer).invoke();
3570	>	}
3571	>
3572	>	/**
3573	>	* Returns the result of accumulating the given transformation
3574	>	* of all keys using the given reducer to combine values, and
3575	>	* the given basis as an identity value.
3576	>	*
3577	>	* @param transformer a function returning the transformation
3578	>	* for an element
3579	>	* @param basis the identity (initial default value) for the reduction
3580	>	* @param reducer a commutative associative combining function
3581	>	* @return the result of accumulating the given transformation
3582	>	* of all keys
3583	>	*/
3584	>	public int reduceKeysToInt(ObjectToInt<? super K> transformer,
3585	>	int basis,
3586	>	IntByIntToInt reducer) {
3587	>	return ForkJoinTasks.reduceKeysToInt
3588	>	(this, transformer, basis, reducer).invoke();
3589	>	}
3590	>
3591	>	/**
3592	>	* Performs the given action for each value.
3593	>	*
3594	>	* @param action the action
3595	>	*/
3596	>	public void forEachValue(Action<V> action) {
3597	>	ForkJoinTasks.forEachValue
3598	>	(this, action).invoke();
3599	>	}
3600	>
3601	>	/**
3602	>	* Performs the given action for each non-null transformation
3603	>	* of each value.
3604	>	*
3605	>	* @param transformer a function returning the transformation
3606	>	* for an element, or null of there is no transformation (in
3607	>	* which case the action is not applied).
3608	>	*/
3609	>	public <U> void forEachValue(Fun<? super V, ? extends U> transformer,
3610	>	Action<U> action) {
3611	>	ForkJoinTasks.forEachValue
3612	>	(this, transformer, action).invoke();
3613	>	}
3614	>
3615	>	/**
3616	>	* Returns a non-null result from applying the given search
3617	>	* function on each value, or null if none.  Upon success,
3618	>	* further element processing is suppressed and the results of
3619	>	* any other parallel invocations of the search function are
3620	>	* ignored.
3621	>	*
3622	>	* @param searchFunction a function returning a non-null
3623	>	* result on success, else null
3624	>	* @return a non-null result from applying the given search
3625	>	* function on each value, or null if none
3626	>	*
3627	>	*/
3628	>	public <U> U searchValues(Fun<? super V, ? extends U> searchFunction) {
3629	>	return ForkJoinTasks.searchValues
3630	>	(this, searchFunction).invoke();
3631	>	}
3632	>
3633	>	/**
3634	>	* Returns the result of accumulating all values using the
3635	>	* given reducer to combine values, or null if none.
3636	>	*
3637	>	* @param reducer a commutative associative combining function
3638	>	* @return  the result of accumulating all values
3639	>	*/
3640	>	public V reduceValues(BiFun<? super V, ? super V, ? extends V> reducer) {
3641	>	return ForkJoinTasks.reduceValues
3642	>	(this, reducer).invoke();
3643	>	}
3644	>
3645	>	/**
3646	>	* Returns the result of accumulating the given transformation
3647	>	* of all values using the given reducer to combine values, or
3648	>	* null if none.
3649	>	*
3650	>	* @param transformer a function returning the transformation
3651	>	* for an element, or null of there is no transformation (in
3652	>	* which case it is not combined).
3653	>	* @param reducer a commutative associative combining function
3654	>	* @return the result of accumulating the given transformation
3655	>	* of all values
3656	>	*/
3657	>	public <U> U reduceValues(Fun<? super V, ? extends U> transformer,
3658	>	BiFun<? super U, ? super U, ? extends U> reducer) {
3659	>	return ForkJoinTasks.reduceValues
3660	>	(this, transformer, reducer).invoke();
3661	>	}
3662	>
3663	>	/**
3664	>	* Returns the result of accumulating the given transformation
3665	>	* of all values using the given reducer to combine values,
3666	>	* and the given basis as an identity value.
3667	>	*
3668	>	* @param transformer a function returning the transformation
3669	>	* for an element
3670	>	* @param basis the identity (initial default value) for the reduction
3671	>	* @param reducer a commutative associative combining function
3672	>	* @return the result of accumulating the given transformation
3673	>	* of all values
3674	>	*/
3675	>	public double reduceValuesToDouble(ObjectToDouble<? super V> transformer,
3676	>	double basis,
3677	>	DoubleByDoubleToDouble reducer) {
3678	>	return ForkJoinTasks.reduceValuesToDouble
3679	>	(this, transformer, basis, reducer).invoke();
3680	>	}
3681	>
3682	>	/**
3683	>	* Returns the result of accumulating the given transformation
3684	>	* of all values using the given reducer to combine values,
3685	>	* and the given basis as an identity value.
3686	>	*
3687	>	* @param transformer a function returning the transformation
3688	>	* for an element
3689	>	* @param basis the identity (initial default value) for the reduction
3690	>	* @param reducer a commutative associative combining function
3691	>	* @return the result of accumulating the given transformation
3692	>	* of all values
3693	>	*/
3694	>	public long reduceValuesToLong(ObjectToLong<? super V> transformer,
3695	>	long basis,
3696	>	LongByLongToLong reducer) {
3697	>	return ForkJoinTasks.reduceValuesToLong
3698	>	(this, transformer, basis, reducer).invoke();
3699	>	}
3700	>
3701	>	/**
3702	>	* Returns the result of accumulating the given transformation
3703	>	* of all values using the given reducer to combine values,
3704	>	* and the given basis as an identity value.
3705	>	*
3706	>	* @param transformer a function returning the transformation
3707	>	* for an element
3708	>	* @param basis the identity (initial default value) for the reduction
3709	>	* @param reducer a commutative associative combining function
3710	>	* @return the result of accumulating the given transformation
3711	>	* of all values
3712	>	*/
3713	>	public int reduceValuesToInt(ObjectToInt<? super V> transformer,
3714	>	int basis,
3715	>	IntByIntToInt reducer) {
3716	>	return ForkJoinTasks.reduceValuesToInt
3717	>	(this, transformer, basis, reducer).invoke();
3718	>	}
3719	>
3720	>	/**
3721	>	* Performs the given action for each entry.
3722	>	*
3723	>	* @param action the action
3724	>	*/
3725	>	public void forEachEntry(Action<Map.Entry<K,V>> action) {
3726	>	ForkJoinTasks.forEachEntry
3727	>	(this, action).invoke();
3728	>	}
3729	>
3730	>	/**
3731	>	* Performs the given action for each non-null transformation
3732	>	* of each entry.
3733	>	*
3734	>	* @param transformer a function returning the transformation
3735	>	* for an element, or null of there is no transformation (in
3736	>	* which case the action is not applied).
3737	>	* @param action the action
3738	>	*/
3739	>	public <U> void forEachEntry(Fun<Map.Entry<K,V>, ? extends U> transformer,
3740	>	Action<U> action) {
3741	>	ForkJoinTasks.forEachEntry
3742	>	(this, transformer, action).invoke();
3743	>	}
3744	>
3745	>	/**
3746	>	* Returns a non-null result from applying the given search
3747	>	* function on each entry, or null if none.  Upon success,
3748	>	* further element processing is suppressed and the results of
3749	>	* any other parallel invocations of the search function are
3750	>	* ignored.
3751	>	*
3752	>	* @param searchFunction a function returning a non-null
3753	>	* result on success, else null
3754	>	* @return a non-null result from applying the given search
3755	>	* function on each entry, or null if none
3756	>	*/
3757	>	public <U> U searchEntries(Fun<Map.Entry<K,V>, ? extends U> searchFunction) {
3758	>	return ForkJoinTasks.searchEntries
3759	>	(this, searchFunction).invoke();
3760	>	}
3761	>
3762	>	/**
3763	>	* Returns the result of accumulating all entries using the
3764	>	* given reducer to combine values, or null if none.
3765	>	*
3766	>	* @param reducer a commutative associative combining function
3767	>	* @return the result of accumulating all entries
3768	>	*/
3769	>	public Map.Entry<K,V> reduceEntries(BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer) {
3770	>	return ForkJoinTasks.reduceEntries
3771	>	(this, reducer).invoke();
3772	>	}
3773	>
3774	>	/**
3775	>	* Returns the result of accumulating the given transformation
3776	>	* of all entries using the given reducer to combine values,
3777	>	* or null if none.
3778	>	*
3779	>	* @param transformer a function returning the transformation
3780	>	* for an element, or null of there is no transformation (in
3781	>	* which case it is not combined).
3782	>	* @param reducer a commutative associative combining function
3783	>	* @return the result of accumulating the given transformation
3784	>	* of all entries
3785	>	*/
3786	>	public <U> U reduceEntries(Fun<Map.Entry<K,V>, ? extends U> transformer,
3787	>	BiFun<? super U, ? super U, ? extends U> reducer) {
3788	>	return ForkJoinTasks.reduceEntries
3789	>	(this, transformer, reducer).invoke();
3790	>	}
3791	>
3792	>	/**
3793	>	* Returns the result of accumulating the given transformation
3794	>	* of all entries using the given reducer to combine values,
3795	>	* and the given basis as an identity value.
3796	>	*
3797	>	* @param transformer a function returning the transformation
3798	>	* for an element
3799	>	* @param basis the identity (initial default value) for the reduction
3800	>	* @param reducer a commutative associative combining function
3801	>	* @return the result of accumulating the given transformation
3802	>	* of all entries
3803	>	*/
3804	>	public double reduceEntriesToDouble(ObjectToDouble<Map.Entry<K,V>> transformer,
3805	>	double basis,
3806	>	DoubleByDoubleToDouble reducer) {
3807	>	return ForkJoinTasks.reduceEntriesToDouble
3808	>	(this, transformer, basis, reducer).invoke();
3809	>	}
3810	>
3811	>	/**
3812	>	* Returns the result of accumulating the given transformation
3813	>	* of all entries using the given reducer to combine values,
3814	>	* and the given basis as an identity value.
3815	>	*
3816	>	* @param transformer a function returning the transformation
3817	>	* for an element
3818	>	* @param basis the identity (initial default value) for the reduction
3819	>	* @param reducer a commutative associative combining function
3820	>	* @return  the result of accumulating the given transformation
3821	>	* of all entries
3822	>	*/
3823	>	public long reduceEntriesToLong(ObjectToLong<Map.Entry<K,V>> transformer,
3824	>	long basis,
3825	>	LongByLongToLong reducer) {
3826	>	return ForkJoinTasks.reduceEntriesToLong
3827	>	(this, transformer, basis, reducer).invoke();
3828	>	}
3829	>
3830	>	/**
3831	>	* Returns the result of accumulating the given transformation
3832	>	* of all entries using the given reducer to combine values,
3833	>	* and the given basis as an identity value.
3834	>	*
3835	>	* @param transformer a function returning the transformation
3836	>	* for an element
3837	>	* @param basis the identity (initial default value) for the reduction
3838	>	* @param reducer a commutative associative combining function
3839	>	* @return the result of accumulating the given transformation
3840	>	* of all entries
3841	>	*/
3842	>	public int reduceEntriesToInt(ObjectToInt<Map.Entry<K,V>> transformer,
3843	>	int basis,
3844	>	IntByIntToInt reducer) {
3845	>	return ForkJoinTasks.reduceEntriesToInt
3846	>	(this, transformer, basis, reducer).invoke();
3847	>	}
3848	>
3849	>	/* ----------------Views -------------- */
3850	>
3851	>	/**
3852	>	* Base class for views.
3853	>	*/
3854	>	static abstract class CHMView<K, V> {
3855	>	final ConcurrentHashMapV8<K, V> map;
3856	>	CHMView(ConcurrentHashMapV8<K, V> map)  { this.map = map; }
3857	>
3858	>	/**
3859	>	* Returns the map backing this view.
3860	>	*
3861	>	* @return the map backing this view
3862	>	*/
3863	>	public ConcurrentHashMapV8<K,V> getMap() { return map; }
3864	>
3865	>	public final int size()                 { return map.size(); }
3866	>	public final boolean isEmpty()          { return map.isEmpty(); }
3867	>	public final void clear()               { map.clear(); }
3868	>
3869	>	// implementations below rely on concrete classes supplying these
3870	>	abstract public Iterator<?> iterator();
3871	>	abstract public boolean contains(Object o);
3872	>	abstract public boolean remove(Object o);
3873	>
3874	>	private static final String oomeMsg = "Required array size too large";
3875	>
3876	>	public final Object[] toArray() {
3877	>	long sz = map.mappingCount();
3878	>	if (sz > (long)(MAX_ARRAY_SIZE))
3879	>	throw new OutOfMemoryError(oomeMsg);
3880	>	int n = (int)sz;
3881	>	Object[] r = new Object[n];
3882	>	int i = 0;
3883	>	Iterator<?> it = iterator();
3884	>	while (it.hasNext()) {
3885	>	if (i == n) {
3886	>	if (n >= MAX_ARRAY_SIZE)
3887	>	throw new OutOfMemoryError(oomeMsg);
3888	>	if (n >= MAX_ARRAY_SIZE - (MAX_ARRAY_SIZE >>> 1) - 1)
3889	>	n = MAX_ARRAY_SIZE;
3890	>	else
3891	>	n += (n >>> 1) + 1;
3892	>	r = Arrays.copyOf(r, n);
3893	>	}
3894	>	r[i++] = it.next();
3895	>	}
3896	>	return (i == n) ? r : Arrays.copyOf(r, i);
3897	>	}
3898	>
3899	>	@SuppressWarnings("unchecked") public final <T> T[] toArray(T[] a) {
3900	>	long sz = map.mappingCount();
3901	>	if (sz > (long)(MAX_ARRAY_SIZE))
3902	>	throw new OutOfMemoryError(oomeMsg);
3903	>	int m = (int)sz;
3904	>	T[] r = (a.length >= m) ? a :
3905	>	(T[])java.lang.reflect.Array
3906	>	.newInstance(a.getClass().getComponentType(), m);
3907	>	int n = r.length;
3908	>	int i = 0;
3909	>	Iterator<?> it = iterator();
3910	>	while (it.hasNext()) {
3911	>	if (i == n) {
3912	>	if (n >= MAX_ARRAY_SIZE)
3913	>	throw new OutOfMemoryError(oomeMsg);
3914	>	if (n >= MAX_ARRAY_SIZE - (MAX_ARRAY_SIZE >>> 1) - 1)
3915	>	n = MAX_ARRAY_SIZE;
3916	>	else
3917	>	n += (n >>> 1) + 1;
3918	>	r = Arrays.copyOf(r, n);
3919	>	}
3920	>	r[i++] = (T)it.next();
3921	>	}
3922	>	if (a == r && i < n) {
3923	>	r[i] = null; // null-terminate
3924	>	return r;
3925	>	}
3926	>	return (i == n) ? r : Arrays.copyOf(r, i);
3927	>	}
3928	>
3929	>	public final int hashCode() {
3930	>	int h = 0;
3931	>	for (Iterator<?> it = iterator(); it.hasNext();)
3932	>	h += it.next().hashCode();
3933	>	return h;
3934	>	}
3935	>
3936	>	public final String toString() {
3937	>	StringBuilder sb = new StringBuilder();
3938	>	sb.append('[');
3939	>	Iterator<?> it = iterator();
3940	>	if (it.hasNext()) {
3941	>	for (;;) {
3942	>	Object e = it.next();
3943	>	sb.append(e == this ? "(this Collection)" : e);
3944	>	if (!it.hasNext())
3945	>	break;
3946	>	sb.append(',').append(' ');
3947	>	}
3948	>	}
3949	>	return sb.append(']').toString();
3950	>	}
3951	>
3952	>	public final boolean containsAll(Collection<?> c) {
3953	>	if (c != this) {
3954	>	for (Iterator<?> it = c.iterator(); it.hasNext();) {
3955	>	Object e = it.next();
3956	>	if (e == null || !contains(e))
3957	>	return false;
3958	>	}
3959	>	}
3960	>	return true;
3961	>	}
3962	>
3963	>	public final boolean removeAll(Collection<?> c) {
3964	>	boolean modified = false;
3965	>	for (Iterator<?> it = iterator(); it.hasNext();) {
3966	>	if (c.contains(it.next())) {
3967	>	it.remove();
3968	>	modified = true;
3969	>	}
3970	>	}
3971	>	return modified;
3972	>	}
3973	>
3974	>	public final boolean retainAll(Collection<?> c) {
3975	>	boolean modified = false;
3976	>	for (Iterator<?> it = iterator(); it.hasNext();) {
3977	>	if (!c.contains(it.next())) {
3978	>	it.remove();
3979	>	modified = true;
3980	>	}
3981	>	}
3982	>	return modified;
3983	>	}
3984	>
3985	>	}
3986	>
3987	>	/**
3988	>	* A view of a ConcurrentHashMapV8 as a {@link Set} of keys, in
3989	>	* which additions may optionally be enabled by mapping to a
3990	>	* common value.  This class cannot be directly instantiated. See
3991	>	* {@link #keySet}, {@link #keySet(Object)}, {@link #newKeySet()},
3992	>	* {@link #newKeySet(int)}.
