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 1.12 by jsr166, Tue Aug 30 18:31:54 2011 UTC vs.
Revision 1.125 by jsr166, Sun Sep 13 16:28:14 2015 UTC

#	Line 5 | Line 5
5		*/
6
7		package jsr166e;
8	<	import jsr166e.LongAdder;
9	<	import java.util.Map;
10	<	import java.util.Set;
11	<	import java.util.Collection;
8	>
9	>	import jsr166e.ForkJoinPool;
10	>
11	>	import java.io.ObjectStreamField;
12	>	import java.io.Serializable;
13	>	import java.lang.reflect.ParameterizedType;
14	>	import java.lang.reflect.Type;
15		import java.util.AbstractMap;
16	<	import java.util.AbstractSet;
17	<	import java.util.AbstractCollection;
18	<	import java.util.Hashtable;
16	>	import java.util.Arrays;
17	>	import java.util.Collection;
18	>	import java.util.ConcurrentModificationException;
19	>	import java.util.Enumeration;
20		import java.util.HashMap;
21	+	import java.util.Hashtable;
22		import java.util.Iterator;
23	<	import java.util.Enumeration;
19	<	import java.util.ConcurrentModificationException;
23	>	import java.util.Map;
24		import java.util.NoSuchElementException;
25	+	import java.util.Set;
26		import java.util.concurrent.ConcurrentMap;
27	<	import java.io.Serializable;
27	>	import java.util.concurrent.atomic.AtomicReference;
28	>	import java.util.concurrent.atomic.AtomicInteger;
29	>	import java.util.concurrent.locks.LockSupport;
30	>	import java.util.concurrent.locks.ReentrantLock;
31
32		/**
33		* A hash table supporting full concurrency of retrievals and
#	Line 33 | Line 41 | import java.io.Serializable;
41		* interoperable with {@code Hashtable} in programs that rely on its
42		* thread safety but not on its synchronization details.
43		*
44	<	* <p> Retrieval operations (including {@code get}) generally do not
44	>	* <p>Retrieval operations (including {@code get}) generally do not
45		* block, so may overlap with update operations (including {@code put}
46		* and {@code remove}). Retrievals reflect the results of the most
47		* recently <em>completed</em> update operations holding upon their
48	<	* onset.  For aggregate operations such as {@code putAll} and {@code
49	<	* clear}, concurrent retrievals may reflect insertion or removal of
50	<	* only some entries.  Similarly, Iterators and Enumerations return
51	<	* elements reflecting the state of the hash table at some point at or
52	<	* since the creation of the iterator/enumeration.  They do
53	<	* <em>not</em> throw {@link ConcurrentModificationException}.
54	<	* However, iterators are designed to be used by only one thread at a
55	<	* time.  Bear in mind that the results of aggregate status methods
56	<	* including {@code size}, {@code isEmpty}, and {@code containsValue}
57	<	* are typically useful only when a map is not undergoing concurrent
58	<	* updates in other threads.  Otherwise the results of these methods
59	<	* reflect transient states that may be adequate for monitoring
60	<	* purposes, but not for program control.
61	<	*
62	<	* <p> Resizing this or any other kind of hash table is a relatively
63	<	* slow operation, so, when possible, it is a good idea to provide
64	<	* estimates of expected table sizes in constructors. Also, for
65	<	* compatibility with previous versions of this class, constructors
66	<	* may optionally specify an expected {@code concurrencyLevel} as an
67	<	* additional hint for internal sizing.
48	>	* onset. (More formally, an update operation for a given key bears a
49	>	* <em>happens-before</em> relation with any (non-null) retrieval for
50	>	* that key reporting the updated value.)  For aggregate operations
51	>	* such as {@code putAll} and {@code clear}, concurrent retrievals may
52	>	* reflect insertion or removal of only some entries.  Similarly,
53	>	* Iterators and Enumerations return elements reflecting the state of
54	>	* the hash table at some point at or since the creation of the
55	>	* iterator/enumeration.  They do <em>not</em> throw {@link
56	>	* ConcurrentModificationException}.  However, iterators are designed
57	>	* to be used by only one thread at a time.  Bear in mind that the
58	>	* results of aggregate status methods including {@code size}, {@code
59	>	* isEmpty}, and {@code containsValue} are typically useful only when
60	>	* a map is not undergoing concurrent updates in other threads.
61	>	* Otherwise the results of these methods reflect transient states
62	>	* that may be adequate for monitoring or estimation purposes, but not
63	>	* for program control.
64	>	*
65	>	* <p>The table is dynamically expanded when there are too many
66	>	* collisions (i.e., keys that have distinct hash codes but fall into
67	>	* the same slot modulo the table size), with the expected average
68	>	* effect of maintaining roughly two bins per mapping (corresponding
69	>	* to a 0.75 load factor threshold for resizing). There may be much
70	>	* variance around this average as mappings are added and removed, but
71	>	* overall, this maintains a commonly accepted time/space tradeoff for
72	>	* hash tables.  However, resizing this or any other kind of hash
73	>	* table may be a relatively slow operation. When possible, it is a
74	>	* good idea to provide a size estimate as an optional {@code
75	>	* initialCapacity} constructor argument. An additional optional
76	>	* {@code loadFactor} constructor argument provides a further means of
77	>	* customizing initial table capacity by specifying the table density
78	>	* to be used in calculating the amount of space to allocate for the
79	>	* given number of elements.  Also, for compatibility with previous
80	>	* versions of this class, constructors may optionally specify an
81	>	* expected {@code concurrencyLevel} as an additional hint for
82	>	* internal sizing.  Note that using many keys with exactly the same
83	>	* {@code hashCode()} is a sure way to slow down performance of any
84	>	* hash table. To ameliorate impact, when keys are {@link Comparable},
85	>	* this class may use comparison order among keys to help break ties.
86	>	*
87	>	* <p>A {@link Set} projection of a ConcurrentHashMapV8 may be created
88	>	* (using {@link #newKeySet()} or {@link #newKeySet(int)}), or viewed
89	>	* (using {@link #keySet(Object)} when only keys are of interest, and the
90	>	* mapped values are (perhaps transiently) not used or all take the
91	>	* same mapping value.
92		*
93		* <p>This class and its views and iterators implement all of the
94		* <em>optional</em> methods of the {@link Map} and {@link Iterator}
95		* interfaces.
96		*
97	<	* <p> Like {@link Hashtable} but unlike {@link HashMap}, this class
97	>	* <p>Like {@link Hashtable} but unlike {@link HashMap}, this class
98		* does <em>not</em> allow {@code null} to be used as a key or value.
99		*
100	+	* <p>ConcurrentHashMapV8s support a set of sequential and parallel bulk
101	+	* operations that are designed
102	+	* to be safely, and often sensibly, applied even with maps that are
103	+	* being concurrently updated by other threads; for example, when
104	+	* computing a snapshot summary of the values in a shared registry.
105	+	* There are three kinds of operation, each with four forms, accepting
106	+	* functions with Keys, Values, Entries, and (Key, Value) arguments
107	+	* and/or return values. Because the elements of a ConcurrentHashMapV8
108	+	* are not ordered in any particular way, and may be processed in
109	+	* different orders in different parallel executions, the correctness
110	+	* of supplied functions should not depend on any ordering, or on any
111	+	* other objects or values that may transiently change while
112	+	* computation is in progress; and except for forEach actions, should
113	+	* ideally be side-effect-free. Bulk operations on {@link java.util.Map.Entry}
114	+	* objects do not support method {@code setValue}.
115	+	*
116	+	* <ul>
117	+	* <li>forEach: Perform a given action on each element.
118	+	* A variant form applies a given transformation on each element
119	+	* before performing the action.
120	+	*
121	+	* <li>search: Return the first available non-null result of
122	+	* applying a given function on each element; skipping further
123	+	* search when a result is found.
124	+	*
125	+	* <li>reduce: Accumulate each element.  The supplied reduction
126	+	* function cannot rely on ordering (more formally, it should be
127	+	* both associative and commutative).  There are five variants:
128	+	*
129	+	* <ul>
130	+	*
131	+	* <li>Plain reductions. (There is not a form of this method for
132	+	* (key, value) function arguments since there is no corresponding
133	+	* return type.)
134	+	*
135	+	* <li>Mapped reductions that accumulate the results of a given
136	+	* function applied to each element.
137	+	*
138	+	* <li>Reductions to scalar doubles, longs, and ints, using a
139	+	* given basis value.
140	+	*
141	+	* </ul>
142	+	* </ul>
143	+	*
144	+	* <p>These bulk operations accept a {@code parallelismThreshold}
145	+	* argument. Methods proceed sequentially if the current map size is
146	+	* estimated to be less than the given threshold. Using a value of
147	+	* {@code Long.MAX_VALUE} suppresses all parallelism.  Using a value
148	+	* of {@code 1} results in maximal parallelism by partitioning into
149	+	* enough subtasks to fully utilize the {@link
150	+	* ForkJoinPool#commonPool()} that is used for all parallel
151	+	* computations. Normally, you would initially choose one of these
152	+	* extreme values, and then measure performance of using in-between
153	+	* values that trade off overhead versus throughput.
154	+	*
155	+	* <p>The concurrency properties of bulk operations follow
156	+	* from those of ConcurrentHashMapV8: Any non-null result returned
157	+	* from {@code get(key)} and related access methods bears a
158	+	* happens-before relation with the associated insertion or
159	+	* update.  The result of any bulk operation reflects the
160	+	* composition of these per-element relations (but is not
161	+	* necessarily atomic with respect to the map as a whole unless it
162	+	* is somehow known to be quiescent).  Conversely, because keys
163	+	* and values in the map are never null, null serves as a reliable
164	+	* atomic indicator of the current lack of any result.  To
165	+	* maintain this property, null serves as an implicit basis for
166	+	* all non-scalar reduction operations. For the double, long, and
167	+	* int versions, the basis should be one that, when combined with
168	+	* any other value, returns that other value (more formally, it
169	+	* should be the identity element for the reduction). Most common
170	+	* reductions have these properties; for example, computing a sum
171	+	* with basis 0 or a minimum with basis MAX_VALUE.
172	+	*
173	+	* <p>Search and transformation functions provided as arguments
174	+	* should similarly return null to indicate the lack of any result
175	+	* (in which case it is not used). In the case of mapped
176	+	* reductions, this also enables transformations to serve as
177	+	* filters, returning null (or, in the case of primitive
178	+	* specializations, the identity basis) if the element should not
179	+	* be combined. You can create compound transformations and
180	+	* filterings by composing them yourself under this "null means
181	+	* there is nothing there now" rule before using them in search or
182	+	* reduce operations.
183	+	*
184	+	* <p>Methods accepting and/or returning Entry arguments maintain
185	+	* key-value associations. They may be useful for example when
186	+	* finding the key for the greatest value. Note that "plain" Entry
187	+	* arguments can be supplied using {@code new
188	+	* AbstractMap.SimpleEntry(k,v)}.
189	+	*
190	+	* <p>Bulk operations may complete abruptly, throwing an
191	+	* exception encountered in the application of a supplied
192	+	* function. Bear in mind when handling such exceptions that other
193	+	* concurrently executing functions could also have thrown
194	+	* exceptions, or would have done so if the first exception had
195	+	* not occurred.
196	+	*
197	+	* <p>Speedups for parallel compared to sequential forms are common
198	+	* but not guaranteed.  Parallel operations involving brief functions
199	+	* on small maps may execute more slowly than sequential forms if the
200	+	* underlying work to parallelize the computation is more expensive
201	+	* than the computation itself.  Similarly, parallelization may not
202	+	* lead to much actual parallelism if all processors are busy
203	+	* performing unrelated tasks.
204	+	*
205	+	* <p>All arguments to all task methods must be non-null.
206	+	*
207	+	* <p><em>jsr166e note: During transition, this class
208	+	* uses nested functional interfaces with different names but the
209	+	* same forms as those expected for JDK8.</em>
210	+	*
211		* <p>This class is a member of the
212		* <a href="{@docRoot}/../technotes/guides/collections/index.html">
213		* Java Collections Framework</a>.