3993	>	*/
3994	>	public static class KeySetView<K,V> extends CHMView<K,V>
3995	>	implements Set<K>, java.io.Serializable {
3996	>	private static final long serialVersionUID = 7249069246763182397L;
3997	>	private final V value;
3998	>	KeySetView(ConcurrentHashMapV8<K, V> map, V value) {  // non-public
3999	>	super(map);
4000	>	this.value = value;
4001	>	}
4002	>
4003	>	/**
4004	>	* Returns the default mapped value for additions,
4005	>	* or {@code null} if additions are not supported.
4006	>	*
4007	>	* @return the default mapped value for additions, or {@code null}
4008	>	* if not supported.
4009	>	*/
4010	>	public V getMappedValue() { return value; }
4011	>
4012	>	// implement Set API
4013	>
4014	>	public boolean contains(Object o) { return map.containsKey(o); }
4015	>	public boolean remove(Object o)   { return map.remove(o) != null; }
4016	>
4017	>	/**
4018	>	* Returns a "weakly consistent" iterator that will never
4019	>	* throw {@link ConcurrentModificationException}, and
4020	>	* guarantees to traverse elements as they existed upon
4021	>	* construction of the iterator, and may (but is not
4022	>	* guaranteed to) reflect any modifications subsequent to
4023	>	* construction.
4024	>	*
4025	>	* @return an iterator over the keys of this map
4026	>	*/
4027	>	public Iterator<K> iterator()     { return new KeyIterator<K,V>(map); }
4028	>	public boolean add(K e) {
4029	>	V v;
4030	>	if ((v = value) == null)
4031	>	throw new UnsupportedOperationException();
4032	>	if (e == null)
4033	>	throw new NullPointerException();
4034	>	return map.internalPut(e, v, true) == null;
4035	>	}
4036	>	public boolean addAll(Collection<? extends K> c) {
4037	>	boolean added = false;
4038	>	V v;
4039	>	if ((v = value) == null)
4040	>	throw new UnsupportedOperationException();
4041	>	for (K e : c) {
4042	>	if (e == null)
4043	>	throw new NullPointerException();
4044	>	if (map.internalPut(e, v, true) == null)
4045	>	added = true;
4046	>	}
4047	>	return added;
4048	>	}
4049	>	public boolean equals(Object o) {
4050	>	Set<?> c;
4051	>	return ((o instanceof Set) &&
4052	>	((c = (Set<?>)o) == this ||
4053	>	(containsAll(c) && c.containsAll(this))));
4054	>	}
4055	>
4056	>	/**
4057	>	* Performs the given action for each key.
4058	>	*
4059	>	* @param action the action
4060	>	*/
4061	>	public void forEach(Action<K> action) {
4062	>	ForkJoinTasks.forEachKey
4063	>	(map, action).invoke();
4064	>	}
4065	>
4066	>	/**
4067	>	* Performs the given action for each non-null transformation
4068	>	* of each key.
4069	>	*
4070	>	* @param transformer a function returning the transformation
4071	>	* for an element, or null of there is no transformation (in
4072	>	* which case the action is not applied).
4073	>	* @param action the action
4074	>	*/
4075	>	public <U> void forEach(Fun<? super K, ? extends U> transformer,
4076	>	Action<U> action) {
4077	>	ForkJoinTasks.forEachKey
4078	>	(map, transformer, action).invoke();
4079	>	}
4080	>
4081	>	/**
4082	>	* Returns a non-null result from applying the given search
4083	>	* function on each key, or null if none. Upon success,
4084	>	* further element processing is suppressed and the results of
4085	>	* any other parallel invocations of the search function are
4086	>	* ignored.
4087	>	*
4088	>	* @param searchFunction a function returning a non-null
4089	>	* result on success, else null
4090	>	* @return a non-null result from applying the given search
4091	>	* function on each key, or null if none
4092	>	*/
4093	>	public <U> U search(Fun<? super K, ? extends U> searchFunction) {
4094	>	return ForkJoinTasks.searchKeys
4095	>	(map, searchFunction).invoke();
4096	>	}
4097	>
4098	>	/**
4099	>	* Returns the result of accumulating all keys using the given
4100	>	* reducer to combine values, or null if none.
4101	>	*
4102	>	* @param reducer a commutative associative combining function
4103	>	* @return the result of accumulating all keys using the given
4104	>	* reducer to combine values, or null if none
4105	>	*/
4106	>	public K reduce(BiFun<? super K, ? super K, ? extends K> reducer) {
4107	>	return ForkJoinTasks.reduceKeys
4108	>	(map, reducer).invoke();
4109	>	}
4110	>
4111	>	/**
4112	>	* Returns the result of accumulating the given transformation
4113	>	* of all keys using the given reducer to combine values, and
4114	>	* the given basis as an identity value.
4115	>	*
4116	>	* @param transformer a function returning the transformation
4117	>	* for an element
4118	>	* @param basis the identity (initial default value) for the reduction
4119	>	* @param reducer a commutative associative combining function
4120	>	* @return  the result of accumulating the given transformation
4121	>	* of all keys
4122	>	*/
4123	>	public double reduceToDouble(ObjectToDouble<? super K> transformer,
4124	>	double basis,
4125	>	DoubleByDoubleToDouble reducer) {
4126	>	return ForkJoinTasks.reduceKeysToDouble
4127	>	(map, transformer, basis, reducer).invoke();
4128	>	}
4129	>
4130	>	/**
4131	>	* Returns the result of accumulating the given transformation
4132	>	* of all keys using the given reducer to combine values, and
4133	>	* the given basis as an identity value.
4134	>	*
4135	>	* @param transformer a function returning the transformation
4136	>	* for an element
4137	>	* @param basis the identity (initial default value) for the reduction
4138	>	* @param reducer a commutative associative combining function
4139	>	* @return the result of accumulating the given transformation
4140	>	* of all keys
4141	>	*/
4142	>	public long reduceToLong(ObjectToLong<? super K> transformer,
4143	>	long basis,
4144	>	LongByLongToLong reducer) {
4145	>	return ForkJoinTasks.reduceKeysToLong
4146	>	(map, transformer, basis, reducer).invoke();
4147	>	}
4148	>
4149	>	/**
4150	>	* Returns the result of accumulating the given transformation
4151	>	* of all keys using the given reducer to combine values, and
4152	>	* the given basis as an identity value.
4153	>	*
4154	>	* @param transformer a function returning the transformation
4155	>	* for an element
4156	>	* @param basis the identity (initial default value) for the reduction
4157	>	* @param reducer a commutative associative combining function
4158	>	* @return the result of accumulating the given transformation
4159	>	* of all keys
4160	>	*/
4161	>	public int reduceToInt(ObjectToInt<? super K> transformer,
4162	>	int basis,
4163	>	IntByIntToInt reducer) {
4164	>	return ForkJoinTasks.reduceKeysToInt
4165	>	(map, transformer, basis, reducer).invoke();
4166	>	}
4167	>
4168	>	}
4169	>
4170	>	/**
4171	>	* A view of a ConcurrentHashMapV8 as a {@link Collection} of
4172	>	* values, in which additions are disabled. This class cannot be
4173	>	* directly instantiated. See {@link #values},
4174	>	*
4175	>	* <p>The view's {@code iterator} is a "weakly consistent" iterator
4176	>	* that will never throw {@link ConcurrentModificationException},
4177	>	* and guarantees to traverse elements as they existed upon
4178	>	* construction of the iterator, and may (but is not guaranteed to)
4179	>	* reflect any modifications subsequent to construction.
4180	>	*/
4181	>	public static final class ValuesView<K,V> extends CHMView<K,V>
4182	>	implements Collection<V> {
4183	>	ValuesView(ConcurrentHashMapV8<K, V> map)   { super(map); }
4184	>	public final boolean contains(Object o) { return map.containsValue(o); }
4185	>	public final boolean remove(Object o) {
4186	>	if (o != null) {
4187	>	Iterator<V> it = new ValueIterator<K,V>(map);
4188	>	while (it.hasNext()) {
4189	>	if (o.equals(it.next())) {
4190	>	it.remove();
4191	>	return true;
4192	>	}
4193	>	}
4194	>	}
4195	>	return false;
4196	>	}
4197	>
4198	>	/**
4199	>	* Returns a "weakly consistent" iterator that will never
4200	>	* throw {@link ConcurrentModificationException}, and
4201	>	* guarantees to traverse elements as they existed upon
4202	>	* construction of the iterator, and may (but is not
4203	>	* guaranteed to) reflect any modifications subsequent to
4204	>	* construction.
4205	>	*
4206	>	* @return an iterator over the values of this map
4207	>	*/
4208	>	public final Iterator<V> iterator() {
4209	>	return new ValueIterator<K,V>(map);
4210	>	}
4211	>	public final boolean add(V e) {
4212	>	throw new UnsupportedOperationException();
4213	>	}
4214	>	public final boolean addAll(Collection<? extends V> c) {
4215	>	throw new UnsupportedOperationException();
4216	>	}
4217	>
4218	>	/**
4219	>	* Performs the given action for each value.
4220	>	*
4221	>	* @param action the action
4222	>	*/
4223	>	public void forEach(Action<V> action) {
4224	>	ForkJoinTasks.forEachValue
4225	>	(map, action).invoke();
4226	>	}
4227	>
4228	>	/**
4229	>	* Performs the given action for each non-null transformation
4230	>	* of each value.
4231	>	*
4232	>	* @param transformer a function returning the transformation
4233	>	* for an element, or null of there is no transformation (in
4234	>	* which case the action is not applied).
4235	>	*/
4236	>	public <U> void forEach(Fun<? super V, ? extends U> transformer,
4237	>	Action<U> action) {
4238	>	ForkJoinTasks.forEachValue
4239	>	(map, transformer, action).invoke();
4240	>	}
4241	>
4242	>	/**
4243	>	* Returns a non-null result from applying the given search
4244	>	* function on each value, or null if none.  Upon success,
4245	>	* further element processing is suppressed and the results of
4246	>	* any other parallel invocations of the search function are
4247	>	* ignored.
4248	>	*
4249	>	* @param searchFunction a function returning a non-null
4250	>	* result on success, else null
4251	>	* @return a non-null result from applying the given search
4252	>	* function on each value, or null if none
4253	>	*
4254	>	*/
4255	>	public <U> U search(Fun<? super V, ? extends U> searchFunction) {
4256	>	return ForkJoinTasks.searchValues
4257	>	(map, searchFunction).invoke();
4258	>	}
4259	>
4260	>	/**
4261	>	* Returns the result of accumulating all values using the
4262	>	* given reducer to combine values, or null if none.
4263	>	*
4264	>	* @param reducer a commutative associative combining function
4265	>	* @return  the result of accumulating all values
4266	>	*/
4267	>	public V reduce(BiFun<? super V, ? super V, ? extends V> reducer) {
4268	>	return ForkJoinTasks.reduceValues
4269	>	(map, reducer).invoke();
4270	>	}
4271	>
4272	>	/**
4273	>	* Returns the result of accumulating the given transformation
4274	>	* of all values using the given reducer to combine values, or
4275	>	* null if none.
4276	>	*
4277	>	* @param transformer a function returning the transformation
4278	>	* for an element, or null of there is no transformation (in
4279	>	* which case it is not combined).
4280	>	* @param reducer a commutative associative combining function
4281	>	* @return the result of accumulating the given transformation
4282	>	* of all values
4283	>	*/
4284	>	public <U> U reduce(Fun<? super V, ? extends U> transformer,
4285	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4286	>	return ForkJoinTasks.reduceValues
4287	>	(map, transformer, reducer).invoke();
4288	>	}
4289	>
4290	>	/**
4291	>	* Returns the result of accumulating the given transformation
4292	>	* of all values using the given reducer to combine values,
4293	>	* and the given basis as an identity value.
4294	>	*
4295	>	* @param transformer a function returning the transformation
4296	>	* for an element
4297	>	* @param basis the identity (initial default value) for the reduction
4298	>	* @param reducer a commutative associative combining function
4299	>	* @return the result of accumulating the given transformation
4300	>	* of all values
4301	>	*/
4302	>	public double reduceToDouble(ObjectToDouble<? super V> transformer,
4303	>	double basis,
4304	>	DoubleByDoubleToDouble reducer) {
4305	>	return ForkJoinTasks.reduceValuesToDouble
4306	>	(map, transformer, basis, reducer).invoke();
4307	>	}
4308	>
4309	>	/**
4310	>	* Returns the result of accumulating the given transformation
4311	>	* of all values using the given reducer to combine values,
4312	>	* and the given basis as an identity value.
4313	>	*
4314	>	* @param transformer a function returning the transformation
4315	>	* for an element
4316	>	* @param basis the identity (initial default value) for the reduction
4317	>	* @param reducer a commutative associative combining function
4318	>	* @return the result of accumulating the given transformation
4319	>	* of all values
4320	>	*/
4321	>	public long reduceToLong(ObjectToLong<? super V> transformer,
4322	>	long basis,
4323	>	LongByLongToLong reducer) {
4324	>	return ForkJoinTasks.reduceValuesToLong
4325	>	(map, transformer, basis, reducer).invoke();
4326	>	}
4327	>
4328	>	/**
4329	>	* Returns the result of accumulating the given transformation
4330	>	* of all values using the given reducer to combine values,
4331	>	* and the given basis as an identity value.
4332	>	*
4333	>	* @param transformer a function returning the transformation
4334	>	* for an element
4335	>	* @param basis the identity (initial default value) for the reduction
4336	>	* @param reducer a commutative associative combining function
4337	>	* @return the result of accumulating the given transformation
4338	>	* of all values
4339	>	*/
4340	>	public int reduceToInt(ObjectToInt<? super V> transformer,
4341	>	int basis,
4342	>	IntByIntToInt reducer) {
4343	>	return ForkJoinTasks.reduceValuesToInt
4344	>	(map, transformer, basis, reducer).invoke();
4345	>	}
4346	>
4347	>	}
4348	>
4349	>	/**
4350	>	* A view of a ConcurrentHashMapV8 as a {@link Set} of (key, value)
4351	>	* entries.  This class cannot be directly instantiated. See
4352	>	* {@link #entrySet}.