214		*
72	–	* <p><em>jsr166e note: This class is a candidate replacement for
73	–	* java.util.concurrent.ConcurrentHashMap.<em>
74	–	*
215		* @since 1.5
216		* @author Doug Lea
217		* @param <K> the type of keys maintained by this map
218		* @param <V> the type of mapped values
219		*/
220	<	public class ConcurrentHashMapV8<K, V>
221	<	implements ConcurrentMap<K, V>, Serializable {
220	>	public class ConcurrentHashMapV8<K,V> extends AbstractMap<K,V>
221	>	implements ConcurrentMap<K,V>, Serializable {
222		private static final long serialVersionUID = 7249069246763182397L;
223
224		/**
225	<	* A function computing a mapping from the given key to a value,
226	<	* or {@code null} if there is no mapping. This is a place-holder
227	<	* for an upcoming JDK8 interface.
228	<	*/
229	<	public static interface MappingFunction<K, V> {
230	<	/**
231	<	* Returns a value for the given key, or null if there is no
232	<	* mapping. If this function throws an (unchecked) exception,
233	<	* the exception is rethrown to its caller, and no mapping is
234	<	* recorded.  Because this function is invoked within
95	<	* atomicity control, the computation should be short and
96	<	* simple. The most common usage is to construct a new object
97	<	* serving as an initial mapped value.
98	<	*
99	<	* @param key the (non-null) key
100	<	* @return a value, or null if none
225	>	* An object for traversing and partitioning elements of a source.
226	>	* This interface provides a subset of the functionality of JDK8
227	>	* java.util.Spliterator.
228	>	*/
229	>	public static interface ConcurrentHashMapSpliterator<T> {
230	>	/**
231	>	* If possible, returns a new spliterator covering
232	>	* approximately one half of the elements, which will not be
233	>	* covered by this spliterator. Returns null if cannot be
234	>	* split.
235		*/
236	<	V map(K key);
237	<	}
236	>	ConcurrentHashMapSpliterator<T> trySplit();
237	>	/**
238	>	* Returns an estimate of the number of elements covered by
239	>	* this Spliterator.
240	>	*/
241	>	long estimateSize();
242	>
243	>	/** Applies the action to each untraversed element */
244	>	void forEachRemaining(Action<? super T> action);
245	>	/** If an element remains, applies the action and returns true. */
246	>	boolean tryAdvance(Action<? super T> action);
247	>	}
248	>
249	>	// Sams
250	>	/** Interface describing a void action of one argument */
251	>	public interface Action<A> { void apply(A a); }
252	>	/** Interface describing a void action of two arguments */
253	>	public interface BiAction<A,B> { void apply(A a, B b); }
254	>	/** Interface describing a function of one argument */
255	>	public interface Fun<A,T> { T apply(A a); }
256	>	/** Interface describing a function of two arguments */
257	>	public interface BiFun<A,B,T> { T apply(A a, B b); }
258	>	/** Interface describing a function mapping its argument to a double */
259	>	public interface ObjectToDouble<A> { double apply(A a); }
260	>	/** Interface describing a function mapping its argument to a long */
261	>	public interface ObjectToLong<A> { long apply(A a); }
262	>	/** Interface describing a function mapping its argument to an int */
263	>	public interface ObjectToInt<A> {int apply(A a); }
264	>	/** Interface describing a function mapping two arguments to a double */
265	>	public interface ObjectByObjectToDouble<A,B> { double apply(A a, B b); }
266	>	/** Interface describing a function mapping two arguments to a long */
267	>	public interface ObjectByObjectToLong<A,B> { long apply(A a, B b); }
268	>	/** Interface describing a function mapping two arguments to an int */
269	>	public interface ObjectByObjectToInt<A,B> {int apply(A a, B b); }
270	>	/** Interface describing a function mapping two doubles to a double */
271	>	public interface DoubleByDoubleToDouble { double apply(double a, double b); }
272	>	/** Interface describing a function mapping two longs to a long */
273	>	public interface LongByLongToLong { long apply(long a, long b); }
274	>	/** Interface describing a function mapping two ints to an int */
275	>	public interface IntByIntToInt { int apply(int a, int b); }
276	>
277
278		/*
279		* Overview:
#	Line 108 | Line 281 | public class ConcurrentHashMapV8<K, V>
281		* The primary design goal of this hash table is to maintain
282		* concurrent readability (typically method get(), but also
283		* iterators and related methods) while minimizing update
284	<	* contention.
285	<	*
286	<	* Each key-value mapping is held in a Node.  Because Node fields
287	<	* can contain special values, they are defined using plain Object
288	<	* types. Similarly in turn, all internal methods that use them
289	<	* work off Object types. All public generic-typed methods relay
290	<	* in/out of these internal methods, supplying casts as needed.
284	>	* contention. Secondary goals are to keep space consumption about
285	>	* the same or better than java.util.HashMap, and to support high
286	>	* initial insertion rates on an empty table by many threads.
287	>	*
288	>	* This map usually acts as a binned (bucketed) hash table.  Each
289	>	* key-value mapping is held in a Node.  Most nodes are instances
290	>	* of the basic Node class with hash, key, value, and next
291	>	* fields. However, various subclasses exist: TreeNodes are
292	>	* arranged in balanced trees, not lists.  TreeBins hold the roots
293	>	* of sets of TreeNodes. ForwardingNodes are placed at the heads
294	>	* of bins during resizing. ReservationNodes are used as
295	>	* placeholders while establishing values in computeIfAbsent and
296	>	* related methods.  The types TreeBin, ForwardingNode, and
297	>	* ReservationNode do not hold normal user keys, values, or
298	>	* hashes, and are readily distinguishable during search etc
299	>	* because they have negative hash fields and null key and value
300	>	* fields. (These special nodes are either uncommon or transient,
301	>	* so the impact of carrying around some unused fields is
302	>	* insignificant.)
303		*
304		* The table is lazily initialized to a power-of-two size upon the
305	<	* first insertion.  Each bin in the table contains a (typically
306	<	* short) list of Nodes.  Table accesses require volatile/atomic
307	<	* reads, writes, and CASes.  Because there is no other way to
308	<	* arrange this without adding further indirections, we use
309	<	* intrinsics (sun.misc.Unsafe) operations.  The lists of nodes
310	<	* within bins are always accurately traversable under volatile
311	<	* reads, so long as lookups check hash code and non-nullness of
312	<	* key and value before checking key equality. (All valid hash
313	<	* codes are nonnegative. Negative values are reserved for special
314	<	* forwarding nodes; see below.)
315	<	*
316	<	* A bin may be locked during update (insert, delete, and replace)
317	<	* operations.  We do not want to waste the space required to
318	<	* associate a distinct lock object with each bin, so instead use
319	<	* the first node of a bin list itself as a lock, using builtin
320	<	* "synchronized" locks. These save space and we can live with
321	<	* only plain block-structured lock/unlock operations. Using the
322	<	* first node of a list as a lock does not by itself suffice
323	<	* though: When a node is locked, any update must first validate
324	<	* that it is still the first node, and retry if not. (Because new
325	<	* nodes are always appended to lists, once a node is first in a
326	<	* bin, it remains first until deleted or the bin becomes
327	<	* invalidated.)  However, update operations can and sometimes do
328	<	* still traverse the bin until the point of update, which helps
329	<	* reduce cache misses on retries.  This is a converse of sorts to
330	<	* the lazy locking technique described by Herlihy & Shavit. If
331	<	* there is no existing node during a put operation, then one can
332	<	* be CAS'ed in (without need for lock except in computeIfAbsent);
333	<	* the CAS serves as validation. This is on average the most
334	<	* common case for put operations -- under random hash codes, the
335	<	* distribution of nodes in bins follows a Poisson distribution
336	<	* (see http://en.wikipedia.org/wiki/Poisson_distribution) with a
337	<	* parameter of 0.5 on average under the default loadFactor of
338	<	* 0.75.  The expected number of locks covering different elements
339	<	* (i.e., bins with 2 or more nodes) is approximately 10% at
340	<	* steady state under default settings.  Lock contention
341	<	* probability for two threads accessing arbitrary distinct
342	<	* elements is, roughly, 1 / (8 * #elements).
343	<	*
344	<	* The table is resized when occupancy exceeds a threshold.  Only
345	<	* a single thread performs the resize (using field "resizing", to
346	<	* arrange exclusion), but the table otherwise remains usable for
347	<	* both reads and updates. Resizing proceeds by transferring bins,
348	<	* one by one, from the table to the next table.  Upon transfer,
349	<	* the old table bin contains only a special forwarding node (with
350	<	* negative hash code ("MOVED")) that contains the next table as
351	<	* its key. On encountering a forwarding node, access and update
352	<	* operations restart, using the new table. To ensure concurrent
353	<	* readability of traversals, transfers must proceed from the last
354	<	* bin (table.length - 1) up towards the first.  Any traversal
355	<	* starting from the first bin can then arrange to move to the new
356	<	* table for the rest of the traversal without revisiting nodes.
357	<	* This constrains bin transfers to a particular order, and so can
358	<	* block indefinitely waiting for the next lock, and other threads
359	<	* cannot help with the transfer. However, expected stalls are
360	<	* infrequent enough to not warrant the additional overhead and
361	<	* complexity of access and iteration schemes that could admit
362	<	* out-of-order or concurrent bin transfers.
363	<	*
364	<	* A similar traversal scheme (not yet implemented) can apply to
365	<	* partial traversals during partitioned aggregate operations.
366	<	* Also, read-only operations give up if ever forwarded to a null
367	<	* table, which provides support for shutdown-style clearing,
368	<	* which is also not currently implemented.
369	<	*
370	<	* The element count is maintained using a LongAdder, which avoids
371	<	* contention on updates but can encounter cache thrashing if read
372	<	* too frequently during concurrent updates. To avoid reading so
373	<	* often, resizing is normally attempted only upon adding to a bin
374	<	* already holding two or more nodes. Under the default threshold
375	<	* (0.75), and uniform hash distributions, the probability of this
376	<	* occurring at threshold is around 13%, meaning that only about 1
377	<	* in 8 puts check threshold (and after resizing, many fewer do
378	<	* so). But this approximation has high variance for small table
379	<	* sizes, so we check on any collision for sizes <= 64.  Further,
380	<	* to increase the probability that a resize occurs soon enough, we
381	<	* offset the threshold (see THRESHOLD_OFFSET) by the expected
382	<	* number of puts between checks. This is currently set to 8, in
383	<	* accord with the default load factor. In practice, this is
384	<	* rarely overridden, and in any case is close enough to other
385	<	* plausible values not to waste dynamic probability computation
386	<	* for more precision.
305	>	* first insertion.  Each bin in the table normally contains a
306	>	* list of Nodes (most often, the list has only zero or one Node).
307	>	* Table accesses require volatile/atomic reads, writes, and
308	>	* CASes.  Because there is no other way to arrange this without
309	>	* adding further indirections, we use intrinsics
310	>	* (sun.misc.Unsafe) operations.
311	>	*
312	>	* We use the top (sign) bit of Node hash fields for control
313	>	* purposes -- it is available anyway because of addressing
314	>	* constraints.  Nodes with negative hash fields are specially
315	>	* handled or ignored in map methods.
316	>	*
317	>	* Insertion (via put or its variants) of the first node in an
318	>	* empty bin is performed by just CASing it to the bin.  This is
319	>	* by far the most common case for put operations under most
320	>	* key/hash distributions.  Other update operations (insert,
321	>	* delete, and replace) require locks.  We do not want to waste
322	>	* the space required to associate a distinct lock object with
323	>	* each bin, so instead use the first node of a bin list itself as
324	>	* a lock. Locking support for these locks relies on builtin
325	>	* "synchronized" monitors.