4353	>	*/
4354	>	public static final class EntrySetView<K,V> extends CHMView<K,V>
4355	>	implements Set<Map.Entry<K,V>> {
4356	>	EntrySetView(ConcurrentHashMapV8<K, V> map) { super(map); }
4357	>	public final boolean contains(Object o) {
4358	>	Object k, v, r; Map.Entry<?,?> e;
4359	>	return ((o instanceof Map.Entry) &&
4360	>	(k = (e = (Map.Entry<?,?>)o).getKey()) != null &&
4361	>	(r = map.get(k)) != null &&
4362	>	(v = e.getValue()) != null &&
4363	>	(v == r || v.equals(r)));
4364	>	}
4365	>	public final boolean remove(Object o) {
4366	>	Object k, v; Map.Entry<?,?> e;
4367	>	return ((o instanceof Map.Entry) &&
4368	>	(k = (e = (Map.Entry<?,?>)o).getKey()) != null &&
4369	>	(v = e.getValue()) != null &&
4370	>	map.remove(k, v));
4371	>	}
4372	>
4373	>	/**
4374	>	* Returns a "weakly consistent" iterator that will never
4375	>	* throw {@link ConcurrentModificationException}, and
4376	>	* guarantees to traverse elements as they existed upon
4377	>	* construction of the iterator, and may (but is not
4378	>	* guaranteed to) reflect any modifications subsequent to
4379	>	* construction.
4380	>	*
4381	>	* @return an iterator over the entries of this map
4382	>	*/
4383	>	public final Iterator<Map.Entry<K,V>> iterator() {
4384	>	return new EntryIterator<K,V>(map);
4385	>	}
4386	>
4387	>	public final boolean add(Entry<K,V> e) {
4388	>	K key = e.getKey();
4389	>	V value = e.getValue();
4390	>	if (key == null || value == null)
4391	>	throw new NullPointerException();
4392	>	return map.internalPut(key, value, false) == null;
4393	>	}
4394	>	public final boolean addAll(Collection<? extends Entry<K,V>> c) {
4395	>	boolean added = false;
4396	>	for (Entry<K,V> e : c) {
4397	>	if (add(e))
4398	>	added = true;
4399	>	}
4400	>	return added;
4401	>	}
4402	>	public boolean equals(Object o) {
4403	>	Set<?> c;
4404	>	return ((o instanceof Set) &&
4405	>	((c = (Set<?>)o) == this ||
4406	>	(containsAll(c) && c.containsAll(this))));
4407	>	}
4408	>
4409	>	/**
4410	>	* Performs the given action for each entry.
4411	>	*
4412	>	* @param action the action
4413	>	*/
4414	>	public void forEach(Action<Map.Entry<K,V>> action) {
4415	>	ForkJoinTasks.forEachEntry
4416	>	(map, action).invoke();
4417	>	}
4418	>
4419	>	/**
4420	>	* Performs the given action for each non-null transformation
4421	>	* of each entry.
4422	>	*
4423	>	* @param transformer a function returning the transformation
4424	>	* for an element, or null of there is no transformation (in
4425	>	* which case the action is not applied).
4426	>	* @param action the action
4427	>	*/
4428	>	public <U> void forEach(Fun<Map.Entry<K,V>, ? extends U> transformer,
4429	>	Action<U> action) {
4430	>	ForkJoinTasks.forEachEntry
4431	>	(map, transformer, action).invoke();
4432	>	}
4433	>
4434	>	/**
4435	>	* Returns a non-null result from applying the given search
4436	>	* function on each entry, or null if none.  Upon success,
4437	>	* further element processing is suppressed and the results of
4438	>	* any other parallel invocations of the search function are
4439	>	* ignored.
4440	>	*
4441	>	* @param searchFunction a function returning a non-null
4442	>	* result on success, else null
4443	>	* @return a non-null result from applying the given search
4444	>	* function on each entry, or null if none
4445	>	*/
4446	>	public <U> U search(Fun<Map.Entry<K,V>, ? extends U> searchFunction) {
4447	>	return ForkJoinTasks.searchEntries
4448	>	(map, searchFunction).invoke();
4449	>	}
4450	>
4451	>	/**
4452	>	* Returns the result of accumulating all entries using the
4453	>	* given reducer to combine values, or null if none.
4454	>	*
4455	>	* @param reducer a commutative associative combining function
4456	>	* @return the result of accumulating all entries
4457	>	*/
4458	>	public Map.Entry<K,V> reduce(BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer) {
4459	>	return ForkJoinTasks.reduceEntries
4460	>	(map, reducer).invoke();
4461	>	}
4462	>
4463	>	/**
4464	>	* Returns the result of accumulating the given transformation
4465	>	* of all entries using the given reducer to combine values,
4466	>	* or null if none.
4467	>	*
4468	>	* @param transformer a function returning the transformation
4469	>	* for an element, or null of there is no transformation (in
4470	>	* which case it is not combined).
4471	>	* @param reducer a commutative associative combining function
4472	>	* @return the result of accumulating the given transformation
4473	>	* of all entries
4474	>	*/
4475	>	public <U> U reduce(Fun<Map.Entry<K,V>, ? extends U> transformer,
4476	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4477	>	return ForkJoinTasks.reduceEntries
4478	>	(map, transformer, reducer).invoke();
4479	>	}
4480	>
4481	>	/**
4482	>	* Returns the result of accumulating the given transformation
4483	>	* of all entries using the given reducer to combine values,
4484	>	* and the given basis as an identity value.
4485	>	*
4486	>	* @param transformer a function returning the transformation
4487	>	* for an element
4488	>	* @param basis the identity (initial default value) for the reduction
4489	>	* @param reducer a commutative associative combining function
4490	>	* @return the result of accumulating the given transformation
4491	>	* of all entries
4492	>	*/
4493	>	public double reduceToDouble(ObjectToDouble<Map.Entry<K,V>> transformer,
4494	>	double basis,
4495	>	DoubleByDoubleToDouble reducer) {
4496	>	return ForkJoinTasks.reduceEntriesToDouble
4497	>	(map, transformer, basis, reducer).invoke();
4498	>	}
4499	>
4500	>	/**
4501	>	* Returns the result of accumulating the given transformation
4502	>	* of all entries using the given reducer to combine values,
4503	>	* and the given basis as an identity value.
4504	>	*
4505	>	* @param transformer a function returning the transformation
4506	>	* for an element
4507	>	* @param basis the identity (initial default value) for the reduction
4508	>	* @param reducer a commutative associative combining function
4509	>	* @return  the result of accumulating the given transformation
4510	>	* of all entries
4511	>	*/
4512	>	public long reduceToLong(ObjectToLong<Map.Entry<K,V>> transformer,
4513	>	long basis,
4514	>	LongByLongToLong reducer) {
4515	>	return ForkJoinTasks.reduceEntriesToLong
4516	>	(map, transformer, basis, reducer).invoke();
4517	>	}
4518	>
4519	>	/**
4520	>	* Returns the result of accumulating the given transformation
4521	>	* of all entries using the given reducer to combine values,
4522	>	* and the given basis as an identity value.
4523	>	*
4524	>	* @param transformer a function returning the transformation
4525	>	* for an element
4526	>	* @param basis the identity (initial default value) for the reduction
4527	>	* @param reducer a commutative associative combining function
4528	>	* @return the result of accumulating the given transformation
4529	>	* of all entries
4530	>	*/
4531	>	public int reduceToInt(ObjectToInt<Map.Entry<K,V>> transformer,
4532	>	int basis,
4533	>	IntByIntToInt reducer) {
4534	>	return ForkJoinTasks.reduceEntriesToInt
4535	>	(map, transformer, basis, reducer).invoke();
4536	>	}
4537	>
4538	>	}
4539	>
4540	>	// ---------------------------------------------------------------------
4541	>
4542	>	/**
4543	>	* Predefined tasks for performing bulk parallel operations on
4544	>	* ConcurrentHashMapV8s. These tasks follow the forms and rules used
4545	>	* for bulk operations. Each method has the same name, but returns
4546	>	* a task rather than invoking it. These methods may be useful in
4547	>	* custom applications such as submitting a task without waiting
4548	>	* for completion, using a custom pool, or combining with other
4549	>	* tasks.
4550	>	*/
4551	>	public static class ForkJoinTasks {
4552	>	private ForkJoinTasks() {}
4553	>
4554	>	/**
4555	>	* Returns a task that when invoked, performs the given
4556	>	* action for each (key, value)
4557	>	*
4558	>	* @param map the map
4559	>	* @param action the action
4560	>	* @return the task
4561	>	*/
4562	>	public static <K,V> ForkJoinTask<Void> forEach
4563	>	(ConcurrentHashMapV8<K,V> map,
4564	>	BiAction<K,V> action) {
4565	>	if (action == null) throw new NullPointerException();
4566	>	return new ForEachMappingTask<K,V>(map, null, -1, action);
4567	>	}
4568	>
4569	>	/**
4570	>	* Returns a task that when invoked, performs the given
4571	>	* action for each non-null transformation of each (key, value)
4572	>	*
4573	>	* @param map the map
4574	>	* @param transformer a function returning the transformation
4575	>	* for an element, or null if there is no transformation (in
4576	>	* which case the action is not applied)
4577	>	* @param action the action
4578	>	* @return the task
4579	>	*/
4580	>	public static <K,V,U> ForkJoinTask<Void> forEach
4581	>	(ConcurrentHashMapV8<K,V> map,
4582	>	BiFun<? super K, ? super V, ? extends U> transformer,
4583	>	Action<U> action) {
4584	>	if (transformer == null || action == null)
4585	>	throw new NullPointerException();
4586	>	return new ForEachTransformedMappingTask<K,V,U>
4587	>	(map, null, -1, transformer, action);
4588	>	}
4589	>
4590	>	/**
4591	>	* Returns a task that when invoked, returns a non-null result
4592	>	* from applying the given search function on each (key,
4593	>	* value), or null if none. Upon success, further element
4594	>	* processing is suppressed and the results of any other
4595	>	* parallel invocations of the search function are ignored.
4596	>	*
4597	>	* @param map the map
4598	>	* @param searchFunction a function returning a non-null
4599	>	* result on success, else null
4600	>	* @return the task
4601	>	*/
4602	>	public static <K,V,U> ForkJoinTask<U> search
4603	>	(ConcurrentHashMapV8<K,V> map,
4604	>	BiFun<? super K, ? super V, ? extends U> searchFunction) {
4605	>	if (searchFunction == null) throw new NullPointerException();
4606	>	return new SearchMappingsTask<K,V,U>
4607	>	(map, null, -1, searchFunction,
4608	>	new AtomicReference<U>());
4609	>	}
4610	>
4611	>	/**
4612	>	* Returns a task that when invoked, returns the result of
4613	>	* accumulating the given transformation of all (key, value) pairs
4614	>	* using the given reducer to combine values, or null if none.
4615	>	*
4616	>	* @param map the map
4617	>	* @param transformer a function returning the transformation
4618	>	* for an element, or null if there is no transformation (in
4619	>	* which case it is not combined).
4620	>	* @param reducer a commutative associative combining function
4621	>	* @return the task
4622	>	*/
4623	>	public static <K,V,U> ForkJoinTask<U> reduce
4624	>	(ConcurrentHashMapV8<K,V> map,
4625	>	BiFun<? super K, ? super V, ? extends U> transformer,
4626	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4627	>	if (transformer == null || reducer == null)
4628	>	throw new NullPointerException();
4629	>	return new MapReduceMappingsTask<K,V,U>
4630	>	(map, null, -1, null, transformer, reducer);
4631	>	}
4632	>
4633	>	/**
4634	>	* Returns a task that when invoked, returns the result of
4635	>	* accumulating the given transformation of all (key, value) pairs
4636	>	* using the given reducer to combine values, and the given
4637	>	* basis as an identity value.
4638	>	*
4639	>	* @param map the map
4640	>	* @param transformer a function returning the transformation
4641	>	* for an element
4642	>	* @param basis the identity (initial default value) for the reduction
4643	>	* @param reducer a commutative associative combining function
4644	>	* @return the task
4645	>	*/
4646	>	public static <K,V> ForkJoinTask<Double> reduceToDouble
4647	>	(ConcurrentHashMapV8<K,V> map,
4648	>	ObjectByObjectToDouble<? super K, ? super V> transformer,
4649	>	double basis,
4650	>	DoubleByDoubleToDouble reducer) {
4651	>	if (transformer == null || reducer == null)
4652	>	throw new NullPointerException();
4653	>	return new MapReduceMappingsToDoubleTask<K,V>
4654	>	(map, null, -1, null, transformer, basis, reducer);
4655	>	}
4656	>
4657	>	/**
4658	>	* Returns a task that when invoked, returns the result of
4659	>	* accumulating the given transformation of all (key, value) pairs
4660	>	* using the given reducer to combine values, and the given
4661	>	* basis as an identity value.
4662	>	*
4663	>	* @param map the map
4664	>	* @param transformer a function returning the transformation
4665	>	* for an element
4666	>	* @param basis the identity (initial default value) for the reduction
4667	>	* @param reducer a commutative associative combining function
4668	>	* @return the task
4669	>	*/
4670	>	public static <K,V> ForkJoinTask<Long> reduceToLong
4671	>	(ConcurrentHashMapV8<K,V> map,
4672	>	ObjectByObjectToLong<? super K, ? super V> transformer,
4673	>	long basis,
4674	>	LongByLongToLong reducer) {
4675	>	if (transformer == null || reducer == null)
4676	>	throw new NullPointerException();
4677	>	return new MapReduceMappingsToLongTask<K,V>
4678	>	(map, null, -1, null, transformer, basis, reducer);
4679	>	}
4680	>
4681	>	/**
4682	>	* Returns a task that when invoked, returns the result of
4683	>	* accumulating the given transformation of all (key, value) pairs
4684	>	* using the given reducer to combine values, and the given
4685	>	* basis as an identity value.
4686	>	*
4687	>	* @param transformer a function returning the transformation
4688	>	* for an element
4689	>	* @param basis the identity (initial default value) for the reduction
4690	>	* @param reducer a commutative associative combining function
4691	>	* @return the task
4692	>	*/
4693	>	public static <K,V> ForkJoinTask<Integer> reduceToInt
4694	>	(ConcurrentHashMapV8<K,V> map,
4695	>	ObjectByObjectToInt<? super K, ? super V> transformer,
4696	>	int basis,
4697	>	IntByIntToInt reducer) {
4698	>	if (transformer == null || reducer == null)
4699	>	throw new NullPointerException();
4700	>	return new MapReduceMappingsToIntTask<K,V>
4701	>	(map, null, -1, null, transformer, basis, reducer);
4702	>	}
4703	>
4704	>	/**
4705	>	* Returns a task that when invoked, performs the given action
4706	>	* for each key.
4707	>	*
4708	>	* @param map the map
4709	>	* @param action the action
4710	>	* @return the task
4711	>	*/
4712	>	public static <K,V> ForkJoinTask<Void> forEachKey
4713	>	(ConcurrentHashMapV8<K,V> map,
4714	>	Action<K> action) {
4715	>	if (action == null) throw new NullPointerException();
4716	>	return new ForEachKeyTask<K,V>(map, null, -1, action);
4717	>	}
4718	>
4719	>	/**
4720	>	* Returns a task that when invoked, performs the given action
4721	>	* for each non-null transformation of each key.