326	>	*
327	>	* Using the first node of a list as a lock does not by itself
328	>	* suffice though: When a node is locked, any update must first
329	>	* validate that it is still the first node after locking it, and
330	>	* retry if not. Because new nodes are always appended to lists,
331	>	* once a node is first in a bin, it remains first until deleted
332	>	* or the bin becomes invalidated (upon resizing).
333	>	*
334	>	* The main disadvantage of per-bin locks is that other update
335	>	* operations on other nodes in a bin list protected by the same
336	>	* lock can stall, for example when user equals() or mapping
337	>	* functions take a long time.  However, statistically, under
338	>	* random hash codes, this is not a common problem.  Ideally, the
339	>	* frequency of nodes in bins follows a Poisson distribution
340	>	* (http://en.wikipedia.org/wiki/Poisson_distribution) with a
341	>	* parameter of about 0.5 on average, given the resizing threshold
342	>	* of 0.75, although with a large variance because of resizing
343	>	* granularity. Ignoring variance, the expected occurrences of
344	>	* list size k are (exp(-0.5) * pow(0.5, k) / factorial(k)). The
345	>	* first values are:
346	>	*
347	>	* 0:    0.60653066
348	>	* 1:    0.30326533
349	>	* 2:    0.07581633
350	>	* 3:    0.01263606
351	>	* 4:    0.00157952
352	>	* 5:    0.00015795
353	>	* 6:    0.00001316
354	>	* 7:    0.00000094
355	>	* 8:    0.00000006
356	>	* more: less than 1 in ten million
357	>	*
358	>	* Lock contention probability for two threads accessing distinct
359	>	* elements is roughly 1 / (8 * #elements) under random hashes.
360	>	*
361	>	* Actual hash code distributions encountered in practice
362	>	* sometimes deviate significantly from uniform randomness.  This
363	>	* includes the case when N > (1<<30), so some keys MUST collide.
364	>	* Similarly for dumb or hostile usages in which multiple keys are
365	>	* designed to have identical hash codes or ones that differs only
366	>	* in masked-out high bits. So we use a secondary strategy that
367	>	* applies when the number of nodes in a bin exceeds a
368	>	* threshold. These TreeBins use a balanced tree to hold nodes (a
369	>	* specialized form of red-black trees), bounding search time to
370	>	* O(log N).  Each search step in a TreeBin is at least twice as
371	>	* slow as in a regular list, but given that N cannot exceed
372	>	* (1<<64) (before running out of addresses) this bounds search
373	>	* steps, lock hold times, etc, to reasonable constants (roughly
374	>	* 100 nodes inspected per operation worst case) so long as keys
375	>	* are Comparable (which is very common -- String, Long, etc).
376	>	* TreeBin nodes (TreeNodes) also maintain the same "next"
377	>	* traversal pointers as regular nodes, so can be traversed in
378	>	* iterators in the same way.
379	>	*
380	>	* The table is resized when occupancy exceeds a percentage
381	>	* threshold (nominally, 0.75, but see below).  Any thread
382	>	* noticing an overfull bin may assist in resizing after the
383	>	* initiating thread allocates and sets up the replacement array.
384	>	* However, rather than stalling, these other threads may proceed
385	>	* with insertions etc.  The use of TreeBins shields us from the
386	>	* worst case effects of overfilling while resizes are in
387	>	* progress.  Resizing proceeds by transferring bins, one by one,
388	>	* from the table to the next table. However, threads claim small
389	>	* blocks of indices to transfer (via field transferIndex) before
390	>	* doing so, reducing contention.  A generation stamp in field
391	>	* sizeCtl ensures that resizings do not overlap. Because we are
392	>	* using power-of-two expansion, the elements from each bin must
393	>	* either stay at same index, or move with a power of two
394	>	* offset. We eliminate unnecessary node creation by catching
395	>	* cases where old nodes can be reused because their next fields
396	>	* won't change.  On average, only about one-sixth of them need
397	>	* cloning when a table doubles. The nodes they replace will be
398	>	* garbage collectable as soon as they are no longer referenced by
399	>	* any reader thread that may be in the midst of concurrently
400	>	* traversing table.  Upon transfer, the old table bin contains
401	>	* only a special forwarding node (with hash field "MOVED") that
402	>	* contains the next table as its key. On encountering a
403	>	* forwarding node, access and update operations restart, using
404	>	* the new table.
405	>	*
406	>	* Each bin transfer requires its bin lock, which can stall
407	>	* waiting for locks while resizing. However, because other
408	>	* threads can join in and help resize rather than contend for
409	>	* locks, average aggregate waits become shorter as resizing
410	>	* progresses.  The transfer operation must also ensure that all
411	>	* accessible bins in both the old and new table are usable by any
412	>	* traversal.  This is arranged in part by proceeding from the
413	>	* last bin (table.length - 1) up towards the first.  Upon seeing
414	>	* a forwarding node, traversals (see class Traverser) arrange to
415	>	* move to the new table without revisiting nodes.  To ensure that
416	>	* no intervening nodes are skipped even when moved out of order,
417	>	* a stack (see class TableStack) is created on first encounter of
418	>	* a forwarding node during a traversal, to maintain its place if
419	>	* later processing the current table. The need for these
420	>	* save/restore mechanics is relatively rare, but when one
421	>	* forwarding node is encountered, typically many more will be.
422	>	* So Traversers use a simple caching scheme to avoid creating so
423	>	* many new TableStack nodes. (Thanks to Peter Levart for
424	>	* suggesting use of a stack here.)
425	>	*
426	>	* The traversal scheme also applies to partial traversals of
427	>	* ranges of bins (via an alternate Traverser constructor)
428	>	* to support partitioned aggregate operations.  Also, read-only
429	>	* operations give up if ever forwarded to a null table, which
430	>	* provides support for shutdown-style clearing, which is also not
431	>	* currently implemented.
432	>	*
433	>	* Lazy table initialization minimizes footprint until first use,
434	>	* and also avoids resizings when the first operation is from a
435	>	* putAll, constructor with map argument, or deserialization.
436	>	* These cases attempt to override the initial capacity settings,
437	>	* but harmlessly fail to take effect in cases of races.
438	>	*
439	>	* The element count is maintained using a specialization of
440	>	* LongAdder. We need to incorporate a specialization rather than
441	>	* just use a LongAdder in order to access implicit
442	>	* contention-sensing that leads to creation of multiple
443	>	* CounterCells.  The counter mechanics avoid contention on
444	>	* updates but can encounter cache thrashing if read too
445	>	* frequently during concurrent access. To avoid reading so often,
446	>	* resizing under contention is attempted only upon adding to a
447	>	* bin already holding two or more nodes. Under uniform hash
448	>	* distributions, the probability of this occurring at threshold
449	>	* is around 13%, meaning that only about 1 in 8 puts check
450	>	* threshold (and after resizing, many fewer do so).
451	>	*
452	>	* TreeBins use a special form of comparison for search and
453	>	* related operations (which is the main reason we cannot use
454	>	* existing collections such as TreeMaps). TreeBins contain
455	>	* Comparable elements, but may contain others, as well as
456	>	* elements that are Comparable but not necessarily Comparable for
457	>	* the same T, so we cannot invoke compareTo among them. To handle
458	>	* this, the tree is ordered primarily by hash value, then by
459	>	* Comparable.compareTo order if applicable.  On lookup at a node,
460	>	* if elements are not comparable or compare as 0 then both left
461	>	* and right children may need to be searched in the case of tied
462	>	* hash values. (This corresponds to the full list search that
463	>	* would be necessary if all elements were non-Comparable and had
464	>	* tied hashes.) On insertion, to keep a total ordering (or as
465	>	* close as is required here) across rebalancings, we compare
466	>	* classes and identityHashCodes as tie-breakers. The red-black
467	>	* balancing code is updated from pre-jdk-collections
468	>	* (http://gee.cs.oswego.edu/dl/classes/collections/RBCell.java)
469	>	* based in turn on Cormen, Leiserson, and Rivest "Introduction to
470	>	* Algorithms" (CLR).
471	>	*
472	>	* TreeBins also require an additional locking mechanism.  While
473	>	* list traversal is always possible by readers even during
474	>	* updates, tree traversal is not, mainly because of tree-rotations
475	>	* that may change the root node and/or its linkages.  TreeBins
476	>	* include a simple read-write lock mechanism parasitic on the
477	>	* main bin-synchronization strategy: Structural adjustments
478	>	* associated with an insertion or removal are already bin-locked
479	>	* (and so cannot conflict with other writers) but must wait for
480	>	* ongoing readers to finish. Since there can be only one such
481	>	* waiter, we use a simple scheme using a single "waiter" field to
482	>	* block writers.  However, readers need never block.  If the root
483	>	* lock is held, they proceed along the slow traversal path (via
484	>	* next-pointers) until the lock becomes available or the list is
485	>	* exhausted, whichever comes first. These cases are not fast, but
486	>	* maximize aggregate expected throughput.
487	>	*
488	>	* Maintaining API and serialization compatibility with previous
489	>	* versions of this class introduces several oddities. Mainly: We
490	>	* leave untouched but unused constructor arguments referring to
491	>	* concurrencyLevel. We accept a loadFactor constructor argument,
492	>	* but apply it only to initial table capacity (which is the only
493	>	* time that we can guarantee to honor it.) We also declare an
494	>	* unused "Segment" class that is instantiated in minimal form
495	>	* only when serializing.
496	>	*
497	>	* Also, solely for compatibility with previous versions of this
498	>	* class, it extends AbstractMap, even though all of its methods
499	>	* are overridden, so it is just useless baggage.
500	>	*
501	>	* This file is organized to make things a little easier to follow
502	>	* while reading than they might otherwise: First the main static
503	>	* declarations and utilities, then fields, then main public
504	>	* methods (with a few factorings of multiple public methods into
505	>	* internal ones), then sizing methods, trees, traversers, and
506	>	* bulk operations.
507		*/
508
509		/* ---------------- Constants -------------- */
510
511		/**
512	<	* The smallest allowed table capacity.  Must be a power of 2, at
513	<	* least 2.
512	>	* The largest possible table capacity.  This value must be
513	>	* exactly 1<<30 to stay within Java array allocation and indexing
514	>	* bounds for power of two table sizes, and is further required
515	>	* because the top two bits of 32bit hash fields are used for
516	>	* control purposes.
517		*/
518	<	static final int MINIMUM_CAPACITY = 2;
518	>	private static final int MAXIMUM_CAPACITY = 1 << 30;
519
520		/**
521	<	* The largest allowed table capacity.  Must be a power of 2, at
522	<	* most 1<<30.
521	>	* The default initial table capacity.  Must be a power of 2
522	>	* (i.e., at least 1) and at most MAXIMUM_CAPACITY.
523		*/
524	<	static final int MAXIMUM_CAPACITY = 1 << 30;
524	>	private static final int DEFAULT_CAPACITY = 16;
525
526		/**
527	<	* The default initial table capacity.  Must be a power of 2, at
528	<	* least MINIMUM_CAPACITY and at most MAXIMUM_CAPACITY.
527	>	* The largest possible (non-power of two) array size.
528	>	* Needed by toArray and related methods.
529		*/
530	<	static final int DEFAULT_CAPACITY = 16;
530	>	static final int MAX_ARRAY_SIZE = Integer.MAX_VALUE - 8;
531
532		/**
533	<	* The default load factor for this table, used when not otherwise
534	<	* specified in a constructor.
533	>	* The default concurrency level for this table. Unused but
534	>	* defined for compatibility with previous versions of this class.
535		*/
536	<	static final float DEFAULT_LOAD_FACTOR = 0.75f;
536	>	private static final int DEFAULT_CONCURRENCY_LEVEL = 16;
537
538		/**
539	<	* The default concurrency level for this table. Unused, but
540	<	* defined for compatibility with previous versions of this class.
539	>	* The load factor for this table. Overrides of this value in
540	>	* constructors affect only the initial table capacity.  The
541	>	* actual floating point value isn't normally used -- it is
542	>	* simpler to use expressions such as {@code n - (n >>> 2)} for
543	>	* the associated resizing threshold.
544		*/
545	<	static final int DEFAULT_CONCURRENCY_LEVEL = 16;
545	>	private static final float LOAD_FACTOR = 0.75f;
546
547		/**
548	<	* The count value to offset thresholds to compensate for checking
549	<	* for resizing only when inserting into bins with two or more
550	<	* elements. See above for explanation.
548	>	* The bin count threshold for using a tree rather than list for a
549	>	* bin.  Bins are converted to trees when adding an element to a
550	>	* bin with at least this many nodes. The value must be greater
551	>	* than 2, and should be at least 8 to mesh with assumptions in
552	>	* tree removal about conversion back to plain bins upon
553	>	* shrinkage.
554		*/
555	<	static final int THRESHOLD_OFFSET = 8;
555	>	static final int TREEIFY_THRESHOLD = 8;
556
557		/**
558	<	* Special node hash value indicating to use table in node.key
559	<	* Must be negative.
558	>	* The bin count threshold for untreeifying a (split) bin during a
559	>	* resize operation. Should be less than TREEIFY_THRESHOLD, and at
560	>	* most 6 to mesh with shrinkage detection under removal.
561		*/
562	<	static final int MOVED = -1;
248	<
249	<	/* ---------------- Fields -------------- */
562	>	static final int UNTREEIFY_THRESHOLD = 6;
563
564		/**
565	<	* The array of bins. Lazily initialized upon first insertion.
566	<	* Size is always a power of two. Accessed directly by inner
567	<	* classes.
565	>	* The smallest table capacity for which bins may be treeified.
566	>	* (Otherwise the table is resized if too many nodes in a bin.)
567	>	* The value should be at least 4 * TREEIFY_THRESHOLD to avoid
568	>	* conflicts between resizing and treeification thresholds.
569		*/
570	<	transient volatile Node[] table;
570	>	static final int MIN_TREEIFY_CAPACITY = 64;
571
572	<	/** The counter maintaining number of elements. */
573	<	private transient final LongAdder counter;
574	<	/** Nonzero when table is being initialized or resized. Updated via CAS. */
575	<	private transient volatile int resizing;
576	<	/** The target load factor for the table. */
577	<	private transient float loadFactor;
578	<	/** The next element count value upon which to resize the table. */
579	<	private transient int threshold;
266	<	/** The initial capacity of the table. */
267	<	private transient int initCap;
268	<
269	<	// views
270	<	transient Set<K> keySet;
271	<	transient Set<Map.Entry<K,V>> entrySet;
272	<	transient Collection<V> values;
572	>	/**
573	>	* Minimum number of rebinnings per transfer step. Ranges are
574	>	* subdivided to allow multiple resizer threads.  This value
575	>	* serves as a lower bound to avoid resizers encountering
576	>	* excessive memory contention.  The value should be at least
577	>	* DEFAULT_CAPACITY.
578	>	*/
579	>	private static final int MIN_TRANSFER_STRIDE = 16;
580
581	<	/** For serialization compatibility. Null unless serialized; see below */
582	<	Segment<K,V>[] segments;
581	>	/**
582	>	* The number of bits used for generation stamp in sizeCtl.
583	>	* Must be at least 6 for 32bit arrays.
584	>	*/
585	>	private static int RESIZE_STAMP_BITS = 16;
586
587		/**
588	<	* Applies a supplemental hash function to a given hashCode, which
589	<	* defends against poor quality hash functions.  The result must
280	<	* be non-negative, and for reasonable performance must have good
281	<	* avalanche properties; i.e., that each bit of the argument
282	<	* affects each bit (except sign bit) of the result.
588	>	* The maximum number of threads that can help resize.
589	>	* Must fit in 32 - RESIZE_STAMP_BITS bits.
590		*/
591	<	private static final int spread(int h) {
285	<	// Apply base step of MurmurHash; see http://code.google.com/p/smhasher/
286	<	h ^= h >>> 16;
287	<	h *= 0x85ebca6b;
288	<	h ^= h >>> 13;
289	<	h *= 0xc2b2ae35;
290	<	return (h >>> 16) ^ (h & 0x7fffffff); // mask out sign bit
291	<	}
591	>	private static final int MAX_RESIZERS = (1 << (32 - RESIZE_STAMP_BITS)) - 1;
592
593		/**
594	<	* Key-value entry. Note that this is never exported out as a
295	<	* user-visible Map.Entry.
594	>	* The bit shift for recording size stamp in sizeCtl.
595		*/
596	<	static final class Node {
596	>	private static final int RESIZE_STAMP_SHIFT = 32 - RESIZE_STAMP_BITS;
597	>
598	>	/*
599	>	* Encodings for Node hash fields. See above for explanation.
600	>	*/
601	>	static final int MOVED     = -1; // hash for forwarding nodes
602	>	static final int TREEBIN   = -2; // hash for roots of trees
603	>	static final int RESERVED  = -3; // hash for transient reservations
604	>	static final int HASH_BITS = 0x7fffffff; // usable bits of normal node hash
605	>
606	>	/** Number of CPUS, to place bounds on some sizings */
607	>	static final int NCPU = Runtime.getRuntime().availableProcessors();
608	>
609	>	/** For serialization compatibility. */
610	>	private static final ObjectStreamField[] serialPersistentFields = {
611	>	new ObjectStreamField("segments", Segment[].class),
612	>	new ObjectStreamField("segmentMask", Integer.TYPE),
613	>	new ObjectStreamField("segmentShift", Integer.TYPE)
614	>	};
615	>
616	>	/* ---------------- Nodes -------------- */
617	>
618	>	/**
619	>	* Key-value entry.  This class is never exported out as a
620	>	* user-mutable Map.Entry (i.e., one supporting setValue; see
621	>	* MapEntry below), but can be used for read-only traversals used
622	>	* in bulk tasks.  Subclasses of Node with a negative hash field
623	>	* are special, and contain null keys and values (but are never
624	>	* exported).  Otherwise, keys and vals are never null.
625	>	*/
626	>	static class Node<K,V> implements Map.Entry<K,V> {
627		final int hash;
628	<	final Object key;
629	<	volatile Object val;
630	<	volatile Node next;
628	>	final K key;
629	>	volatile V val;
630	>	volatile Node<K,V> next;
631
632	<	Node(int hash, Object key, Object val, Node next) {
632	>	Node(int hash, K key, V val, Node<K,V> next) {
633		this.hash = hash;
634		this.key = key;
635		this.val = val;
636		this.next = next;
637		}
638	+
639	+	public final K getKey()     { return key; }
640	+	public final V getValue()   { return val; }
641	+	public final int hashCode() { return key.hashCode() ^ val.hashCode(); }
642	+	public final String toString() { return key + "=" + val; }
643	+	public final V setValue(V value) {
644	+	throw new UnsupportedOperationException();
645	+	}
646	+
647	+	public final boolean equals(Object o) {
648	+	Object k, v, u; Map.Entry<?,?> e;
649	+	return ((o instanceof Map.Entry) &&
650	+	(k = (e = (Map.Entry<?,?>)o).getKey()) != null &&
651	+	(v = e.getValue()) != null &&
652	+	(k == key || k.equals(key)) &&
653	+	(v == (u = val) || v.equals(u)));
654	+	}
655	+
656	+	/**
657	+	* Virtualized support for map.get(); overridden in subclasses.
658	+	*/
659	+	Node<K,V> find(int h, Object k) {
660	+	Node<K,V> e = this;
661	+	if (k != null) {
662	+	do {
663	+	K ek;
664	+	if (e.hash == h &&
665	+	((ek = e.key) == k || (ek != null && k.equals(ek))))
666	+	return e;
667	+	} while ((e = e.next) != null);
668	+	}
669	+	return null;
670	+	}
671	+	}
672	+
673	+	/* ---------------- Static utilities -------------- */
674	+
675	+	/**
676	+	* Spreads (XORs) higher bits of hash to lower and also forces top
677	+	* bit to 0. Because the table uses power-of-two masking, sets of
678	+	* hashes that vary only in bits above the current mask will
679	+	* always collide. (Among known examples are sets of Float keys
680	+	* holding consecutive whole numbers in small tables.)  So we
681	+	* apply a transform that spreads the impact of higher bits
682	+	* downward. There is a tradeoff between speed, utility, and
683	+	* quality of bit-spreading. Because many common sets of hashes
684	+	* are already reasonably distributed (so don't benefit from
685	+	* spreading), and because we use trees to handle large sets of
686	+	* collisions in bins, we just XOR some shifted bits in the
687	+	* cheapest possible way to reduce systematic lossage, as well as
688	+	* to incorporate impact of the highest bits that would otherwise
689	+	* never be used in index calculations because of table bounds.
690	+	*/
691	+	static final int spread(int h) {
692	+	return (h ^ (h >>> 16)) & HASH_BITS;
693	+	}
694	+
695	+	/**
696	+	* Returns a power of two table size for the given desired capacity.
697	+	* See Hackers Delight, sec 3.2
698	+	*/
699	+	private static final int tableSizeFor(int c) {
700	+	int n = c - 1;
701	+	n |= n >>> 1;
702	+	n |= n >>> 2;
703	+	n |= n >>> 4;
704	+	n |= n >>> 8;
705	+	n |= n >>> 16;
706	+	return (n < 0) ? 1 : (n >= MAXIMUM_CAPACITY) ? MAXIMUM_CAPACITY : n + 1;
707	+	}
708	+
709	+	/**
710	+	* Returns x's Class if it is of the form "class C implements
711	+	* Comparable<C>", else null.
712	+	*/
713	+	static Class<?> comparableClassFor(Object x) {
714	+	if (x instanceof Comparable) {
715	+	Class<?> c; Type[] ts, as; Type t; ParameterizedType p;
716	+	if ((c = x.getClass()) == String.class) // bypass checks
717	+	return c;
718	+	if ((ts = c.getGenericInterfaces()) != null) {
719	+	for (int i = 0; i < ts.length; ++i) {
720	+	if (((t = ts[i]) instanceof ParameterizedType) &&
721	+	((p = (ParameterizedType)t).getRawType() ==
722	+	Comparable.class) &&
723	+	(as = p.getActualTypeArguments()) != null &&
724	+	as.length == 1 && as[0] == c) // type arg is c
725	+	return c;
726	+	}
727	+	}
728	+	}
729	+	return null;
730		}
731
732	+	/**
733	+	* Returns k.compareTo(x) if x matches kc (k's screened comparable
734	+	* class), else 0.
735	+	*/
736	+	@SuppressWarnings({"rawtypes","unchecked"}) // for cast to Comparable
737	+	static int compareComparables(Class<?> kc, Object k, Object x) {
738	+	return (x == null || x.getClass() != kc ? 0 :
739	+	((Comparable)k).compareTo(x));
740	+	}
741	+
742	+	/* ---------------- Table element access -------------- */
743	+
744		/*
745		* Volatile access methods are used for table elements as well as
746	<	* elements of in-progress next table while resizing.  Uses in
747	<	* access and update methods are null checked by callers, and
748	<	* implicitly bounds-checked, relying on the invariants that tab
749	<	* arrays have non-zero size, and all indices are masked with
750	<	* (tab.length - 1) which is never negative and always less than
751	<	* length. The "relaxed" non-volatile forms are used only during
752	<	* table initialization. The only other usage is in
753	<	* HashIterator.advance, which performs explicit checks.
746	>	* elements of in-progress next table while resizing.  All uses of
747	>	* the tab arguments must be null checked by callers.  All callers
748	>	* also paranoically precheck that tab's length is not zero (or an
749	>	* equivalent check), thus ensuring that any index argument taking
750	>	* the form of a hash value anded with (length - 1) is a valid
751	>	* index.  Note that, to be correct wrt arbitrary concurrency
752	>	* errors by users, these checks must operate on local variables,
753	>	* which accounts for some odd-looking inline assignments below.