4722	>	*
4723	>	* @param map the map
4724	>	* @param transformer a function returning the transformation
4725	>	* for an element, or null if there is no transformation (in
4726	>	* which case the action is not applied)
4727	>	* @param action the action
4728	>	* @return the task
4729	>	*/
4730	>	public static <K,V,U> ForkJoinTask<Void> forEachKey
4731	>	(ConcurrentHashMapV8<K,V> map,
4732	>	Fun<? super K, ? extends U> transformer,
4733	>	Action<U> action) {
4734	>	if (transformer == null || action == null)
4735	>	throw new NullPointerException();
4736	>	return new ForEachTransformedKeyTask<K,V,U>
4737	>	(map, null, -1, transformer, action);
4738	>	}
4739	>
4740	>	/**
4741	>	* Returns a task that when invoked, returns a non-null result
4742	>	* from applying the given search function on each key, or
4743	>	* null if none.  Upon success, further element processing is
4744	>	* suppressed and the results of any other parallel
4745	>	* invocations of the search function are ignored.
4746	>	*
4747	>	* @param map the map
4748	>	* @param searchFunction a function returning a non-null
4749	>	* result on success, else null
4750	>	* @return the task
4751	>	*/
4752	>	public static <K,V,U> ForkJoinTask<U> searchKeys
4753	>	(ConcurrentHashMapV8<K,V> map,
4754	>	Fun<? super K, ? extends U> searchFunction) {
4755	>	if (searchFunction == null) throw new NullPointerException();
4756	>	return new SearchKeysTask<K,V,U>
4757	>	(map, null, -1, searchFunction,
4758	>	new AtomicReference<U>());
4759	>	}
4760	>
4761	>	/**
4762	>	* Returns a task that when invoked, returns the result of
4763	>	* accumulating all keys using the given reducer to combine
4764	>	* values, or null if none.
4765	>	*
4766	>	* @param map the map
4767	>	* @param reducer a commutative associative combining function
4768	>	* @return the task
4769	>	*/
4770	>	public static <K,V> ForkJoinTask<K> reduceKeys
4771	>	(ConcurrentHashMapV8<K,V> map,
4772	>	BiFun<? super K, ? super K, ? extends K> reducer) {
4773	>	if (reducer == null) throw new NullPointerException();
4774	>	return new ReduceKeysTask<K,V>
4775	>	(map, null, -1, null, reducer);
4776	>	}
4777	>
4778	>	/**
4779	>	* Returns a task that when invoked, returns the result of
4780	>	* accumulating the given transformation of all keys using the given
4781	>	* reducer to combine values, or null if none.
4782	>	*
4783	>	* @param map the map
4784	>	* @param transformer a function returning the transformation
4785	>	* for an element, or null if there is no transformation (in
4786	>	* which case it is not combined).
4787	>	* @param reducer a commutative associative combining function
4788	>	* @return the task
4789	>	*/
4790	>	public static <K,V,U> ForkJoinTask<U> reduceKeys
4791	>	(ConcurrentHashMapV8<K,V> map,
4792	>	Fun<? super K, ? extends U> transformer,
4793	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4794	>	if (transformer == null || reducer == null)
4795	>	throw new NullPointerException();
4796	>	return new MapReduceKeysTask<K,V,U>
4797	>	(map, null, -1, null, transformer, reducer);
4798	>	}
4799	>
4800	>	/**
4801	>	* Returns a task that when invoked, returns the result of
4802	>	* accumulating the given transformation of all keys using the given
4803	>	* reducer to combine values, and the given basis as an
4804	>	* identity value.
4805	>	*
4806	>	* @param map the map
4807	>	* @param transformer a function returning the transformation
4808	>	* for an element
4809	>	* @param basis the identity (initial default value) for the reduction
4810	>	* @param reducer a commutative associative combining function
4811	>	* @return the task
4812	>	*/
4813	>	public static <K,V> ForkJoinTask<Double> reduceKeysToDouble
4814	>	(ConcurrentHashMapV8<K,V> map,
4815	>	ObjectToDouble<? super K> transformer,
4816	>	double basis,
4817	>	DoubleByDoubleToDouble reducer) {
4818	>	if (transformer == null || reducer == null)
4819	>	throw new NullPointerException();
4820	>	return new MapReduceKeysToDoubleTask<K,V>
4821	>	(map, null, -1, null, transformer, basis, reducer);
4822	>	}
4823	>
4824	>	/**
4825	>	* Returns a task that when invoked, returns the result of
4826	>	* accumulating the given transformation of all keys using the given
4827	>	* reducer to combine values, and the given basis as an
4828	>	* identity value.
4829	>	*
4830	>	* @param map the map
4831	>	* @param transformer a function returning the transformation
4832	>	* for an element
4833	>	* @param basis the identity (initial default value) for the reduction
4834	>	* @param reducer a commutative associative combining function
4835	>	* @return the task
4836	>	*/
4837	>	public static <K,V> ForkJoinTask<Long> reduceKeysToLong
4838	>	(ConcurrentHashMapV8<K,V> map,
4839	>	ObjectToLong<? super K> transformer,
4840	>	long basis,
4841	>	LongByLongToLong reducer) {
4842	>	if (transformer == null || reducer == null)
4843	>	throw new NullPointerException();
4844	>	return new MapReduceKeysToLongTask<K,V>
4845	>	(map, null, -1, null, transformer, basis, reducer);
4846	>	}
4847	>
4848	>	/**
4849	>	* Returns a task that when invoked, returns the result of
4850	>	* accumulating the given transformation of all keys using the given
4851	>	* reducer to combine values, and the given basis as an
4852	>	* identity value.
4853	>	*
4854	>	* @param map the map
4855	>	* @param transformer a function returning the transformation
4856	>	* for an element
4857	>	* @param basis the identity (initial default value) for the reduction
4858	>	* @param reducer a commutative associative combining function
4859	>	* @return the task
4860	>	*/
4861	>	public static <K,V> ForkJoinTask<Integer> reduceKeysToInt
4862	>	(ConcurrentHashMapV8<K,V> map,
4863	>	ObjectToInt<? super K> transformer,
4864	>	int basis,
4865	>	IntByIntToInt reducer) {
4866	>	if (transformer == null || reducer == null)
4867	>	throw new NullPointerException();
4868	>	return new MapReduceKeysToIntTask<K,V>
4869	>	(map, null, -1, null, transformer, basis, reducer);
4870	>	}
4871	>
4872	>	/**
4873	>	* Returns a task that when invoked, performs the given action
4874	>	* for each value.
4875	>	*
4876	>	* @param map the map
4877	>	* @param action the action
4878	>	*/
4879	>	public static <K,V> ForkJoinTask<Void> forEachValue
4880	>	(ConcurrentHashMapV8<K,V> map,
4881	>	Action<V> action) {
4882	>	if (action == null) throw new NullPointerException();
4883	>	return new ForEachValueTask<K,V>(map, null, -1, action);
4884	>	}
4885	>
4886	>	/**
4887	>	* Returns a task that when invoked, performs the given action
4888	>	* for each non-null transformation of each value.
4889	>	*
4890	>	* @param map the map
4891	>	* @param transformer a function returning the transformation
4892	>	* for an element, or null if there is no transformation (in
4893	>	* which case the action is not applied)
4894	>	* @param action the action
4895	>	*/
4896	>	public static <K,V,U> ForkJoinTask<Void> forEachValue
4897	>	(ConcurrentHashMapV8<K,V> map,
4898	>	Fun<? super V, ? extends U> transformer,
4899	>	Action<U> action) {
4900	>	if (transformer == null || action == null)
4901	>	throw new NullPointerException();
4902	>	return new ForEachTransformedValueTask<K,V,U>
4903	>	(map, null, -1, transformer, action);
4904	>	}
4905	>
4906	>	/**
4907	>	* Returns a task that when invoked, returns a non-null result
4908	>	* from applying the given search function on each value, or
4909	>	* null if none.  Upon success, further element processing is
4910	>	* suppressed and the results of any other parallel
4911	>	* invocations of the search function are ignored.
4912	>	*
4913	>	* @param map the map
4914	>	* @param searchFunction a function returning a non-null
4915	>	* result on success, else null
4916	>	* @return the task
4917	>	*/
4918	>	public static <K,V,U> ForkJoinTask<U> searchValues
4919	>	(ConcurrentHashMapV8<K,V> map,
4920	>	Fun<? super V, ? extends U> searchFunction) {
4921	>	if (searchFunction == null) throw new NullPointerException();
4922	>	return new SearchValuesTask<K,V,U>
4923	>	(map, null, -1, searchFunction,
4924	>	new AtomicReference<U>());
4925	>	}
4926	>
4927	>	/**
4928	>	* Returns a task that when invoked, returns the result of
4929	>	* accumulating all values using the given reducer to combine
4930	>	* values, or null if none.
4931	>	*
4932	>	* @param map the map
4933	>	* @param reducer a commutative associative combining function
4934	>	* @return the task
4935	>	*/
4936	>	public static <K,V> ForkJoinTask<V> reduceValues
4937	>	(ConcurrentHashMapV8<K,V> map,
4938	>	BiFun<? super V, ? super V, ? extends V> reducer) {
4939	>	if (reducer == null) throw new NullPointerException();
4940	>	return new ReduceValuesTask<K,V>
4941	>	(map, null, -1, null, reducer);
4942	>	}
4943	>
4944	>	/**
4945	>	* Returns a task that when invoked, returns the result of
4946	>	* accumulating the given transformation of all values using the
4947	>	* given reducer to combine values, or null if none.
4948	>	*
4949	>	* @param map the map
4950	>	* @param transformer a function returning the transformation
4951	>	* for an element, or null if there is no transformation (in
4952	>	* which case it is not combined).
4953	>	* @param reducer a commutative associative combining function
4954	>	* @return the task
4955	>	*/
4956	>	public static <K,V,U> ForkJoinTask<U> reduceValues
4957	>	(ConcurrentHashMapV8<K,V> map,
4958	>	Fun<? super V, ? extends U> transformer,
4959	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4960	>	if (transformer == null || reducer == null)
4961	>	throw new NullPointerException();
4962	>	return new MapReduceValuesTask<K,V,U>
4963	>	(map, null, -1, null, transformer, reducer);
4964	>	}
4965	>
4966	>	/**
4967	>	* Returns a task that when invoked, returns the result of
4968	>	* accumulating the given transformation of all values using the
4969	>	* given reducer to combine values, and the given basis as an
4970	>	* identity value.
4971	>	*
4972	>	* @param map the map
4973	>	* @param transformer a function returning the transformation
4974	>	* for an element
4975	>	* @param basis the identity (initial default value) for the reduction
4976	>	* @param reducer a commutative associative combining function
4977	>	* @return the task
4978	>	*/
4979	>	public static <K,V> ForkJoinTask<Double> reduceValuesToDouble
4980	>	(ConcurrentHashMapV8<K,V> map,
4981	>	ObjectToDouble<? super V> transformer,
4982	>	double basis,
4983	>	DoubleByDoubleToDouble reducer) {
4984	>	if (transformer == null || reducer == null)
4985	>	throw new NullPointerException();
4986	>	return new MapReduceValuesToDoubleTask<K,V>
4987	>	(map, null, -1, null, transformer, basis, reducer);
4988	>	}
4989	>
4990	>	/**
4991	>	* Returns a task that when invoked, returns the result of
4992	>	* accumulating the given transformation of all values using the
4993	>	* given reducer to combine values, and the given basis as an
4994	>	* identity value.
4995	>	*
4996	>	* @param map the map
4997	>	* @param transformer a function returning the transformation
4998	>	* for an element
4999	>	* @param basis the identity (initial default value) for the reduction
5000	>	* @param reducer a commutative associative combining function
5001	>	* @return the task
5002	>	*/
5003	>	public static <K,V> ForkJoinTask<Long> reduceValuesToLong
5004	>	(ConcurrentHashMapV8<K,V> map,
5005	>	ObjectToLong<? super V> transformer,
5006	>	long basis,
5007	>	LongByLongToLong reducer) {
5008	>	if (transformer == null || reducer == null)
5009	>	throw new NullPointerException();
5010	>	return new MapReduceValuesToLongTask<K,V>
5011	>	(map, null, -1, null, transformer, basis, reducer);
5012	>	}
5013	>
5014	>	/**
5015	>	* Returns a task that when invoked, returns the result of
5016	>	* accumulating the given transformation of all values using the
5017	>	* given reducer to combine values, and the given basis as an
5018	>	* identity value.
5019	>	*
5020	>	* @param map the map
5021	>	* @param transformer a function returning the transformation
5022	>	* for an element
5023	>	* @param basis the identity (initial default value) for the reduction
5024	>	* @param reducer a commutative associative combining function
5025	>	* @return the task
5026	>	*/
5027	>	public static <K,V> ForkJoinTask<Integer> reduceValuesToInt
5028	>	(ConcurrentHashMapV8<K,V> map,
5029	>	ObjectToInt<? super V> transformer,
5030	>	int basis,
5031	>	IntByIntToInt reducer) {
5032	>	if (transformer == null || reducer == null)
5033	>	throw new NullPointerException();
5034	>	return new MapReduceValuesToIntTask<K,V>
5035	>	(map, null, -1, null, transformer, basis, reducer);
5036	>	}
5037	>
5038	>	/**
5039	>	* Returns a task that when invoked, perform the given action
5040	>	* for each entry.
5041	>	*
5042	>	* @param map the map
5043	>	* @param action the action
5044	>	*/
5045	>	public static <K,V> ForkJoinTask<Void> forEachEntry
5046	>	(ConcurrentHashMapV8<K,V> map,
5047	>	Action<Map.Entry<K,V>> action) {
5048	>	if (action == null) throw new NullPointerException();
5049	>	return new ForEachEntryTask<K,V>(map, null, -1, action);
5050	>	}
5051	>
5052	>	/**
5053	>	* Returns a task that when invoked, perform the given action
5054	>	* for each non-null transformation of each entry.
5055	>	*
5056	>	* @param map the map
5057	>	* @param transformer a function returning the transformation
5058	>	* for an element, or null if there is no transformation (in
5059	>	* which case the action is not applied)
5060	>	* @param action the action
5061	>	*/
5062	>	public static <K,V,U> ForkJoinTask<Void> forEachEntry
5063	>	(ConcurrentHashMapV8<K,V> map,
5064	>	Fun<Map.Entry<K,V>, ? extends U> transformer,
5065	>	Action<U> action) {
5066	>	if (transformer == null || action == null)
5067	>	throw new NullPointerException();
5068	>	return new ForEachTransformedEntryTask<K,V,U>
5069	>	(map, null, -1, transformer, action);
5070	>	}
5071	>
5072	>	/**
5073	>	* Returns a task that when invoked, returns a non-null result
5074	>	* from applying the given search function on each entry, or
5075	>	* null if none.  Upon success, further element processing is
5076	>	* suppressed and the results of any other parallel
5077	>	* invocations of the search function are ignored.
5078	>	*
5079	>	* @param map the map
5080	>	* @param searchFunction a function returning a non-null
5081	>	* result on success, else null
5082	>	* @return the task
5083	>	*/
5084	>	public static <K,V,U> ForkJoinTask<U> searchEntries
5085	>	(ConcurrentHashMapV8<K,V> map,
5086	>	Fun<Map.Entry<K,V>, ? extends U> searchFunction) {
5087	>	if (searchFunction == null) throw new NullPointerException();
5088	>	return new SearchEntriesTask<K,V,U>
5089	>	(map, null, -1, searchFunction,
5090	>	new AtomicReference<U>());
5091	>	}
5092	>
5093	>	/**
5094	>	* Returns a task that when invoked, returns the result of
5095	>	* accumulating all entries using the given reducer to combine
5096	>	* values, or null if none.