754	>	* Note that calls to setTabAt always occur within locked regions,
755	>	* and so in principle require only release ordering, not
756	>	* full volatile semantics, but are currently coded as volatile
757	>	* writes to be conservative.
758	>	*/
759	>
760	>	@SuppressWarnings("unchecked")
761	>	static final <K,V> Node<K,V> tabAt(Node<K,V>[] tab, int i) {
762	>	return (Node<K,V>)U.getObjectVolatile(tab, ((long)i << ASHIFT) + ABASE);
763	>	}
764	>
765	>	static final <K,V> boolean casTabAt(Node<K,V>[] tab, int i,
766	>	Node<K,V> c, Node<K,V> v) {
767	>	return U.compareAndSwapObject(tab, ((long)i << ASHIFT) + ABASE, c, v);
768	>	}
769	>
770	>	static final <K,V> void setTabAt(Node<K,V>[] tab, int i, Node<K,V> v) {
771	>	U.putObjectVolatile(tab, ((long)i << ASHIFT) + ABASE, v);
772	>	}
773	>
774	>	/* ---------------- Fields -------------- */
775	>
776	>	/**
777	>	* The array of bins. Lazily initialized upon first insertion.
778	>	* Size is always a power of two. Accessed directly by iterators.
779	>	*/
780	>	transient volatile Node<K,V>[] table;
781	>
782	>	/**
783	>	* The next table to use; non-null only while resizing.
784	>	*/
785	>	private transient volatile Node<K,V>[] nextTable;
786	>
787	>	/**
788	>	* Base counter value, used mainly when there is no contention,
789	>	* but also as a fallback during table initialization
790	>	* races. Updated via CAS.
791	>	*/
792	>	private transient volatile long baseCount;
793	>
794	>	/**
795	>	* Table initialization and resizing control.  When negative, the
796	>	* table is being initialized or resized: -1 for initialization,
797	>	* else -(1 + the number of active resizing threads).  Otherwise,
798	>	* when table is null, holds the initial table size to use upon
799	>	* creation, or 0 for default. After initialization, holds the
800	>	* next element count value upon which to resize the table.
801		*/
802	+	private transient volatile int sizeCtl;
803
804	<	static final Node tabAt(Node[] tab, int i) { // used in HashIterator
805	<	return (Node)UNSAFE.getObjectVolatile(tab, ((long)i<<ASHIFT)+ABASE);
804	>	/**
805	>	* The next table index (plus one) to split while resizing.
806	>	*/
807	>	private transient volatile int transferIndex;
808	>
809	>	/**
810	>	* Spinlock (locked via CAS) used when resizing and/or creating CounterCells.
811	>	*/
812	>	private transient volatile int cellsBusy;
813	>
814	>	/**
815	>	* Table of counter cells. When non-null, size is a power of 2.
816	>	*/
817	>	private transient volatile CounterCell[] counterCells;
818	>
819	>	// views
820	>	private transient KeySetView<K,V> keySet;
821	>	private transient ValuesView<K,V> values;
822	>	private transient EntrySetView<K,V> entrySet;
823	>
824	>
825	>	/* ---------------- Public operations -------------- */
826	>
827	>	/**
828	>	* Creates a new, empty map with the default initial table size (16).
829	>	*/
830	>	public ConcurrentHashMapV8() {
831	>	}
832	>
833	>	/**
834	>	* Creates a new, empty map with an initial table size
835	>	* accommodating the specified number of elements without the need
836	>	* to dynamically resize.
837	>	*
838	>	* @param initialCapacity The implementation performs internal
839	>	* sizing to accommodate this many elements.
840	>	* @throws IllegalArgumentException if the initial capacity of
841	>	* elements is negative
842	>	*/
843	>	public ConcurrentHashMapV8(int initialCapacity) {
844	>	if (initialCapacity < 0)
845	>	throw new IllegalArgumentException();
846	>	int cap = ((initialCapacity >= (MAXIMUM_CAPACITY >>> 1)) ?
847	>	MAXIMUM_CAPACITY :
848	>	tableSizeFor(initialCapacity + (initialCapacity >>> 1) + 1));
849	>	this.sizeCtl = cap;
850		}
851
852	<	private static final boolean casTabAt(Node[] tab, int i, Node c, Node v) {
853	<	return UNSAFE.compareAndSwapObject(tab, ((long)i<<ASHIFT)+ABASE, c, v);
852	>	/**
853	>	* Creates a new map with the same mappings as the given map.
854	>	*
855	>	* @param m the map
856	>	*/
857	>	public ConcurrentHashMapV8(Map<? extends K, ? extends V> m) {
858	>	this.sizeCtl = DEFAULT_CAPACITY;
859	>	putAll(m);
860		}
861
862	<	private static final void setTabAt(Node[] tab, int i, Node v) {
863	<	UNSAFE.putObjectVolatile(tab, ((long)i<<ASHIFT)+ABASE, v);
862	>	/**
863	>	* Creates a new, empty map with an initial table size based on
864	>	* the given number of elements ({@code initialCapacity}) and
865	>	* initial table density ({@code loadFactor}).
866	>	*
867	>	* @param initialCapacity the initial capacity. The implementation
868	>	* performs internal sizing to accommodate this many elements,
869	>	* given the specified load factor.
870	>	* @param loadFactor the load factor (table density) for
871	>	* establishing the initial table size
872	>	* @throws IllegalArgumentException if the initial capacity of
873	>	* elements is negative or the load factor is nonpositive
874	>	*
875	>	* @since 1.6
876	>	*/
877	>	public ConcurrentHashMapV8(int initialCapacity, float loadFactor) {
878	>	this(initialCapacity, loadFactor, 1);
879		}
880
881	<	private static final Node relaxedTabAt(Node[] tab, int i) {
882	<	return (Node)UNSAFE.getObject(tab, ((long)i<<ASHIFT)+ABASE);
881	>	/**
882	>	* Creates a new, empty map with an initial table size based on
883	>	* the given number of elements ({@code initialCapacity}), table
884	>	* density ({@code loadFactor}), and number of concurrently
885	>	* updating threads ({@code concurrencyLevel}).
886	>	*
887	>	* @param initialCapacity the initial capacity. The implementation
888	>	* performs internal sizing to accommodate this many elements,
889	>	* given the specified load factor.
890	>	* @param loadFactor the load factor (table density) for
891	>	* establishing the initial table size
892	>	* @param concurrencyLevel the estimated number of concurrently
893	>	* updating threads. The implementation may use this value as
894	>	* a sizing hint.
895	>	* @throws IllegalArgumentException if the initial capacity is
896	>	* negative or the load factor or concurrencyLevel are
897	>	* nonpositive
898	>	*/
899	>	public ConcurrentHashMapV8(int initialCapacity,
900	>	float loadFactor, int concurrencyLevel) {
901	>	if (!(loadFactor > 0.0f) || initialCapacity < 0 || concurrencyLevel <= 0)
902	>	throw new IllegalArgumentException();
903	>	if (initialCapacity < concurrencyLevel)   // Use at least as many bins
904	>	initialCapacity = concurrencyLevel;   // as estimated threads
905	>	long size = (long)(1.0 + (long)initialCapacity / loadFactor);
906	>	int cap = (size >= (long)MAXIMUM_CAPACITY) ?
907	>	MAXIMUM_CAPACITY : tableSizeFor((int)size);
908	>	this.sizeCtl = cap;
909		}
910
911	<	private static final void relaxedSetTabAt(Node[] tab, int i, Node v) {
912	<	UNSAFE.putObject(tab, ((long)i<<ASHIFT)+ABASE, v);
911	>	// Original (since JDK1.2) Map methods
912	>
913	>	/**
914	>	* {@inheritDoc}
915	>	*/
916	>	public int size() {
917	>	long n = sumCount();
918	>	return ((n < 0L) ? 0 :
919	>	(n > (long)Integer.MAX_VALUE) ? Integer.MAX_VALUE :
920	>	(int)n);
921		}
922
923	<	/* ---------------- Access and update operations -------------- */
923	>	/**
924	>	* {@inheritDoc}
925	>	*/
926	>	public boolean isEmpty() {
927	>	return sumCount() <= 0L; // ignore transient negative values
928	>	}
929
930	<	/** Implementation for get and containsKey */
931	<	private final Object internalGet(Object k) {
932	<	int h = spread(k.hashCode());
933	<	Node[] tab = table;
934	<	retry: while (tab != null) {
935	<	Node e = tabAt(tab, (tab.length - 1) & h);
936	<	while (e != null) {
937	<	int eh = e.hash;
938	<	if (eh == h) {
939	<	Object ek = e.key, ev = e.val;
940	<	if (ev != null && ek != null && (k == ek || k.equals(ek)))
941	<	return ev;
942	<	}
943	<	else if (eh < 0) { // bin was moved during resize
944	<	tab = (Node[])e.key;
945	<	continue retry;
946	<	}
947	<	e = e.next;
930	>	/**
931	>	* Returns the value to which the specified key is mapped,
932	>	* or {@code null} if this map contains no mapping for the key.
933	>	*
934	>	* <p>More formally, if this map contains a mapping from a key
935	>	* {@code k} to a value {@code v} such that {@code key.equals(k)},
936	>	* then this method returns {@code v}; otherwise it returns
937	>	* {@code null}.  (There can be at most one such mapping.)
938	>	*
939	>	* @throws NullPointerException if the specified key is null
940	>	*/
941	>	public V get(Object key) {
942	>	Node<K,V>[] tab; Node<K,V> e, p; int n, eh; K ek;
943	>	int h = spread(key.hashCode());
944	>	if ((tab = table) != null && (n = tab.length) > 0 &&
945	>	(e = tabAt(tab, (n - 1) & h)) != null) {
946	>	if ((eh = e.hash) == h) {
947	>	if ((ek = e.key) == key || (ek != null && key.equals(ek)))
948	>	return e.val;
949	>	}
950	>	else if (eh < 0)
951	>	return (p = e.find(h, key)) != null ? p.val : null;
952	>	while ((e = e.next) != null) {
953	>	if (e.hash == h &&
954	>	((ek = e.key) == key || (ek != null && key.equals(ek))))
955	>	return e.val;
956		}
364	–	break;
957		}
958		return null;
959		}
960
961	+	/**
962	+	* Tests if the specified object is a key in this table.
963	+	*
964	+	* @param  key possible key
965	+	* @return {@code true} if and only if the specified object
966	+	*         is a key in this table, as determined by the
967	+	*         {@code equals} method; {@code false} otherwise
968	+	* @throws NullPointerException if the specified key is null
969	+	*/
970	+	public boolean containsKey(Object key) {
971	+	return get(key) != null;
972	+	}
973	+
974	+	/**
975	+	* Returns {@code true} if this map maps one or more keys to the
976	+	* specified value. Note: This method may require a full traversal
977	+	* of the map, and is much slower than method {@code containsKey}.
978	+	*
979	+	* @param value value whose presence in this map is to be tested
980	+	* @return {@code true} if this map maps one or more keys to the
981	+	*         specified value
982	+	* @throws NullPointerException if the specified value is null
983	+	*/
984	+	public boolean containsValue(Object value) {
985	+	if (value == null)
986	+	throw new NullPointerException();
987	+	Node<K,V>[] t;
988	+	if ((t = table) != null) {
989	+	Traverser<K,V> it = new Traverser<K,V>(t, t.length, 0, t.length);
990	+	for (Node<K,V> p; (p = it.advance()) != null; ) {
991	+	V v;
992	+	if ((v = p.val) == value || (v != null && value.equals(v)))
993	+	return true;
994	+	}
995	+	}
996	+	return false;
997	+	}
998	+
999	+	/**
1000	+	* Maps the specified key to the specified value in this table.
1001	+	* Neither the key nor the value can be null.
1002	+	*
1003	+	* <p>The value can be retrieved by calling the {@code get} method
1004	+	* with a key that is equal to the original key.
1005	+	*
1006	+	* @param key key with which the specified value is to be associated
1007	+	* @param value value to be associated with the specified key
1008	+	* @return the previous value associated with {@code key}, or
1009	+	*         {@code null} if there was no mapping for {@code key}
1010	+	* @throws NullPointerException if the specified key or value is null
1011	+	*/
1012	+	public V put(K key, V value) {
1013	+	return putVal(key, value, false);
1014	+	}
1015
1016		/** Implementation for put and putIfAbsent */
1017	<	private final Object internalPut(Object k, Object v, boolean replace) {
1018	<	int h = spread(k.hashCode());
1019	<	Object oldVal = null;  // the previous value or null if none
1020	<	Node[] tab = table;
1021	<	for (;;) {
1022	<	Node e; int i;
1023	<	if (tab == null)
1024	<	tab = grow(0);
1025	<	else if ((e = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1026	<	if (casTabAt(tab, i, null, new Node(h, k, v, null)))
1027	<	break;
1017	>	final V putVal(K key, V value, boolean onlyIfAbsent) {
1018	>	if (key == null || value == null) throw new NullPointerException();
1019	>	int hash = spread(key.hashCode());
1020	>	int binCount = 0;
1021	>	for (Node<K,V>[] tab = table;;) {
1022	>	Node<K,V> f; int n, i, fh;
1023	>	if (tab == null || (n = tab.length) == 0)
1024	>	tab = initTable();
1025	>	else if ((f = tabAt(tab, i = (n - 1) & hash)) == null) {
1026	>	if (casTabAt(tab, i, null,
1027	>	new Node<K,V>(hash, key, value, null)))
1028	>	break;                   // no lock when adding to empty bin
1029		}
1030	<	else if (e.hash < 0)
1031	<	tab = (Node[])e.key;
1030	>	else if ((fh = f.hash) == MOVED)
1031	>	tab = helpTransfer(tab, f);
1032		else {
1033	<	boolean validated = false;
1034	<	boolean checkSize = false;
1035	<	synchronized (e) {
1036	<	if (tabAt(tab, i) == e) {
1037	<	validated = true;
1038	<	for (Node first = e;;) {
1039	<	Object ek, ev;
1040	<	if (e.hash == h &&
1041	<	(ek = e.key) != null &&
1042	<	(ev = e.val) != null &&
1043	<	(k == ek || k.equals(ek))) {
1044	<	oldVal = ev;
1045	<	if (replace)
1046	<	e.val = v;
1047	<	break;
1033	>	V oldVal = null;
1034	>	synchronized (f) {
1035	>	if (tabAt(tab, i) == f) {
1036	>	if (fh >= 0) {
1037	>	binCount = 1;
1038	>	for (Node<K,V> e = f;; ++binCount) {
1039	>	K ek;
1040	>	if (e.hash == hash &&
1041	>	((ek = e.key) == key ||
1042	>	(ek != null && key.equals(ek)))) {
1043	>	oldVal = e.val;
1044	>	if (!onlyIfAbsent)
1045	>	e.val = value;
1046	>	break;
1047	>	}
1048	>	Node<K,V> pred = e;
1049	>	if ((e = e.next) == null) {
1050	>	pred.next = new Node<K,V>(hash, key,
1051	>	value, null);
1052	>	break;
1053	>	}
1054		}
1055	<	Node last = e;
1056	<	if ((e = e.next) == null) {
1057	<	last.next = new Node(h, k, v, null);
1058	<	if (last != first || tab.length <= 64)
1059	<	checkSize = true;
1060	<	break;
1055	>	}
1056	>	else if (f instanceof TreeBin) {
1057	>	Node<K,V> p;
1058	>	binCount = 2;
1059	>	if ((p = ((TreeBin<K,V>)f).putTreeVal(hash, key,
1060	>	value)) != null) {
1061	>	oldVal = p.val;
1062	>	if (!onlyIfAbsent)
1063	>	p.val = value;
1064		}
1065		}
1066		}
1067		}
1068	<	if (validated) {
1069	<	if (checkSize && tab.length < MAXIMUM_CAPACITY &&
1070	<	resizing == 0 && counter.sum() >= threshold)
1071	<	grow(0);
1068	>	if (binCount != 0) {
1069	>	if (binCount >= TREEIFY_THRESHOLD)
1070	>	treeifyBin(tab, i);
1071	>	if (oldVal != null)
1072	>	return oldVal;
1073		break;
1074		}
1075		}
1076		}
1077	<	if (oldVal == null)
1078	<	counter.increment();
1079	<	return oldVal;
1077	>	addCount(1L, binCount);
1078	>	return null;
1079	>	}
1080	>
1081	>	/**
1082	>	* Copies all of the mappings from the specified map to this one.
1083	>	* These mappings replace any mappings that this map had for any of the
1084	>	* keys currently in the specified map.
1085	>	*
1086	>	* @param m mappings to be stored in this map
1087	>	*/
1088	>	public void putAll(Map<? extends K, ? extends V> m) {
1089	>	tryPresize(m.size());
1090	>	for (Map.Entry<? extends K, ? extends V> e : m.entrySet())
1091	>	putVal(e.getKey(), e.getValue(), false);
1092	>	}
1093	>
1094	>	/**
1095	>	* Removes the key (and its corresponding value) from this map.
1096	>	* This method does nothing if the key is not in the map.
1097	>	*
1098	>	* @param  key the key that needs to be removed
1099	>	* @return the previous value associated with {@code key}, or
1100	>	*         {@code null} if there was no mapping for {@code key}
1101	>	* @throws NullPointerException if the specified key is null
1102	>	*/
1103	>	public V remove(Object key) {
1104	>	return replaceNode(key, null, null);
1105		}
1106
1107		/**
#	Line 427 | Line 1109 | public class ConcurrentHashMapV8<K, V>
1109		* Replaces node value with v, conditional upon match of cv if
1110		* non-null.  If resulting value is null, delete.
1111		*/
1112	<	private final Object internalReplace(Object k, Object v, Object cv) {
1113	<	int h = spread(k.hashCode());
1114	<	Object oldVal = null;
1115	<	Node e; int i;
1116	<	Node[] tab = table;
1117	<	while (tab != null &&
1118	<	(e = tabAt(tab, i = (tab.length - 1) & h)) != null) {
1119	<	if (e.hash < 0)
1120	<	tab = (Node[])e.key;
1112	>	final V replaceNode(Object key, V value, Object cv) {
1113	>	int hash = spread(key.hashCode());
1114	>	for (Node<K,V>[] tab = table;;) {
1115	>	Node<K,V> f; int n, i, fh;
1116	>	if (tab == null || (n = tab.length) == 0 ||
1117	>	(f = tabAt(tab, i = (n - 1) & hash)) == null)
1118	>	break;
1119	>	else if ((fh = f.hash) == MOVED)
1120	>	tab = helpTransfer(tab, f);
1121		else {
1122	+	V oldVal = null;
1123		boolean validated = false;
1124	<	boolean deleted = false;
1125	<	synchronized (e) {
1126	<	if (tabAt(tab, i) == e) {
1127	<	validated = true;
1128	<	Node pred = null;
1129	<	do {
1130	<	Object ek, ev;
1131	<	if (e.hash == h &&
1132	<	(ek = e.key) != null &&
1133	<	(ev = e.val) != null &&
1134	<	(k == ek || k.equals(ek))) {
1135	<	if (cv == null || cv == ev || cv.equals(ev)) {
1136	<	oldVal = ev;
1137	<	if ((e.val = v) == null) {
1138	<	deleted = true;
1139	<	Node en = e.next;
1140	<	if (pred != null)
458	<	pred.next = en;
1124	>	synchronized (f) {
1125	>	if (tabAt(tab, i) == f) {
1126	>	if (fh >= 0) {
1127	>	validated = true;
1128	>	for (Node<K,V> e = f, pred = null;;) {
1129	>	K ek;
1130	>	if (e.hash == hash &&
1131	>	((ek = e.key) == key ||
1132	>	(ek != null && key.equals(ek)))) {
1133	>	V ev = e.val;
1134	>	if (cv == null || cv == ev ||
1135	>	(ev != null && cv.equals(ev))) {
1136	>	oldVal = ev;
1137	>	if (value != null)
1138	>	e.val = value;
1139	>	else if (pred != null)
1140	>	pred.next = e.next;
1141		else
1142	<	setTabAt(tab, i, en);
1142	>	setTabAt(tab, i, e.next);
1143		}
1144	+	break;
1145		}
1146	<	break;
1146	>	pred = e;
1147	>	if ((e = e.next) == null)
1148	>	break;
1149		}
1150	<	pred = e;
1151	<	} while ((e = e.next) != null);
1150	>	}
1151	>	else if (f instanceof TreeBin) {
1152	>	validated = true;
1153	>	TreeBin<K,V> t = (TreeBin<K,V>)f;
1154	>	TreeNode<K,V> r, p;
1155	>	if ((r = t.root) != null &&
1156	>	(p = r.findTreeNode(hash, key, null)) != null) {
1157	>	V pv = p.val;
1158	>	if (cv == null || cv == pv ||
1159	>	(pv != null && cv.equals(pv))) {
1160	>	oldVal = pv;
1161	>	if (value != null)
1162	>	p.val = value;
1163	>	else if (t.removeTreeNode(p))
1164	>	setTabAt(tab, i, untreeify(t.first));
1165	>	}
1166	>	}
1167	>	}
1168		}
1169		}
1170		if (validated) {
1171	<	if (deleted)
1172	<	counter.decrement();
1171	>	if (oldVal != null) {
1172	>	if (value == null)
1173	>	addCount(-1L, -1);
1174	>	return oldVal;
1175	>	}
1176		break;
1177		}
1178		}
1179		}
1180	<	return oldVal;
1180	>	return null;
1181		}
1182
1183	<	/** Implementation for computeIfAbsent and compute */
1184	<	@SuppressWarnings("unchecked")
1185	<	private final V internalCompute(K k,
1186	<	MappingFunction<? super K, ? extends V> f,
1187	<	boolean replace) {
1188	<	int h = spread(k.hashCode());
1189	<	V val = null;
1190	<	boolean added = false;
1191	<	Node[] tab = table;
1183	>	/**
1184	>	* Removes all of the mappings from this map.
1185	>	*/
1186	>	public void clear() {
1187	>	long delta = 0L; // negative number of deletions
1188	>	int i = 0;
1189	>	Node<K,V>[] tab = table;
1190	>	while (tab != null && i < tab.length) {
1191	>	int fh;
1192	>	Node<K,V> f = tabAt(tab, i);
1193	>	if (f == null)
1194	>	++i;
1195	>	else if ((fh = f.hash) == MOVED) {
1196	>	tab = helpTransfer(tab, f);
1197	>	i = 0; // restart
1198	>	}
1199	>	else {
1200	>	synchronized (f) {
1201	>	if (tabAt(tab, i) == f) {
1202	>	Node<K,V> p = (fh >= 0 ? f :
1203	>	(f instanceof TreeBin) ?
1204	>	((TreeBin<K,V>)f).first : null);
1205	>	while (p != null) {
1206	>	--delta;
1207	>	p = p.next;
1208	>	}
1209	>	setTabAt(tab, i++, null);
1210	>	}
1211	>	}
1212	>	}
1213	>	}
1214	>	if (delta != 0L)
1215	>	addCount(delta, -1);
1216	>	}
1217	>
1218	>	/**
1219	>	* Returns a {@link Set} view of the keys contained in this map.
1220	>	* The set is backed by the map, so changes to the map are
1221	>	* reflected in the set, and vice-versa. The set supports element
1222	>	* removal, which removes the corresponding mapping from this map,
1223	>	* via the {@code Iterator.remove}, {@code Set.remove},
1224	>	* {@code removeAll}, {@code retainAll}, and {@code clear}
1225	>	* operations.  It does not support the {@code add} or
1226	>	* {@code addAll} operations.
1227	>	*
1228	>	* <p>The view's {@code iterator} is a "weakly consistent" iterator
1229	>	* that will never throw {@link ConcurrentModificationException},
1230	>	* and guarantees to traverse elements as they existed upon
1231	>	* construction of the iterator, and may (but is not guaranteed to)
1232	>	* reflect any modifications subsequent to construction.