5097	>	*
5098	>	* @param map the map
5099	>	* @param reducer a commutative associative combining function
5100	>	* @return the task
5101	>	*/
5102	>	public static <K,V> ForkJoinTask<Map.Entry<K,V>> reduceEntries
5103	>	(ConcurrentHashMapV8<K,V> map,
5104	>	BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer) {
5105	>	if (reducer == null) throw new NullPointerException();
5106	>	return new ReduceEntriesTask<K,V>
5107	>	(map, null, -1, null, reducer);
5108	>	}
5109	>
5110	>	/**
5111	>	* Returns a task that when invoked, returns the result of
5112	>	* accumulating the given transformation of all entries using the
5113	>	* given reducer to combine values, or null if none.
5114	>	*
5115	>	* @param map the map
5116	>	* @param transformer a function returning the transformation
5117	>	* for an element, or null if there is no transformation (in
5118	>	* which case it is not combined).
5119	>	* @param reducer a commutative associative combining function
5120	>	* @return the task
5121	>	*/
5122	>	public static <K,V,U> ForkJoinTask<U> reduceEntries
5123	>	(ConcurrentHashMapV8<K,V> map,
5124	>	Fun<Map.Entry<K,V>, ? extends U> transformer,
5125	>	BiFun<? super U, ? super U, ? extends U> reducer) {
5126	>	if (transformer == null || reducer == null)
5127	>	throw new NullPointerException();
5128	>	return new MapReduceEntriesTask<K,V,U>
5129	>	(map, null, -1, null, transformer, reducer);
5130	>	}
5131	>
5132	>	/**
5133	>	* Returns a task that when invoked, returns the result of
5134	>	* accumulating the given transformation of all entries using the
5135	>	* given reducer to combine values, and the given basis as an
5136	>	* identity value.
5137	>	*
5138	>	* @param map the map
5139	>	* @param transformer a function returning the transformation
5140	>	* for an element
5141	>	* @param basis the identity (initial default value) for the reduction
5142	>	* @param reducer a commutative associative combining function
5143	>	* @return the task
5144	>	*/
5145	>	public static <K,V> ForkJoinTask<Double> reduceEntriesToDouble
5146	>	(ConcurrentHashMapV8<K,V> map,
5147	>	ObjectToDouble<Map.Entry<K,V>> transformer,
5148	>	double basis,
5149	>	DoubleByDoubleToDouble reducer) {
5150	>	if (transformer == null || reducer == null)
5151	>	throw new NullPointerException();
5152	>	return new MapReduceEntriesToDoubleTask<K,V>
5153	>	(map, null, -1, null, transformer, basis, reducer);
5154	>	}
5155	>
5156	>	/**
5157	>	* Returns a task that when invoked, returns the result of
5158	>	* accumulating the given transformation of all entries using the
5159	>	* given reducer to combine values, and the given basis as an
5160	>	* identity value.
5161	>	*
5162	>	* @param map the map
5163	>	* @param transformer a function returning the transformation
5164	>	* for an element
5165	>	* @param basis the identity (initial default value) for the reduction
5166	>	* @param reducer a commutative associative combining function
5167	>	* @return the task
5168	>	*/
5169	>	public static <K,V> ForkJoinTask<Long> reduceEntriesToLong
5170	>	(ConcurrentHashMapV8<K,V> map,
5171	>	ObjectToLong<Map.Entry<K,V>> transformer,
5172	>	long basis,
5173	>	LongByLongToLong reducer) {
5174	>	if (transformer == null || reducer == null)
5175	>	throw new NullPointerException();
5176	>	return new MapReduceEntriesToLongTask<K,V>
5177	>	(map, null, -1, null, transformer, basis, reducer);
5178	>	}
5179	>
5180	>	/**
5181	>	* Returns a task that when invoked, returns the result of
5182	>	* accumulating the given transformation of all entries using the
5183	>	* given reducer to combine values, and the given basis as an
5184	>	* identity value.
5185	>	*
5186	>	* @param map the map
5187	>	* @param transformer a function returning the transformation
5188	>	* for an element
5189	>	* @param basis the identity (initial default value) for the reduction
5190	>	* @param reducer a commutative associative combining function
5191	>	* @return the task
5192	>	*/
5193	>	public static <K,V> ForkJoinTask<Integer> reduceEntriesToInt
5194	>	(ConcurrentHashMapV8<K,V> map,
5195	>	ObjectToInt<Map.Entry<K,V>> transformer,
5196	>	int basis,
5197	>	IntByIntToInt reducer) {
5198	>	if (transformer == null || reducer == null)
5199	>	throw new NullPointerException();
5200	>	return new MapReduceEntriesToIntTask<K,V>
5201	>	(map, null, -1, null, transformer, basis, reducer);
5202	>	}
5203	>	}
5204	>
5205	>	// -------------------------------------------------------
5206	>
5207	>	/*
5208	>	* Task classes. Coded in a regular but ugly format/style to
5209	>	* simplify checks that each variant differs in the right way from
5210	>	* others. The null screenings exist because compilers cannot tell
5211	>	* that we've already null-checked task arguments, so we force
5212	>	* simplest hoisted bypass to help avoid convoluted traps.
5213	>	*/
5214	>
5215	>	@SuppressWarnings("serial") static final class ForEachKeyTask<K,V>
5216	>	extends Traverser<K,V,Void> {
5217	>	final Action<K> action;
5218	>	ForEachKeyTask
5219	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5220	>	Action<K> action) {
5221	>	super(m, p, b);
5222	>	this.action = action;
5223	>	}
5224	>	@SuppressWarnings("unchecked") public final void compute() {
5225	>	final Action<K> action;
5226	>	if ((action = this.action) != null) {
5227	>	for (int b; (b = preSplit()) > 0;)
5228	>	new ForEachKeyTask<K,V>(map, this, b, action).fork();
5229	>	while (advance() != null)
5230	>	action.apply((K)nextKey);
5231	>	propagateCompletion();
5232	>	}
5233	>	}
5234	>	}
5235	>
5236	>	@SuppressWarnings("serial") static final class ForEachValueTask<K,V>
5237	>	extends Traverser<K,V,Void> {
5238	>	final Action<V> action;
5239	>	ForEachValueTask
5240	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5241	>	Action<V> action) {
5242	>	super(m, p, b);
5243	>	this.action = action;
5244	>	}
5245	>	@SuppressWarnings("unchecked") public final void compute() {
5246	>	final Action<V> action;
5247	>	if ((action = this.action) != null) {
5248	>	for (int b; (b = preSplit()) > 0;)
5249	>	new ForEachValueTask<K,V>(map, this, b, action).fork();
5250	>	Object v;
5251	>	while ((v = advance()) != null)
5252	>	action.apply((V)v);
5253	>	propagateCompletion();
5254	>	}
5255	>	}
5256	>	}
5257	>
5258	>	@SuppressWarnings("serial") static final class ForEachEntryTask<K,V>
5259	>	extends Traverser<K,V,Void> {
5260	>	final Action<Entry<K,V>> action;
5261	>	ForEachEntryTask
5262	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5263	>	Action<Entry<K,V>> action) {
5264	>	super(m, p, b);
5265	>	this.action = action;
5266	>	}
5267	>	@SuppressWarnings("unchecked") public final void compute() {
5268	>	final Action<Entry<K,V>> action;
5269	>	if ((action = this.action) != null) {
5270	>	for (int b; (b = preSplit()) > 0;)
5271	>	new ForEachEntryTask<K,V>(map, this, b, action).fork();
5272	>	Object v;
5273	>	while ((v = advance()) != null)
5274	>	action.apply(entryFor((K)nextKey, (V)v));
5275	>	propagateCompletion();
5276	>	}
5277	>	}
5278	>	}
5279	>
5280	>	@SuppressWarnings("serial") static final class ForEachMappingTask<K,V>
5281	>	extends Traverser<K,V,Void> {
5282	>	final BiAction<K,V> action;
5283	>	ForEachMappingTask
5284	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5285	>	BiAction<K,V> action) {
5286	>	super(m, p, b);
5287	>	this.action = action;
5288	>	}
5289	>	@SuppressWarnings("unchecked") public final void compute() {
5290	>	final BiAction<K,V> action;
5291	>	if ((action = this.action) != null) {
5292	>	for (int b; (b = preSplit()) > 0;)
5293	>	new ForEachMappingTask<K,V>(map, this, b, action).fork();
5294	>	Object v;
5295	>	while ((v = advance()) != null)
5296	>	action.apply((K)nextKey, (V)v);
5297	>	propagateCompletion();
5298	>	}
5299	>	}
5300	>	}
5301	>
5302	>	@SuppressWarnings("serial") static final class ForEachTransformedKeyTask<K,V,U>
5303	>	extends Traverser<K,V,Void> {
5304	>	final Fun<? super K, ? extends U> transformer;
5305	>	final Action<U> action;
5306	>	ForEachTransformedKeyTask
5307	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5308	>	Fun<? super K, ? extends U> transformer, Action<U> action) {
5309	>	super(m, p, b);
5310	>	this.transformer = transformer; this.action = action;
5311	>	}
5312	>	@SuppressWarnings("unchecked") public final void compute() {
5313	>	final Fun<? super K, ? extends U> transformer;
5314	>	final Action<U> action;
5315	>	if ((transformer = this.transformer) != null &&
5316	>	(action = this.action) != null) {
5317	>	for (int b; (b = preSplit()) > 0;)
5318	>	new ForEachTransformedKeyTask<K,V,U>
5319	>	(map, this, b, transformer, action).fork();
5320	>	U u;
5321	>	while (advance() != null) {
5322	>	if ((u = transformer.apply((K)nextKey)) != null)
5323	>	action.apply(u);
5324	>	}
5325	>	propagateCompletion();
5326	>	}
5327	>	}
5328	>	}
5329	>
5330	>	@SuppressWarnings("serial") static final class ForEachTransformedValueTask<K,V,U>
5331	>	extends Traverser<K,V,Void> {
5332	>	final Fun<? super V, ? extends U> transformer;
5333	>	final Action<U> action;
5334	>	ForEachTransformedValueTask
5335	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5336	>	Fun<? super V, ? extends U> transformer, Action<U> action) {
5337	>	super(m, p, b);
5338	>	this.transformer = transformer; this.action = action;
5339	>	}
5340	>	@SuppressWarnings("unchecked") public final void compute() {
5341	>	final Fun<? super V, ? extends U> transformer;
5342	>	final Action<U> action;
5343	>	if ((transformer = this.transformer) != null &&
5344	>	(action = this.action) != null) {
5345	>	for (int b; (b = preSplit()) > 0;)
5346	>	new ForEachTransformedValueTask<K,V,U>
5347	>	(map, this, b, transformer, action).fork();
5348	>	Object v; U u;
5349	>	while ((v = advance()) != null) {
5350	>	if ((u = transformer.apply((V)v)) != null)
5351	>	action.apply(u);
5352	>	}
5353	>	propagateCompletion();
5354	>	}
5355	>	}
5356	>	}
5357	>
5358	>	@SuppressWarnings("serial") static final class ForEachTransformedEntryTask<K,V,U>
5359	>	extends Traverser<K,V,Void> {
5360	>	final Fun<Map.Entry<K,V>, ? extends U> transformer;
5361	>	final Action<U> action;
5362	>	ForEachTransformedEntryTask
5363	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5364	>	Fun<Map.Entry<K,V>, ? extends U> transformer, Action<U> action) {
5365	>	super(m, p, b);
5366	>	this.transformer = transformer; this.action = action;
5367	>	}
5368	>	@SuppressWarnings("unchecked") public final void compute() {
5369	>	final Fun<Map.Entry<K,V>, ? extends U> transformer;
5370	>	final Action<U> action;
5371	>	if ((transformer = this.transformer) != null &&
5372	>	(action = this.action) != null) {
5373	>	for (int b; (b = preSplit()) > 0;)
5374	>	new ForEachTransformedEntryTask<K,V,U>
5375	>	(map, this, b, transformer, action).fork();
5376	>	Object v; U u;
5377	>	while ((v = advance()) != null) {
5378	>	if ((u = transformer.apply(entryFor((K)nextKey,
5379	>	(V)v))) != null)
5380	>	action.apply(u);
5381	>	}
5382	>	propagateCompletion();
5383	>	}
5384	>	}
5385	>	}
5386	>
5387	>	@SuppressWarnings("serial") static final class ForEachTransformedMappingTask<K,V,U>
5388	>	extends Traverser<K,V,Void> {
5389	>	final BiFun<? super K, ? super V, ? extends U> transformer;
5390	>	final Action<U> action;
5391	>	ForEachTransformedMappingTask
5392	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5393	>	BiFun<? super K, ? super V, ? extends U> transformer,
5394	>	Action<U> action) {
5395	>	super(m, p, b);
5396	>	this.transformer = transformer; this.action = action;
5397	>	}
5398	>	@SuppressWarnings("unchecked") public final void compute() {
5399	>	final BiFun<? super K, ? super V, ? extends U> transformer;
5400	>	final Action<U> action;
5401	>	if ((transformer = this.transformer) != null &&
5402	>	(action = this.action) != null) {
5403	>	for (int b; (b = preSplit()) > 0;)
5404	>	new ForEachTransformedMappingTask<K,V,U>
5405	>	(map, this, b, transformer, action).fork();
5406	>	Object v; U u;
5407	>	while ((v = advance()) != null) {
5408	>	if ((u = transformer.apply((K)nextKey, (V)v)) != null)
5409	>	action.apply(u);
5410	>	}
5411	>	propagateCompletion();
5412	>	}
5413	>	}
5414	>	}
5415	>
5416	>	@SuppressWarnings("serial") static final class SearchKeysTask<K,V,U>
5417	>	extends Traverser<K,V,U> {
5418	>	final Fun<? super K, ? extends U> searchFunction;
5419	>	final AtomicReference<U> result;
5420	>	SearchKeysTask
5421	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5422	>	Fun<? super K, ? extends U> searchFunction,
5423	>	AtomicReference<U> result) {
5424	>	super(m, p, b);
5425	>	this.searchFunction = searchFunction; this.result = result;
5426	>	}
5427	>	public final U getRawResult() { return result.get(); }
5428	>	@SuppressWarnings("unchecked") public final void compute() {
5429	>	final Fun<? super K, ? extends U> searchFunction;
5430	>	final AtomicReference<U> result;
5431	>	if ((searchFunction = this.searchFunction) != null &&
5432	>	(result = this.result) != null) {
5433	>	for (int b;;) {
5434	>	if (result.get() != null)
5435	>	return;
5436	>	if ((b = preSplit()) <= 0)
5437	>	break;
5438	>	new SearchKeysTask<K,V,U>
5439	>	(map, this, b, searchFunction, result).fork();
5440	>	}
5441	>	while (result.get() == null) {
5442	>	U u;
5443	>	if (advance() == null) {
5444	>	propagateCompletion();
5445	>	break;
5446	>	}
5447	>	if ((u = searchFunction.apply((K)nextKey)) != null) {
5448	>	if (result.compareAndSet(null, u))
5449	>	quietlyCompleteRoot();
5450	>	break;
5451	>	}
5452	>	}
5453	>	}
5454	>	}
5455	>	}
5456	>
5457	>	@SuppressWarnings("serial") static final class SearchValuesTask<K,V,U>
5458	>	extends Traverser<K,V,U> {
5459	>	final Fun<? super V, ? extends U> searchFunction;
5460	>	final AtomicReference<U> result;
5461	>	SearchValuesTask
5462	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5463	>	Fun<? super V, ? extends U> searchFunction,
5464	>	AtomicReference<U> result) {
5465	>	super(m, p, b);
5466	>	this.searchFunction = searchFunction; this.result = result;
5467	>	}
5468	>	public final U getRawResult() { return result.get(); }
5469	>	@SuppressWarnings("unchecked") public final void compute() {
5470	>	final Fun<? super V, ? extends U> searchFunction;
5471	>	final AtomicReference<U> result;
5472	>	if ((searchFunction = this.searchFunction) != null &&
5473	>	(result = this.result) != null) {
5474	>	for (int b;;) {
5475	>	if (result.get() != null)
5476	>	return;
5477	>	if ((b = preSplit()) <= 0)
5478	>	break;
5479	>	new SearchValuesTask<K,V,U>
5480	>	(map, this, b, searchFunction, result).fork();
5481	>	}
5482	>	while (result.get() == null) {
5483	>	Object v; U u;
5484	>	if ((v = advance()) == null) {
5485	>	propagateCompletion();
5486	>	break;
5487	>	}
5488	>	if ((u = searchFunction.apply((V)v)) != null) {
5489	>	if (result.compareAndSet(null, u))
5490	>	quietlyCompleteRoot();
5491	>	break;
5492	>	}
5493	>	}
5494	>	}
5495	>	}
5496	>	}
5497	>
5498	>	@SuppressWarnings("serial") static final class SearchEntriesTask<K,V,U>
5499	>	extends Traverser<K,V,U> {
5500	>	final Fun<Entry<K,V>, ? extends U> searchFunction;
5501	>	final AtomicReference<U> result;
5502	>	SearchEntriesTask
5503	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5504	>	Fun<Entry<K,V>, ? extends U> searchFunction,
5505	>	AtomicReference<U> result) {
5506	>	super(m, p, b);
5507	>	this.searchFunction = searchFunction; this.result = result;
5508	>	}
5509	>	public final U getRawResult() { return result.get(); }
5510	>	@SuppressWarnings("unchecked") public final void compute() {
5511	>	final Fun<Entry<K,V>, ? extends U> searchFunction;
5512	>	final AtomicReference<U> result;
5513	>	if ((searchFunction = this.searchFunction) != null &&
5514	>	(result = this.result) != null) {
5515	>	for (int b;;) {
5516	>	if (result.get() != null)
5517	>	return;
5518	>	if ((b = preSplit()) <= 0)
5519	>	break;
5520	>	new SearchEntriesTask<K,V,U>
5521	>	(map, this, b, searchFunction, result).fork();
5522	>	}
5523	>	while (result.get() == null) {
5524	>	Object v; U u;
5525	>	if ((v = advance()) == null) {
5526	>	propagateCompletion();
5527	>	break;
5528	>	}
5529	>	if ((u = searchFunction.apply(entryFor((K)nextKey,
5530	>	(V)v))) != null) {
5531	>	if (result.compareAndSet(null, u))
5532	>	quietlyCompleteRoot();
5533	>	return;
5534	>	}
5535	>	}
5536	>	}
5537	>	}
5538	>	}
5539	>
5540	>	@SuppressWarnings("serial") static final class SearchMappingsTask<K,V,U>
5541	>	extends Traverser<K,V,U> {
5542	>	final BiFun<? super K, ? super V, ? extends U> searchFunction;
5543	>	final AtomicReference<U> result;
5544	>	SearchMappingsTask
5545	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5546	>	BiFun<? super K, ? super V, ? extends U> searchFunction,
5547	>	AtomicReference<U> result) {
5548	>	super(m, p, b);
5549	>	this.searchFunction = searchFunction; this.result = result;
5550	>	}
5551	>	public final U getRawResult() { return result.get(); }
5552	>	@SuppressWarnings("unchecked") public final void compute() {
5553	>	final BiFun<? super K, ? super V, ? extends U> searchFunction;
5554	>	final AtomicReference<U> result;
5555	>	if ((searchFunction = this.searchFunction) != null &&
5556	>	(result = this.result) != null) {
5557	>	for (int b;;) {
5558	>	if (result.get() != null)
5559	>	return;
5560	>	if ((b = preSplit()) <= 0)
5561	>	break;
5562	>	new SearchMappingsTask<K,V,U>
5563	>	(map, this, b, searchFunction, result).fork();
5564	>	}
5565	>	while (result.get() == null) {
5566	>	Object v; U u;
5567	>	if ((v = advance()) == null) {
5568	>	propagateCompletion();
5569	>	break;
5570	>	}
5571	>	if ((u = searchFunction.apply((K)nextKey, (V)v)) != null) {
5572	>	if (result.compareAndSet(null, u))
5573	>	quietlyCompleteRoot();
5574	>	break;
5575	>	}
5576	>	}
5577	>	}
5578	>	}
5579	>	}
5580	>
5581	>	@SuppressWarnings("serial") static final class ReduceKeysTask<K,V>
5582	>	extends Traverser<K,V,K> {
5583	>	final BiFun<? super K, ? super K, ? extends K> reducer;
5584	>	K result;
5585	>	ReduceKeysTask<K,V> rights, nextRight;
5586	>	ReduceKeysTask
5587	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5588	>	ReduceKeysTask<K,V> nextRight,
5589	>	BiFun<? super K, ? super K, ? extends K> reducer) {
5590	>	super(m, p, b); this.nextRight = nextRight;
5591	>	this.reducer = reducer;
5592	>	}
5593	>	public final K getRawResult() { return result; }
5594	>	@SuppressWarnings("unchecked") public final void compute() {
5595	>	final BiFun<? super K, ? super K, ? extends K> reducer;
5596	>	if ((reducer = this.reducer) != null) {
5597	>	for (int b; (b = preSplit()) > 0;)
5598	>	(rights = new ReduceKeysTask<K,V>
5599	>	(map, this, b, rights, reducer)).fork();
5600	>	K r = null;
5601	>	while (advance() != null) {
5602	>	K u = (K)nextKey;
5603	>	r = (r == null) ? u : reducer.apply(r, u);
5604	>	}
5605	>	result = r;
5606	>	CountedCompleter<?> c;
5607	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5608	>	ReduceKeysTask<K,V>
5609	>	t = (ReduceKeysTask<K,V>)c,
5610	>	s = t.rights;
5611	>	while (s != null) {
5612	>	K tr, sr;
5613	>	if ((sr = s.result) != null)
5614	>	t.result = (((tr = t.result) == null) ? sr :
5615	>	reducer.apply(tr, sr));
5616	>	s = t.rights = s.nextRight;
5617	>	}
5618	>	}
5619	>	}
5620	>	}
5621	>	}
5622	>
5623	>	@SuppressWarnings("serial") static final class ReduceValuesTask<K,V>
5624	>	extends Traverser<K,V,V> {
5625	>	final BiFun<? super V, ? super V, ? extends V> reducer;
5626	>	V result;
5627	>	ReduceValuesTask<K,V> rights, nextRight;
5628	>	ReduceValuesTask
5629	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5630	>	ReduceValuesTask<K,V> nextRight,
5631	>	BiFun<? super V, ? super V, ? extends V> reducer) {
5632	>	super(m, p, b); this.nextRight = nextRight;
5633	>	this.reducer = reducer;
5634	>	}
5635	>	public final V getRawResult() { return result; }
5636	>	@SuppressWarnings("unchecked") public final void compute() {
5637	>	final BiFun<? super V, ? super V, ? extends V> reducer;
5638	>	if ((reducer = this.reducer) != null) {
5639	>	for (int b; (b = preSplit()) > 0;)
5640	>	(rights = new ReduceValuesTask<K,V>
5641	>	(map, this, b, rights, reducer)).fork();
5642	>	V r = null;
5643	>	Object v;
5644	>	while ((v = advance()) != null) {
5645	>	V u = (V)v;
5646	>	r = (r == null) ? u : reducer.apply(r, u);
5647	>	}
5648	>	result = r;
5649	>	CountedCompleter<?> c;
5650	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5651	>	ReduceValuesTask<K,V>
5652	>	t = (ReduceValuesTask<K,V>)c,
5653	>	s = t.rights;
5654	>	while (s != null) {
5655	>	V tr, sr;
5656	>	if ((sr = s.result) != null)
5657	>	t.result = (((tr = t.result) == null) ? sr :
5658	>	reducer.apply(tr, sr));
5659	>	s = t.rights = s.nextRight;
5660	>	}
5661	>	}
5662	>	}
5663	>	}
5664	>	}
5665	>
5666	>	@SuppressWarnings("serial") static final class ReduceEntriesTask<K,V>
5667	>	extends Traverser<K,V,Map.Entry<K,V>> {
5668	>	final BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer;
5669	>	Map.Entry<K,V> result;
5670	>	ReduceEntriesTask<K,V> rights, nextRight;
5671	>	ReduceEntriesTask
5672	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5673	>	ReduceEntriesTask<K,V> nextRight,
5674	>	BiFun<Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer) {
5675	>	super(m, p, b); this.nextRight = nextRight;
5676	>	this.reducer = reducer;
5677	>	}
5678	>	public final Map.Entry<K,V> getRawResult() { return result; }
5679	>	@SuppressWarnings("unchecked") public final void compute() {
5680	>	final BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer;
5681	>	if ((reducer = this.reducer) != null) {
5682	>	for (int b; (b = preSplit()) > 0;)
5683	>	(rights = new ReduceEntriesTask<K,V>
5684	>	(map, this, b, rights, reducer)).fork();
5685	>	Map.Entry<K,V> r = null;
5686	>	Object v;
5687	>	while ((v = advance()) != null) {
5688	>	Map.Entry<K,V> u = entryFor((K)nextKey, (V)v);
5689	>	r = (r == null) ? u : reducer.apply(r, u);
5690	>	}
5691	>	result = r;
5692	>	CountedCompleter<?> c;
5693	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5694	>	ReduceEntriesTask<K,V>
5695	>	t = (ReduceEntriesTask<K,V>)c,
5696	>	s = t.rights;
5697	>	while (s != null) {
5698	>	Map.Entry<K,V> tr, sr;
5699	>	if ((sr = s.result) != null)
5700	>	t.result = (((tr = t.result) == null) ? sr :
5701	>	reducer.apply(tr, sr));
5702	>	s = t.rights = s.nextRight;
5703	>	}
5704	>	}
5705	>	}
5706	>	}
5707	>	}
5708	>
5709	>	@SuppressWarnings("serial") static final class MapReduceKeysTask<K,V,U>
5710	>	extends Traverser<K,V,U> {
5711	>	final Fun<? super K, ? extends U> transformer;
5712	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5713	>	U result;
5714	>	MapReduceKeysTask<K,V,U> rights, nextRight;
5715	>	MapReduceKeysTask
5716	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5717	>	MapReduceKeysTask<K,V,U> nextRight,
5718	>	Fun<? super K, ? extends U> transformer,
5719	>	BiFun<? super U, ? super U, ? extends U> reducer) {
5720	>	super(m, p, b); this.nextRight = nextRight;
5721	>	this.transformer = transformer;
5722	>	this.reducer = reducer;
5723	>	}
5724	>	public final U getRawResult() { return result; }
5725	>	@SuppressWarnings("unchecked") public final void compute() {
5726	>	final Fun<? super K, ? extends U> transformer;
5727	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5728	>	if ((transformer = this.transformer) != null &&
5729	>	(reducer = this.reducer) != null) {
5730	>	for (int b; (b = preSplit()) > 0;)
5731	>	(rights = new MapReduceKeysTask<K,V,U>
5732	>	(map, this, b, rights, transformer, reducer)).fork();
5733	>	U r = null, u;
5734	>	while (advance() != null) {
5735	>	if ((u = transformer.apply((K)nextKey)) != null)
5736	>	r = (r == null) ? u : reducer.apply(r, u);
5737	>	}
5738	>	result = r;
5739	>	CountedCompleter<?> c;
5740	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5741	>	MapReduceKeysTask<K,V,U>
5742	>	t = (MapReduceKeysTask<K,V,U>)c,
5743	>	s = t.rights;
5744	>	while (s != null) {
5745	>	U tr, sr;
5746	>	if ((sr = s.result) != null)
5747	>	t.result = (((tr = t.result) == null) ? sr :
5748	>	reducer.apply(tr, sr));
5749	>	s = t.rights = s.nextRight;
5750	>	}
5751	>	}
5752	>	}
5753	>	}
5754	>	}
5755	>
5756	>	@SuppressWarnings("serial") static final class MapReduceValuesTask<K,V,U>
5757	>	extends Traverser<K,V,U> {
5758	>	final Fun<? super V, ? extends U> transformer;
5759	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5760	>	U result;
5761	>	MapReduceValuesTask<K,V,U> rights, nextRight;
5762	>	MapReduceValuesTask
5763	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5764	>	MapReduceValuesTask<K,V,U> nextRight,
5765	>	Fun<? super V, ? extends U> transformer,
5766	>	BiFun<? super U, ? super U, ? extends U> reducer) {
5767	>	super(m, p, b); this.nextRight = nextRight;
5768	>	this.transformer = transformer;
5769	>	this.reducer = reducer;
5770	>	}
5771	>	public final U getRawResult() { return result; }
5772	>	@SuppressWarnings("unchecked") public final void compute() {
5773	>	final Fun<? super V, ? extends U> transformer;
5774	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5775	>	if ((transformer = this.transformer) != null &&
5776	>	(reducer = this.reducer) != null) {
5777	>	for (int b; (b = preSplit()) > 0;)
5778	>	(rights = new MapReduceValuesTask<K,V,U>
5779	>	(map, this, b, rights, transformer, reducer)).fork();
5780	>	U r = null, u;
5781	>	Object v;
5782	>	while ((v = advance()) != null) {
5783	>	if ((u = transformer.apply((V)v)) != null)
5784	>	r = (r == null) ? u : reducer.apply(r, u);
5785	>	}
5786	>	result = r;
5787	>	CountedCompleter<?> c;
5788	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5789	>	MapReduceValuesTask<K,V,U>
5790	>	t = (MapReduceValuesTask<K,V,U>)c,
5791	>	s = t.rights;
5792	>	while (s != null) {
5793	>	U tr, sr;
5794	>	if ((sr = s.result) != null)
5795	>	t.result = (((tr = t.result) == null) ? sr :
5796	>	reducer.apply(tr, sr));
5797	>	s = t.rights = s.nextRight;
5798	>	}
5799	>	}
5800	>	}
5801	>	}
5802	>	}
5803	>
5804	>	@SuppressWarnings("serial") static final class MapReduceEntriesTask<K,V,U>
5805	>	extends Traverser<K,V,U> {
5806	>	final Fun<Map.Entry<K,V>, ? extends U> transformer;
5807	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5808	>	U result;
5809	>	MapReduceEntriesTask<K,V,U> rights, nextRight;
5810	>	MapReduceEntriesTask
5811	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5812	>	MapReduceEntriesTask<K,V,U> nextRight,