1233	>	*
1234	>	* @return the set view
1235	>	*/
1236	>	public KeySetView<K,V> keySet() {
1237	>	KeySetView<K,V> ks;
1238	>	return (ks = keySet) != null ? ks : (keySet = new KeySetView<K,V>(this, null));
1239	>	}
1240	>
1241	>	/**
1242	>	* Returns a {@link Collection} view of the values contained in this map.
1243	>	* The collection is backed by the map, so changes to the map are
1244	>	* reflected in the collection, and vice-versa.  The collection
1245	>	* supports element removal, which removes the corresponding
1246	>	* mapping from this map, via the {@code Iterator.remove},
1247	>	* {@code Collection.remove}, {@code removeAll},
1248	>	* {@code retainAll}, and {@code clear} operations.  It does not
1249	>	* support the {@code add} or {@code addAll} operations.
1250	>	*
1251	>	* <p>The view's {@code iterator} is a "weakly consistent" iterator
1252	>	* that will never throw {@link ConcurrentModificationException},
1253	>	* and guarantees to traverse elements as they existed upon
1254	>	* construction of the iterator, and may (but is not guaranteed to)
1255	>	* reflect any modifications subsequent to construction.
1256	>	*
1257	>	* @return the collection view
1258	>	*/
1259	>	public Collection<V> values() {
1260	>	ValuesView<K,V> vs;
1261	>	return (vs = values) != null ? vs : (values = new ValuesView<K,V>(this));
1262	>	}
1263	>
1264	>	/**
1265	>	* Returns a {@link Set} view of the mappings contained in this map.
1266	>	* The set is backed by the map, so changes to the map are
1267	>	* reflected in the set, and vice-versa.  The set supports element
1268	>	* removal, which removes the corresponding mapping from the map,
1269	>	* via the {@code Iterator.remove}, {@code Set.remove},
1270	>	* {@code removeAll}, {@code retainAll}, and {@code clear}
1271	>	* operations.
1272	>	*
1273	>	* <p>The view's {@code iterator} is a "weakly consistent" iterator
1274	>	* that will never throw {@link ConcurrentModificationException},
1275	>	* and guarantees to traverse elements as they existed upon
1276	>	* construction of the iterator, and may (but is not guaranteed to)
1277	>	* reflect any modifications subsequent to construction.
1278	>	*
1279	>	* @return the set view
1280	>	*/
1281	>	public Set<Map.Entry<K,V>> entrySet() {
1282	>	EntrySetView<K,V> es;
1283	>	return (es = entrySet) != null ? es : (entrySet = new EntrySetView<K,V>(this));
1284	>	}
1285	>
1286	>	/**
1287	>	* Returns the hash code value for this {@link Map}, i.e.,
1288	>	* the sum of, for each key-value pair in the map,
1289	>	* {@code key.hashCode() ^ value.hashCode()}.
1290	>	*
1291	>	* @return the hash code value for this map
1292	>	*/
1293	>	public int hashCode() {
1294	>	int h = 0;
1295	>	Node<K,V>[] t;
1296	>	if ((t = table) != null) {
1297	>	Traverser<K,V> it = new Traverser<K,V>(t, t.length, 0, t.length);
1298	>	for (Node<K,V> p; (p = it.advance()) != null; )
1299	>	h += p.key.hashCode() ^ p.val.hashCode();
1300	>	}
1301	>	return h;
1302	>	}
1303	>
1304	>	/**
1305	>	* Returns a string representation of this map.  The string
1306	>	* representation consists of a list of key-value mappings (in no
1307	>	* particular order) enclosed in braces ("{@code {}}").  Adjacent
1308	>	* mappings are separated by the characters {@code ", "} (comma
1309	>	* and space).  Each key-value mapping is rendered as the key
1310	>	* followed by an equals sign ("{@code =}") followed by the
1311	>	* associated value.
1312	>	*
1313	>	* @return a string representation of this map
1314	>	*/
1315	>	public String toString() {
1316	>	Node<K,V>[] t;
1317	>	int f = (t = table) == null ? 0 : t.length;
1318	>	Traverser<K,V> it = new Traverser<K,V>(t, f, 0, f);
1319	>	StringBuilder sb = new StringBuilder();
1320	>	sb.append('{');
1321	>	Node<K,V> p;
1322	>	if ((p = it.advance()) != null) {
1323	>	for (;;) {
1324	>	K k = p.key;
1325	>	V v = p.val;
1326	>	sb.append(k == this ? "(this Map)" : k);
1327	>	sb.append('=');
1328	>	sb.append(v == this ? "(this Map)" : v);
1329	>	if ((p = it.advance()) == null)
1330	>	break;
1331	>	sb.append(',').append(' ');
1332	>	}
1333	>	}
1334	>	return sb.append('}').toString();
1335	>	}
1336	>
1337	>	/**
1338	>	* Compares the specified object with this map for equality.
1339	>	* Returns {@code true} if the given object is a map with the same
1340	>	* mappings as this map.  This operation may return misleading
1341	>	* results if either map is concurrently modified during execution
1342	>	* of this method.
1343	>	*
1344	>	* @param o object to be compared for equality with this map
1345	>	* @return {@code true} if the specified object is equal to this map
1346	>	*/
1347	>	public boolean equals(Object o) {
1348	>	if (o != this) {
1349	>	if (!(o instanceof Map))
1350	>	return false;
1351	>	Map<?,?> m = (Map<?,?>) o;
1352	>	Node<K,V>[] t;
1353	>	int f = (t = table) == null ? 0 : t.length;
1354	>	Traverser<K,V> it = new Traverser<K,V>(t, f, 0, f);
1355	>	for (Node<K,V> p; (p = it.advance()) != null; ) {
1356	>	V val = p.val;
1357	>	Object v = m.get(p.key);
1358	>	if (v == null || (v != val && !v.equals(val)))
1359	>	return false;
1360	>	}
1361	>	for (Map.Entry<?,?> e : m.entrySet()) {
1362	>	Object mk, mv, v;
1363	>	if ((mk = e.getKey()) == null ||
1364	>	(mv = e.getValue()) == null ||
1365	>	(v = get(mk)) == null ||
1366	>	(mv != v && !mv.equals(v)))
1367	>	return false;
1368	>	}
1369	>	}
1370	>	return true;
1371	>	}
1372	>
1373	>	/**
1374	>	* Stripped-down version of helper class used in previous version,
1375	>	* declared for the sake of serialization compatibility
1376	>	*/
1377	>	static class Segment<K,V> extends ReentrantLock implements Serializable {
1378	>	private static final long serialVersionUID = 2249069246763182397L;
1379	>	final float loadFactor;
1380	>	Segment(float lf) { this.loadFactor = lf; }
1381	>	}
1382	>
1383	>	/**
1384	>	* Saves the state of the {@code ConcurrentHashMapV8} instance to a
1385	>	* stream (i.e., serializes it).
1386	>	* @param s the stream
1387	>	* @throws java.io.IOException if an I/O error occurs
1388	>	* @serialData
1389	>	* the key (Object) and value (Object)
1390	>	* for each key-value mapping, followed by a null pair.
1391	>	* The key-value mappings are emitted in no particular order.
1392	>	*/
1393	>	private void writeObject(java.io.ObjectOutputStream s)
1394	>	throws java.io.IOException {
1395	>	// For serialization compatibility
1396	>	// Emulate segment calculation from previous version of this class
1397	>	int sshift = 0;
1398	>	int ssize = 1;
1399	>	while (ssize < DEFAULT_CONCURRENCY_LEVEL) {
1400	>	++sshift;
1401	>	ssize <<= 1;
1402	>	}
1403	>	int segmentShift = 32 - sshift;
1404	>	int segmentMask = ssize - 1;
1405	>	@SuppressWarnings("unchecked") Segment<K,V>[] segments = (Segment<K,V>[])
1406	>	new Segment<?,?>[DEFAULT_CONCURRENCY_LEVEL];
1407	>	for (int i = 0; i < segments.length; ++i)
1408	>	segments[i] = new Segment<K,V>(LOAD_FACTOR);
1409	>	s.putFields().put("segments", segments);
1410	>	s.putFields().put("segmentShift", segmentShift);
1411	>	s.putFields().put("segmentMask", segmentMask);
1412	>	s.writeFields();
1413	>
1414	>	Node<K,V>[] t;
1415	>	if ((t = table) != null) {
1416	>	Traverser<K,V> it = new Traverser<K,V>(t, t.length, 0, t.length);
1417	>	for (Node<K,V> p; (p = it.advance()) != null; ) {
1418	>	s.writeObject(p.key);
1419	>	s.writeObject(p.val);
1420	>	}
1421	>	}
1422	>	s.writeObject(null);
1423	>	s.writeObject(null);
1424	>	segments = null; // throw away
1425	>	}
1426	>
1427	>	/**
1428	>	* Reconstitutes the instance from a stream (that is, deserializes it).
1429	>	* @param s the stream
1430	>	* @throws ClassNotFoundException if the class of a serialized object
1431	>	*         could not be found
1432	>	* @throws java.io.IOException if an I/O error occurs
1433	>	*/
1434	>	private void readObject(java.io.ObjectInputStream s)
1435	>	throws java.io.IOException, ClassNotFoundException {
1436	>	/*
1437	>	* To improve performance in typical cases, we create nodes
1438	>	* while reading, then place in table once size is known.
1439	>	* However, we must also validate uniqueness and deal with
1440	>	* overpopulated bins while doing so, which requires
1441	>	* specialized versions of putVal mechanics.