5813	>	Fun<Map.Entry<K,V>, ? extends U> transformer,
5814	>	BiFun<? super U, ? super U, ? extends U> reducer) {
5815	>	super(m, p, b); this.nextRight = nextRight;
5816	>	this.transformer = transformer;
5817	>	this.reducer = reducer;
5818	>	}
5819	>	public final U getRawResult() { return result; }
5820	>	@SuppressWarnings("unchecked") public final void compute() {
5821	>	final Fun<Map.Entry<K,V>, ? extends U> transformer;
5822	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5823	>	if ((transformer = this.transformer) != null &&
5824	>	(reducer = this.reducer) != null) {
5825	>	for (int b; (b = preSplit()) > 0;)
5826	>	(rights = new MapReduceEntriesTask<K,V,U>
5827	>	(map, this, b, rights, transformer, reducer)).fork();
5828	>	U r = null, u;
5829	>	Object v;
5830	>	while ((v = advance()) != null) {
5831	>	if ((u = transformer.apply(entryFor((K)nextKey,
5832	>	(V)v))) != null)
5833	>	r = (r == null) ? u : reducer.apply(r, u);
5834	>	}
5835	>	result = r;
5836	>	CountedCompleter<?> c;
5837	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5838	>	MapReduceEntriesTask<K,V,U>
5839	>	t = (MapReduceEntriesTask<K,V,U>)c,
5840	>	s = t.rights;
5841	>	while (s != null) {
5842	>	U tr, sr;
5843	>	if ((sr = s.result) != null)
5844	>	t.result = (((tr = t.result) == null) ? sr :
5845	>	reducer.apply(tr, sr));
5846	>	s = t.rights = s.nextRight;
5847	>	}
5848	>	}
5849	>	}
5850	>	}
5851	>	}
5852	>
5853	>	@SuppressWarnings("serial") static final class MapReduceMappingsTask<K,V,U>
5854	>	extends Traverser<K,V,U> {
5855	>	final BiFun<? super K, ? super V, ? extends U> transformer;
5856	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5857	>	U result;
5858	>	MapReduceMappingsTask<K,V,U> rights, nextRight;
5859	>	MapReduceMappingsTask
5860	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5861	>	MapReduceMappingsTask<K,V,U> nextRight,
5862	>	BiFun<? super K, ? super V, ? extends U> transformer,
5863	>	BiFun<? super U, ? super U, ? extends U> reducer) {
5864	>	super(m, p, b); this.nextRight = nextRight;
5865	>	this.transformer = transformer;
5866	>	this.reducer = reducer;
5867	>	}
5868	>	public final U getRawResult() { return result; }
5869	>	@SuppressWarnings("unchecked") public final void compute() {
5870	>	final BiFun<? super K, ? super V, ? extends U> transformer;
5871	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5872	>	if ((transformer = this.transformer) != null &&
5873	>	(reducer = this.reducer) != null) {
5874	>	for (int b; (b = preSplit()) > 0;)
5875	>	(rights = new MapReduceMappingsTask<K,V,U>
5876	>	(map, this, b, rights, transformer, reducer)).fork();
5877	>	U r = null, u;
5878	>	Object v;
5879	>	while ((v = advance()) != null) {
5880	>	if ((u = transformer.apply((K)nextKey, (V)v)) != null)
5881	>	r = (r == null) ? u : reducer.apply(r, u);
5882	>	}
5883	>	result = r;
5884	>	CountedCompleter<?> c;
5885	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5886	>	MapReduceMappingsTask<K,V,U>
5887	>	t = (MapReduceMappingsTask<K,V,U>)c,
5888	>	s = t.rights;
5889	>	while (s != null) {
5890	>	U tr, sr;
5891	>	if ((sr = s.result) != null)
5892	>	t.result = (((tr = t.result) == null) ? sr :
5893	>	reducer.apply(tr, sr));
5894	>	s = t.rights = s.nextRight;
5895	>	}
5896	>	}
5897	>	}
5898	>	}
5899	>	}
5900	>
5901	>	@SuppressWarnings("serial") static final class MapReduceKeysToDoubleTask<K,V>
5902	>	extends Traverser<K,V,Double> {
5903	>	final ObjectToDouble<? super K> transformer;
5904	>	final DoubleByDoubleToDouble reducer;
5905	>	final double basis;
5906	>	double result;
5907	>	MapReduceKeysToDoubleTask<K,V> rights, nextRight;
5908	>	MapReduceKeysToDoubleTask
5909	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5910	>	MapReduceKeysToDoubleTask<K,V> nextRight,
5911	>	ObjectToDouble<? super K> transformer,
5912	>	double basis,
5913	>	DoubleByDoubleToDouble reducer) {
5914	>	super(m, p, b); this.nextRight = nextRight;
5915	>	this.transformer = transformer;
5916	>	this.basis = basis; this.reducer = reducer;
5917	>	}
5918	>	public final Double getRawResult() { return result; }
5919	>	@SuppressWarnings("unchecked") public final void compute() {
5920	>	final ObjectToDouble<? super K> transformer;
5921	>	final DoubleByDoubleToDouble reducer;
5922	>	if ((transformer = this.transformer) != null &&
5923	>	(reducer = this.reducer) != null) {
5924	>	double r = this.basis;
5925	>	for (int b; (b = preSplit()) > 0;)
5926	>	(rights = new MapReduceKeysToDoubleTask<K,V>
5927	>	(map, this, b, rights, transformer, r, reducer)).fork();
5928	>	while (advance() != null)
5929	>	r = reducer.apply(r, transformer.apply((K)nextKey));
5930	>	result = r;
5931	>	CountedCompleter<?> c;
5932	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5933	>	MapReduceKeysToDoubleTask<K,V>
5934	>	t = (MapReduceKeysToDoubleTask<K,V>)c,
5935	>	s = t.rights;
5936	>	while (s != null) {
5937	>	t.result = reducer.apply(t.result, s.result);
5938	>	s = t.rights = s.nextRight;
5939	>	}
5940	>	}
5941	>	}
5942	>	}
5943	>	}
5944	>
5945	>	@SuppressWarnings("serial") static final class MapReduceValuesToDoubleTask<K,V>
5946	>	extends Traverser<K,V,Double> {
5947	>	final ObjectToDouble<? super V> transformer;
5948	>	final DoubleByDoubleToDouble reducer;
5949	>	final double basis;
5950	>	double result;
5951	>	MapReduceValuesToDoubleTask<K,V> rights, nextRight;
5952	>	MapReduceValuesToDoubleTask
5953	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5954	>	MapReduceValuesToDoubleTask<K,V> nextRight,
5955	>	ObjectToDouble<? super V> transformer,
5956	>	double basis,
5957	>	DoubleByDoubleToDouble reducer) {
5958	>	super(m, p, b); this.nextRight = nextRight;
5959	>	this.transformer = transformer;
5960	>	this.basis = basis; this.reducer = reducer;
5961	>	}
5962	>	public final Double getRawResult() { return result; }
5963	>	@SuppressWarnings("unchecked") public final void compute() {
5964	>	final ObjectToDouble<? super V> transformer;
5965	>	final DoubleByDoubleToDouble reducer;
5966	>	if ((transformer = this.transformer) != null &&
5967	>	(reducer = this.reducer) != null) {
5968	>	double r = this.basis;
5969	>	for (int b; (b = preSplit()) > 0;)
5970	>	(rights = new MapReduceValuesToDoubleTask<K,V>
5971	>	(map, this, b, rights, transformer, r, reducer)).fork();
5972	>	Object v;
5973	>	while ((v = advance()) != null)
5974	>	r = reducer.apply(r, transformer.apply((V)v));
5975	>	result = r;
5976	>	CountedCompleter<?> c;
5977	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5978	>	MapReduceValuesToDoubleTask<K,V>
5979	>	t = (MapReduceValuesToDoubleTask<K,V>)c,
5980	>	s = t.rights;
5981	>	while (s != null) {
5982	>	t.result = reducer.apply(t.result, s.result);
5983	>	s = t.rights = s.nextRight;
5984	>	}
5985	>	}
5986	>	}
5987	>	}
5988	>	}
5989	>
5990	>	@SuppressWarnings("serial") static final class MapReduceEntriesToDoubleTask<K,V>
5991	>	extends Traverser<K,V,Double> {
5992	>	final ObjectToDouble<Map.Entry<K,V>> transformer;
5993	>	final DoubleByDoubleToDouble reducer;
5994	>	final double basis;
5995	>	double result;
5996	>	MapReduceEntriesToDoubleTask<K,V> rights, nextRight;
5997	>	MapReduceEntriesToDoubleTask
5998	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
5999	>	MapReduceEntriesToDoubleTask<K,V> nextRight,
6000	>	ObjectToDouble<Map.Entry<K,V>> transformer,
6001	>	double basis,
6002	>	DoubleByDoubleToDouble reducer) {
6003	>	super(m, p, b); this.nextRight = nextRight;
6004	>	this.transformer = transformer;
6005	>	this.basis = basis; this.reducer = reducer;
6006	>	}
6007	>	public final Double getRawResult() { return result; }
6008	>	@SuppressWarnings("unchecked") public final void compute() {
6009	>	final ObjectToDouble<Map.Entry<K,V>> transformer;
6010	>	final DoubleByDoubleToDouble reducer;
6011	>	if ((transformer = this.transformer) != null &&
6012	>	(reducer = this.reducer) != null) {
6013	>	double r = this.basis;
6014	>	for (int b; (b = preSplit()) > 0;)
6015	>	(rights = new MapReduceEntriesToDoubleTask<K,V>
6016	>	(map, this, b, rights, transformer, r, reducer)).fork();
6017	>	Object v;
6018	>	while ((v = advance()) != null)
6019	>	r = reducer.apply(r, transformer.apply(entryFor((K)nextKey,
6020	>	(V)v)));
6021	>	result = r;
6022	>	CountedCompleter<?> c;
6023	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6024	>	MapReduceEntriesToDoubleTask<K,V>
6025	>	t = (MapReduceEntriesToDoubleTask<K,V>)c,
6026	>	s = t.rights;
6027	>	while (s != null) {
6028	>	t.result = reducer.apply(t.result, s.result);
6029	>	s = t.rights = s.nextRight;
6030	>	}
6031	>	}
6032	>	}
6033	>	}
6034	>	}
6035	>
6036	>	@SuppressWarnings("serial") static final class MapReduceMappingsToDoubleTask<K,V>
6037	>	extends Traverser<K,V,Double> {
6038	>	final ObjectByObjectToDouble<? super K, ? super V> transformer;
6039	>	final DoubleByDoubleToDouble reducer;
6040	>	final double basis;
6041	>	double result;
6042	>	MapReduceMappingsToDoubleTask<K,V> rights, nextRight;
6043	>	MapReduceMappingsToDoubleTask
6044	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6045	>	MapReduceMappingsToDoubleTask<K,V> nextRight,
6046	>	ObjectByObjectToDouble<? super K, ? super V> transformer,
6047	>	double basis,
6048	>	DoubleByDoubleToDouble reducer) {
6049	>	super(m, p, b); this.nextRight = nextRight;
6050	>	this.transformer = transformer;
6051	>	this.basis = basis; this.reducer = reducer;
6052	>	}
6053	>	public final Double getRawResult() { return result; }
6054	>	@SuppressWarnings("unchecked") public final void compute() {
6055	>	final ObjectByObjectToDouble<? super K, ? super V> transformer;
6056	>	final DoubleByDoubleToDouble reducer;
6057	>	if ((transformer = this.transformer) != null &&
6058	>	(reducer = this.reducer) != null) {
6059	>	double r = this.basis;
6060	>	for (int b; (b = preSplit()) > 0;)
6061	>	(rights = new MapReduceMappingsToDoubleTask<K,V>
6062	>	(map, this, b, rights, transformer, r, reducer)).fork();
6063	>	Object v;
6064	>	while ((v = advance()) != null)
6065	>	r = reducer.apply(r, transformer.apply((K)nextKey, (V)v));
6066	>	result = r;
6067	>	CountedCompleter<?> c;
6068	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6069	>	MapReduceMappingsToDoubleTask<K,V>
6070	>	t = (MapReduceMappingsToDoubleTask<K,V>)c,
6071	>	s = t.rights;
6072	>	while (s != null) {
6073	>	t.result = reducer.apply(t.result, s.result);
6074	>	s = t.rights = s.nextRight;
6075	>	}
6076	>	}
6077	>	}
6078	>	}
6079	>	}
6080	>
6081	>	@SuppressWarnings("serial") static final class MapReduceKeysToLongTask<K,V>
6082	>	extends Traverser<K,V,Long> {
6083	>	final ObjectToLong<? super K> transformer;
6084	>	final LongByLongToLong reducer;
6085	>	final long basis;
6086	>	long result;
6087	>	MapReduceKeysToLongTask<K,V> rights, nextRight;
6088	>	MapReduceKeysToLongTask
6089	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6090	>	MapReduceKeysToLongTask<K,V> nextRight,
6091	>	ObjectToLong<? super K> transformer,
6092	>	long basis,
6093	>	LongByLongToLong reducer) {
6094	>	super(m, p, b); this.nextRight = nextRight;
6095	>	this.transformer = transformer;
6096	>	this.basis = basis; this.reducer = reducer;
6097	>	}
6098	>	public final Long getRawResult() { return result; }
6099	>	@SuppressWarnings("unchecked") public final void compute() {
6100	>	final ObjectToLong<? super K> transformer;
6101	>	final LongByLongToLong reducer;
6102	>	if ((transformer = this.transformer) != null &&
6103	>	(reducer = this.reducer) != null) {
6104	>	long r = this.basis;
6105	>	for (int b; (b = preSplit()) > 0;)
6106	>	(rights = new MapReduceKeysToLongTask<K,V>
6107	>	(map, this, b, rights, transformer, r, reducer)).fork();
6108	>	while (advance() != null)
6109	>	r = reducer.apply(r, transformer.apply((K)nextKey));
6110	>	result = r;
6111	>	CountedCompleter<?> c;
6112	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6113	>	MapReduceKeysToLongTask<K,V>
6114	>	t = (MapReduceKeysToLongTask<K,V>)c,
6115	>	s = t.rights;
6116	>	while (s != null) {
6117	>	t.result = reducer.apply(t.result, s.result);
6118	>	s = t.rights = s.nextRight;
6119	>	}
6120	>	}
6121	>	}
6122	>	}
6123	>	}
6124	>
6125	>	@SuppressWarnings("serial") static final class MapReduceValuesToLongTask<K,V>
6126	>	extends Traverser<K,V,Long> {
6127	>	final ObjectToLong<? super V> transformer;
6128	>	final LongByLongToLong reducer;
6129	>	final long basis;
6130	>	long result;
6131	>	MapReduceValuesToLongTask<K,V> rights, nextRight;
6132	>	MapReduceValuesToLongTask
6133	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6134	>	MapReduceValuesToLongTask<K,V> nextRight,
6135	>	ObjectToLong<? super V> transformer,
6136	>	long basis,
6137	>	LongByLongToLong reducer) {
6138	>	super(m, p, b); this.nextRight = nextRight;
6139	>	this.transformer = transformer;
6140	>	this.basis = basis; this.reducer = reducer;
6141	>	}
6142	>	public final Long getRawResult() { return result; }
6143	>	@SuppressWarnings("unchecked") public final void compute() {
6144	>	final ObjectToLong<? super V> transformer;
6145	>	final LongByLongToLong reducer;
6146	>	if ((transformer = this.transformer) != null &&
6147	>	(reducer = this.reducer) != null) {
6148	>	long r = this.basis;
6149	>	for (int b; (b = preSplit()) > 0;)
6150	>	(rights = new MapReduceValuesToLongTask<K,V>
6151	>	(map, this, b, rights, transformer, r, reducer)).fork();
6152	>	Object v;
6153	>	while ((v = advance()) != null)
6154	>	r = reducer.apply(r, transformer.apply((V)v));
6155	>	result = r;
6156	>	CountedCompleter<?> c;
6157	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6158	>	MapReduceValuesToLongTask<K,V>
6159	>	t = (MapReduceValuesToLongTask<K,V>)c,
6160	>	s = t.rights;
6161	>	while (s != null) {
6162	>	t.result = reducer.apply(t.result, s.result);
6163	>	s = t.rights = s.nextRight;
6164	>	}
6165	>	}
6166	>	}
6167	>	}
6168	>	}
6169	>
6170	>	@SuppressWarnings("serial") static final class MapReduceEntriesToLongTask<K,V>
6171	>	extends Traverser<K,V,Long> {
6172	>	final ObjectToLong<Map.Entry<K,V>> transformer;
6173	>	final LongByLongToLong reducer;
6174	>	final long basis;
6175	>	long result;
6176	>	MapReduceEntriesToLongTask<K,V> rights, nextRight;
6177	>	MapReduceEntriesToLongTask
6178	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6179	>	MapReduceEntriesToLongTask<K,V> nextRight,
6180	>	ObjectToLong<Map.Entry<K,V>> transformer,
6181	>	long basis,
6182	>	LongByLongToLong reducer) {
6183	>	super(m, p, b); this.nextRight = nextRight;
6184	>	this.transformer = transformer;
6185	>	this.basis = basis; this.reducer = reducer;
6186	>	}
6187	>	public final Long getRawResult() { return result; }
6188	>	@SuppressWarnings("unchecked") public final void compute() {
6189	>	final ObjectToLong<Map.Entry<K,V>> transformer;
6190	>	final LongByLongToLong reducer;
6191	>	if ((transformer = this.transformer) != null &&
6192	>	(reducer = this.reducer) != null) {
6193	>	long r = this.basis;
6194	>	for (int b; (b = preSplit()) > 0;)
6195	>	(rights = new MapReduceEntriesToLongTask<K,V>
6196	>	(map, this, b, rights, transformer, r, reducer)).fork();
6197	>	Object v;
6198	>	while ((v = advance()) != null)
6199	>	r = reducer.apply(r, transformer.apply(entryFor((K)nextKey,
6200	>	(V)v)));
6201	>	result = r;
6202	>	CountedCompleter<?> c;
6203	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6204	>	MapReduceEntriesToLongTask<K,V>
6205	>	t = (MapReduceEntriesToLongTask<K,V>)c,
6206	>	s = t.rights;
6207	>	while (s != null) {
6208	>	t.result = reducer.apply(t.result, s.result);
6209	>	s = t.rights = s.nextRight;
6210	>	}
6211	>	}
6212	>	}
6213	>	}
6214	>	}
6215	>
6216	>	@SuppressWarnings("serial") static final class MapReduceMappingsToLongTask<K,V>
6217	>	extends Traverser<K,V,Long> {
6218	>	final ObjectByObjectToLong<? super K, ? super V> transformer;
6219	>	final LongByLongToLong reducer;
6220	>	final long basis;
6221	>	long result;
6222	>	MapReduceMappingsToLongTask<K,V> rights, nextRight;
6223	>	MapReduceMappingsToLongTask
6224	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6225	>	MapReduceMappingsToLongTask<K,V> nextRight,
6226	>	ObjectByObjectToLong<? super K, ? super V> transformer,
6227	>	long basis,
6228	>	LongByLongToLong reducer) {
6229	>	super(m, p, b); this.nextRight = nextRight;
6230	>	this.transformer = transformer;
6231	>	this.basis = basis; this.reducer = reducer;
6232	>	}
6233	>	public final Long getRawResult() { return result; }
6234	>	@SuppressWarnings("unchecked") public final void compute() {
6235	>	final ObjectByObjectToLong<? super K, ? super V> transformer;
6236	>	final LongByLongToLong reducer;
6237	>	if ((transformer = this.transformer) != null &&
6238	>	(reducer = this.reducer) != null) {
6239	>	long r = this.basis;
6240	>	for (int b; (b = preSplit()) > 0;)
6241	>	(rights = new MapReduceMappingsToLongTask<K,V>
6242	>	(map, this, b, rights, transformer, r, reducer)).fork();
6243	>	Object v;
6244	>	while ((v = advance()) != null)
6245	>	r = reducer.apply(r, transformer.apply((K)nextKey, (V)v));
6246	>	result = r;
6247	>	CountedCompleter<?> c;
6248	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6249	>	MapReduceMappingsToLongTask<K,V>
6250	>	t = (MapReduceMappingsToLongTask<K,V>)c,
6251	>	s = t.rights;
6252	>	while (s != null) {
6253	>	t.result = reducer.apply(t.result, s.result);
6254	>	s = t.rights = s.nextRight;
6255	>	}
6256	>	}
6257	>	}
6258	>	}
6259	>	}
6260	>
6261	>	@SuppressWarnings("serial") static final class MapReduceKeysToIntTask<K,V>
6262	>	extends Traverser<K,V,Integer> {
6263	>	final ObjectToInt<? super K> transformer;
6264	>	final IntByIntToInt reducer;
6265	>	final int basis;
6266	>	int result;
6267	>	MapReduceKeysToIntTask<K,V> rights, nextRight;
6268	>	MapReduceKeysToIntTask
6269	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6270	>	MapReduceKeysToIntTask<K,V> nextRight,
6271	>	ObjectToInt<? super K> transformer,
6272	>	int basis,
6273	>	IntByIntToInt reducer) {
6274	>	super(m, p, b); this.nextRight = nextRight;
6275	>	this.transformer = transformer;
6276	>	this.basis = basis; this.reducer = reducer;
6277	>	}
6278	>	public final Integer getRawResult() { return result; }
6279	>	@SuppressWarnings("unchecked") public final void compute() {
6280	>	final ObjectToInt<? super K> transformer;
6281	>	final IntByIntToInt reducer;
6282	>	if ((transformer = this.transformer) != null &&
6283	>	(reducer = this.reducer) != null) {
6284	>	int r = this.basis;
6285	>	for (int b; (b = preSplit()) > 0;)
6286	>	(rights = new MapReduceKeysToIntTask<K,V>
6287	>	(map, this, b, rights, transformer, r, reducer)).fork();
6288	>	while (advance() != null)
6289	>	r = reducer.apply(r, transformer.apply((K)nextKey));
6290	>	result = r;
6291	>	CountedCompleter<?> c;
6292	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6293	>	MapReduceKeysToIntTask<K,V>
6294	>	t = (MapReduceKeysToIntTask<K,V>)c,
6295	>	s = t.rights;
6296	>	while (s != null) {
6297	>	t.result = reducer.apply(t.result, s.result);
6298	>	s = t.rights = s.nextRight;
6299	>	}
6300	>	}
6301	>	}
6302	>	}
6303	>	}
6304	>
6305	>	@SuppressWarnings("serial") static final class MapReduceValuesToIntTask<K,V>
6306	>	extends Traverser<K,V,Integer> {
6307	>	final ObjectToInt<? super V> transformer;
6308	>	final IntByIntToInt reducer;
6309	>	final int basis;
6310	>	int result;
6311	>	MapReduceValuesToIntTask<K,V> rights, nextRight;
6312	>	MapReduceValuesToIntTask
6313	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6314	>	MapReduceValuesToIntTask<K,V> nextRight,
6315	>	ObjectToInt<? super V> transformer,
6316	>	int basis,
6317	>	IntByIntToInt reducer) {
6318	>	super(m, p, b); this.nextRight = nextRight;
6319	>	this.transformer = transformer;
6320	>	this.basis = basis; this.reducer = reducer;
6321	>	}
6322	>	public final Integer getRawResult() { return result; }
6323	>	@SuppressWarnings("unchecked") public final void compute() {
6324	>	final ObjectToInt<? super V> transformer;
6325	>	final IntByIntToInt reducer;
6326	>	if ((transformer = this.transformer) != null &&
6327	>	(reducer = this.reducer) != null) {
6328	>	int r = this.basis;
6329	>	for (int b; (b = preSplit()) > 0;)
6330	>	(rights = new MapReduceValuesToIntTask<K,V>
6331	>	(map, this, b, rights, transformer, r, reducer)).fork();
6332	>	Object v;
6333	>	while ((v = advance()) != null)
6334	>	r = reducer.apply(r, transformer.apply((V)v));
6335	>	result = r;
6336	>	CountedCompleter<?> c;
6337	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6338	>	MapReduceValuesToIntTask<K,V>
6339	>	t = (MapReduceValuesToIntTask<K,V>)c,
6340	>	s = t.rights;
6341	>	while (s != null) {
6342	>	t.result = reducer.apply(t.result, s.result);
6343	>	s = t.rights = s.nextRight;
6344	>	}
6345	>	}
6346	>	}
6347	>	}
6348	>	}
6349	>
6350	>	@SuppressWarnings("serial") static final class MapReduceEntriesToIntTask<K,V>
6351	>	extends Traverser<K,V,Integer> {
6352	>	final ObjectToInt<Map.Entry<K,V>> transformer;
6353	>	final IntByIntToInt reducer;
6354	>	final int basis;
6355	>	int result;
6356	>	MapReduceEntriesToIntTask<K,V> rights, nextRight;
6357	>	MapReduceEntriesToIntTask
6358	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6359	>	MapReduceEntriesToIntTask<K,V> nextRight,
6360	>	ObjectToInt<Map.Entry<K,V>> transformer,
6361	>	int basis,
6362	>	IntByIntToInt reducer) {
6363	>	super(m, p, b); this.nextRight = nextRight;
6364	>	this.transformer = transformer;
6365	>	this.basis = basis; this.reducer = reducer;
6366	>	}
6367	>	public final Integer getRawResult() { return result; }
6368	>	@SuppressWarnings("unchecked") public final void compute() {
6369	>	final ObjectToInt<Map.Entry<K,V>> transformer;
6370	>	final IntByIntToInt reducer;
6371	>	if ((transformer = this.transformer) != null &&
6372	>	(reducer = this.reducer) != null) {
6373	>	int r = this.basis;
6374	>	for (int b; (b = preSplit()) > 0;)
6375	>	(rights = new MapReduceEntriesToIntTask<K,V>
6376	>	(map, this, b, rights, transformer, r, reducer)).fork();
6377	>	Object v;
6378	>	while ((v = advance()) != null)
6379	>	r = reducer.apply(r, transformer.apply(entryFor((K)nextKey,
6380	>	(V)v)));
6381	>	result = r;
6382	>	CountedCompleter<?> c;
6383	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6384	>	MapReduceEntriesToIntTask<K,V>
6385	>	t = (MapReduceEntriesToIntTask<K,V>)c,
6386	>	s = t.rights;
6387	>	while (s != null) {
6388	>	t.result = reducer.apply(t.result, s.result);
6389	>	s = t.rights = s.nextRight;
6390	>	}
6391	>	}
6392	>	}
6393	>	}
6394	>	}
6395	>
6396	>	@SuppressWarnings("serial") static final class MapReduceMappingsToIntTask<K,V>
6397	>	extends Traverser<K,V,Integer> {
6398	>	final ObjectByObjectToInt<? super K, ? super V> transformer;
6399	>	final IntByIntToInt reducer;
6400	>	final int basis;
6401	>	int result;
6402	>	MapReduceMappingsToIntTask<K,V> rights, nextRight;
6403	>	MapReduceMappingsToIntTask
6404	>	(ConcurrentHashMapV8<K,V> m, Traverser<K,V,?> p, int b,
6405	>	MapReduceMappingsToIntTask<K,V> nextRight,
6406	>	ObjectByObjectToInt<? super K, ? super V> transformer,
6407	>	int basis,
6408	>	IntByIntToInt reducer) {
6409	>	super(m, p, b); this.nextRight = nextRight;
6410	>	this.transformer = transformer;
6411	>	this.basis = basis; this.reducer = reducer;
6412	>	}
6413	>	public final Integer getRawResult() { return result; }
6414	>	@SuppressWarnings("unchecked") public final void compute() {
6415	>	final ObjectByObjectToInt<? super K, ? super V> transformer;
6416	>	final IntByIntToInt reducer;
6417	>	if ((transformer = this.transformer) != null &&
6418	>	(reducer = this.reducer) != null) {
6419	>	int r = this.basis;
6420	>	for (int b; (b = preSplit()) > 0;)
6421	>	(rights = new MapReduceMappingsToIntTask<K,V>
6422	>	(map, this, b, rights, transformer, r, reducer)).fork();
6423	>	Object v;
6424	>	while ((v = advance()) != null)
6425	>	r = reducer.apply(r, transformer.apply((K)nextKey, (V)v));
6426	>	result = r;
6427	>	CountedCompleter<?> c;
6428	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6429	>	MapReduceMappingsToIntTask<K,V>
6430	>	t = (MapReduceMappingsToIntTask<K,V>)c,
6431	>	s = t.rights;
6432	>	while (s != null) {
6433	>	t.result = reducer.apply(t.result, s.result);
6434	>	s = t.rights = s.nextRight;
6435	>	}
6436	>	}
6437	>	}
6438		}
6439		}
6440
6441		// Unsafe mechanics
6442	<	private static final sun.misc.Unsafe UNSAFE;
6443	<	private static final long counterOffset;
6444	<	private static final long resizingOffset;
6442	>	private static final sun.misc.Unsafe U;
6443	>	private static final long SIZECTL;
6444	>	private static final long TRANSFERINDEX;
6445	>	private static final long TRANSFERORIGIN;
6446	>	private static final long BASECOUNT;
6447	>	private static final long COUNTERBUSY;
6448	>	private static final long CELLVALUE;
6449		private static final long ABASE;
6450		private static final int ASHIFT;
6451
6452		static {
6453		int ss;
6454		try {
6455	<	UNSAFE = getUnsafe();
6455	>	U = getUnsafe();
6456		Class<?> k = ConcurrentHashMapV8.class;
6457	<	counterOffset = UNSAFE.objectFieldOffset
6458	<	(k.getDeclaredField("counter"));
6459	<	resizingOffset = UNSAFE.objectFieldOffset
6460	<	(k.getDeclaredField("resizing"));
6457	>	SIZECTL = U.objectFieldOffset
6458	>	(k.getDeclaredField("sizeCtl"));
6459	>	TRANSFERINDEX = U.objectFieldOffset
6460	>	(k.getDeclaredField("transferIndex"));
6461	>	TRANSFERORIGIN = U.objectFieldOffset
6462	>	(k.getDeclaredField("transferOrigin"));
6463	>	BASECOUNT = U.objectFieldOffset
6464	>	(k.getDeclaredField("baseCount"));
6465	>	COUNTERBUSY = U.objectFieldOffset
6466	>	(k.getDeclaredField("counterBusy"));
6467	>	Class<?> ck = CounterCell.class;
6468	>	CELLVALUE = U.objectFieldOffset
6469	>	(ck.getDeclaredField("value"));
6470		Class<?> sc = Node[].class;
6471	<	ABASE = UNSAFE.arrayBaseOffset(sc);
6472	<	ss = UNSAFE.arrayIndexScale(sc);
6471	>	ABASE = U.arrayBaseOffset(sc);
6472	>	ss = U.arrayIndexScale(sc);
6473	>	ASHIFT = 31 - Integer.numberOfLeadingZeros(ss);
6474		} catch (Exception e) {
6475		throw new Error(e);
6476		}
6477		if ((ss & (ss-1)) != 0)
6478		throw new Error("data type scale not a power of two");
1592	–	ASHIFT = 31 - Integer.numberOfLeadingZeros(ss);
6479		}
6480
6481		/**
#	Line 1619 | Line 6505 | public class ConcurrentHashMapV8<K, V>
6505		}
6506		}
6507		}
1622	–
6508		}

Diff Legend

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