1442	>	*/
1443	>	sizeCtl = -1; // force exclusion for table construction
1444	>	s.defaultReadObject();
1445	>	long size = 0L;
1446	>	Node<K,V> p = null;
1447		for (;;) {
1448	<	Node e; int i;
1449	<	if (tab == null)
1450	<	tab = grow(0);
1451	<	else if ((e = tabAt(tab, i = (tab.length - 1) & h)) == null) {
1452	<	Node node = new Node(h, k, null, null);
1453	<	boolean validated = false;
1454	<	synchronized (node) {
1455	<	if (casTabAt(tab, i, null, node)) {
1456	<	validated = true;
1448	>	@SuppressWarnings("unchecked") K k = (K) s.readObject();
1449	>	@SuppressWarnings("unchecked") V v = (V) s.readObject();
1450	>	if (k != null && v != null) {
1451	>	p = new Node<K,V>(spread(k.hashCode()), k, v, p);
1452	>	++size;
1453	>	}
1454	>	else
1455	>	break;
1456	>	}
1457	>	if (size == 0L)
1458	>	sizeCtl = 0;
1459	>	else {
1460	>	int n;
1461	>	if (size >= (long)(MAXIMUM_CAPACITY >>> 1))
1462	>	n = MAXIMUM_CAPACITY;
1463	>	else {
1464	>	int sz = (int)size;
1465	>	n = tableSizeFor(sz + (sz >>> 1) + 1);
1466	>	}
1467	>	@SuppressWarnings("unchecked")
1468	>	Node<K,V>[] tab = (Node<K,V>[])new Node<?,?>[n];
1469	>	int mask = n - 1;
1470	>	long added = 0L;
1471	>	while (p != null) {
1472	>	boolean insertAtFront;
1473	>	Node<K,V> next = p.next, first;
1474	>	int h = p.hash, j = h & mask;
1475	>	if ((first = tabAt(tab, j)) == null)
1476	>	insertAtFront = true;
1477	>	else {
1478	>	K k = p.key;
1479	>	if (first.hash < 0) {
1480	>	TreeBin<K,V> t = (TreeBin<K,V>)first;
1481	>	if (t.putTreeVal(h, k, p.val) == null)
1482	>	++added;
1483	>	insertAtFront = false;
1484	>	}
1485	>	else {
1486	>	int binCount = 0;
1487	>	insertAtFront = true;
1488	>	Node<K,V> q; K qk;
1489	>	for (q = first; q != null; q = q.next) {
1490	>	if (q.hash == h &&
1491	>	((qk = q.key) == k ||
1492	>	(qk != null && k.equals(qk)))) {
1493	>	insertAtFront = false;
1494	>	break;
1495	>	}
1496	>	++binCount;
1497	>	}
1498	>	if (insertAtFront && binCount >= TREEIFY_THRESHOLD) {
1499	>	insertAtFront = false;
1500	>	++added;
1501	>	p.next = first;
1502	>	TreeNode<K,V> hd = null, tl = null;
1503	>	for (q = p; q != null; q = q.next) {
1504	>	TreeNode<K,V> t = new TreeNode<K,V>
1505	>	(q.hash, q.key, q.val, null, null);
1506	>	if ((t.prev = tl) == null)
1507	>	hd = t;
1508	>	else
1509	>	tl.next = t;
1510	>	tl = t;
1511	>	}
1512	>	setTabAt(tab, j, new TreeBin<K,V>(hd));
1513	>	}
1514	>	}
1515	>	}
1516	>	if (insertAtFront) {
1517	>	++added;
1518	>	p.next = first;
1519	>	setTabAt(tab, j, p);
1520	>	}
1521	>	p = next;
1522	>	}
1523	>	table = tab;
1524	>	sizeCtl = n - (n >>> 2);
1525	>	baseCount = added;
1526	>	}
1527	>	}
1528	>
1529	>	// ConcurrentMap methods
1530	>
1531	>	/**
1532	>	* {@inheritDoc}
1533	>	*
1534	>	* @return the previous value associated with the specified key,
1535	>	*         or {@code null} if there was no mapping for the key
1536	>	* @throws NullPointerException if the specified key or value is null
1537	>	*/
1538	>	public V putIfAbsent(K key, V value) {
1539	>	return putVal(key, value, true);
1540	>	}
1541	>
1542	>	/**
1543	>	* {@inheritDoc}
1544	>	*
1545	>	* @throws NullPointerException if the specified key is null
1546	>	*/
1547	>	public boolean remove(Object key, Object value) {
1548	>	if (key == null)
1549	>	throw new NullPointerException();
1550	>	return value != null && replaceNode(key, null, value) != null;
1551	>	}
1552	>
1553	>	/**
1554	>	* {@inheritDoc}
1555	>	*
1556	>	* @throws NullPointerException if any of the arguments are null
1557	>	*/
1558	>	public boolean replace(K key, V oldValue, V newValue) {
1559	>	if (key == null || oldValue == null || newValue == null)
1560	>	throw new NullPointerException();
1561	>	return replaceNode(key, newValue, oldValue) != null;
1562	>	}
1563	>
1564	>	/**
1565	>	* {@inheritDoc}
1566	>	*
1567	>	* @return the previous value associated with the specified key,
1568	>	*         or {@code null} if there was no mapping for the key
1569	>	* @throws NullPointerException if the specified key or value is null
1570	>	*/
1571	>	public V replace(K key, V value) {
1572	>	if (key == null || value == null)
1573	>	throw new NullPointerException();
1574	>	return replaceNode(key, value, null);
1575	>	}
1576	>
1577	>	// Overrides of JDK8+ Map extension method defaults
1578	>
1579	>	/**
1580	>	* Returns the value to which the specified key is mapped, or the
1581	>	* given default value if this map contains no mapping for the
1582	>	* key.
1583	>	*
1584	>	* @param key the key whose associated value is to be returned
1585	>	* @param defaultValue the value to return if this map contains
1586	>	* no mapping for the given key
1587	>	* @return the mapping for the key, if present; else the default value
1588	>	* @throws NullPointerException if the specified key is null
1589	>	*/
1590	>	public V getOrDefault(Object key, V defaultValue) {
1591	>	V v;
1592	>	return (v = get(key)) == null ? defaultValue : v;
1593	>	}
1594	>
1595	>	public void forEach(BiAction<? super K, ? super V> action) {
1596	>	if (action == null) throw new NullPointerException();
1597	>	Node<K,V>[] t;
1598	>	if ((t = table) != null) {
1599	>	Traverser<K,V> it = new Traverser<K,V>(t, t.length, 0, t.length);
1600	>	for (Node<K,V> p; (p = it.advance()) != null; ) {
1601	>	action.apply(p.key, p.val);
1602	>	}
1603	>	}
1604	>	}
1605	>
1606	>	public void replaceAll(BiFun<? super K, ? super V, ? extends V> function) {
1607	>	if (function == null) throw new NullPointerException();
1608	>	Node<K,V>[] t;
1609	>	if ((t = table) != null) {
1610	>	Traverser<K,V> it = new Traverser<K,V>(t, t.length, 0, t.length);
1611	>	for (Node<K,V> p; (p = it.advance()) != null; ) {
1612	>	V oldValue = p.val;
1613	>	for (K key = p.key;;) {
1614	>	V newValue = function.apply(key, oldValue);
1615	>	if (newValue == null)
1616	>	throw new NullPointerException();
1617	>	if (replaceNode(key, newValue, oldValue) != null ||
1618	>	(oldValue = get(key)) == null)
1619	>	break;
1620	>	}
1621	>	}
1622	>	}
1623	>	}
1624	>
1625	>	/**
1626	>	* If the specified key is not already associated with a value,
1627	>	* attempts to compute its value using the given mapping function
1628	>	* and enters it into this map unless {@code null}.  The entire
1629	>	* method invocation is performed atomically, so the function is
1630	>	* applied at most once per key.  Some attempted update operations
1631	>	* on this map by other threads may be blocked while computation
1632	>	* is in progress, so the computation should be short and simple,
1633	>	* and must not attempt to update any other mappings of this map.
1634	>	*
1635	>	* @param key key with which the specified value is to be associated
1636	>	* @param mappingFunction the function to compute a value
1637	>	* @return the current (existing or computed) value associated with
1638	>	*         the specified key, or null if the computed value is null
1639	>	* @throws NullPointerException if the specified key or mappingFunction
1640	>	*         is null
1641	>	* @throws IllegalStateException if the computation detectably
1642	>	*         attempts a recursive update to this map that would
1643	>	*         otherwise never complete
1644	>	* @throws RuntimeException or Error if the mappingFunction does so,
1645	>	*         in which case the mapping is left unestablished
1646	>	*/
1647	>	public V computeIfAbsent(K key, Fun<? super K, ? extends V> mappingFunction) {
1648	>	if (key == null || mappingFunction == null)
1649	>	throw new NullPointerException();
1650	>	int h = spread(key.hashCode());
1651	>	V val = null;
1652	>	int binCount = 0;
1653	>	for (Node<K,V>[] tab = table;;) {
1654	>	Node<K,V> f; int n, i, fh;
1655	>	if (tab == null || (n = tab.length) == 0)
1656	>	tab = initTable();
1657	>	else if ((f = tabAt(tab, i = (n - 1) & h)) == null) {
1658	>	Node<K,V> r = new ReservationNode<K,V>();
1659	>	synchronized (r) {
1660	>	if (casTabAt(tab, i, null, r)) {
1661	>	binCount = 1;
1662	>	Node<K,V> node = null;
1663		try {
1664	<	val = f.map(k);
1665	<	if (val != null) {
1666	<	node.val = val;
1664	>	if ((val = mappingFunction.apply(key)) != null)
1665	>	node = new Node<K,V>(h, key, val, null);
1666	>	} finally {
1667	>	setTabAt(tab, i, node);
1668	>	}
1669	>	}
1670	>	}
1671	>	if (binCount != 0)
1672	>	break;
1673	>	}
1674	>	else if ((fh = f.hash) == MOVED)
1675	>	tab = helpTransfer(tab, f);
1676	>	else {
1677	>	boolean added = false;
1678	>	synchronized (f) {
1679	>	if (tabAt(tab, i) == f) {
1680	>	if (fh >= 0) {
1681	>	binCount = 1;
1682	>	for (Node<K,V> e = f;; ++binCount) {
1683	>	K ek; V ev;
1684	>	if (e.hash == h &&
1685	>	((ek = e.key) == key ||
1686	>	(ek != null && key.equals(ek)))) {
1687	>	val = e.val;
1688	>	break;
1689	>	}
1690	>	Node<K,V> pred = e;
1691	>	if ((e = e.next) == null) {
1692	>	if ((val = mappingFunction.apply(key)) != null) {
1693	>	added = true;
1694	>	pred.next = new Node<K,V>(h, key, val, null);
1695	>	}
1696	>	break;
1697	>	}
1698	>	}
1699	>	}
1700	>	else if (f instanceof TreeBin) {
1701	>	binCount = 2;
1702	>	TreeBin<K,V> t = (TreeBin<K,V>)f;
1703	>	TreeNode<K,V> r, p;
1704	>	if ((r = t.root) != null &&
1705	>	(p = r.findTreeNode(h, key, null)) != null)
1706	>	val = p.val;
1707	>	else if ((val = mappingFunction.apply(key)) != null) {
1708		added = true;
1709	+	t.putTreeVal(h, key, val);
1710	+	}
1711	+	}
1712	+	}
1713	+	}
1714	+	if (binCount != 0) {
1715	+	if (binCount >= TREEIFY_THRESHOLD)
1716	+	treeifyBin(tab, i);
1717	+	if (!added)
1718	+	return val;
1719	+	break;
1720	+	}
1721	+	}
1722	+	}
1723	+	if (val != null)
1724	+	addCount(1L, binCount);
1725	+	return val;
1726	+	}
1727	+
1728	+	/**
1729	+	* If the value for the specified key is present, attempts to
1730	+	* compute a new mapping given the key and its current mapped
1731	+	* value.  The entire method invocation is performed atomically.
1732	+	* Some attempted update operations on this map by other threads
1733	+	* may be blocked while computation is in progress, so the
1734	+	* computation should be short and simple, and must not attempt to
1735	+	* update any other mappings of this map.
1736	+	*
1737	+	* @param key key with which a value may be associated
1738	+	* @param remappingFunction the function to compute a value
1739	+	* @return the new value associated with the specified key, or null if none
1740	+	* @throws NullPointerException if the specified key or remappingFunction
1741	+	*         is null
1742	+	* @throws IllegalStateException if the computation detectably
1743	+	*         attempts a recursive update to this map that would
1744	+	*         otherwise never complete
1745	+	* @throws RuntimeException or Error if the remappingFunction does so,
1746	+	*         in which case the mapping is unchanged
1747	+	*/
1748	+	public V computeIfPresent(K key, BiFun<? super K, ? super V, ? extends V> remappingFunction) {
1749	+	if (key == null || remappingFunction == null)
1750	+	throw new NullPointerException();
1751	+	int h = spread(key.hashCode());
1752	+	V val = null;
1753	+	int delta = 0;
1754	+	int binCount = 0;
1755	+	for (Node<K,V>[] tab = table;;) {
1756	+	Node<K,V> f; int n, i, fh;
1757	+	if (tab == null || (n = tab.length) == 0)
1758	+	tab = initTable();
1759	+	else if ((f = tabAt(tab, i = (n - 1) & h)) == null)
1760	+	break;
1761	+	else if ((fh = f.hash) == MOVED)
1762	+	tab = helpTransfer(tab, f);
1763	+	else {
1764	+	synchronized (f) {
1765	+	if (tabAt(tab, i) == f) {
1766	+	if (fh >= 0) {
1767	+	binCount = 1;
1768	+	for (Node<K,V> e = f, pred = null;; ++binCount) {
1769	+	K ek;
1770	+	if (e.hash == h &&
1771	+	((ek = e.key) == key ||
1772	+	(ek != null && key.equals(ek)))) {
1773	+	val = remappingFunction.apply(key, e.val);
1774	+	if (val != null)
1775	+	e.val = val;
1776	+	else {
1777	+	delta = -1;
1778	+	Node<K,V> en = e.next;
1779	+	if (pred != null)
1780	+	pred.next = en;
1781	+	else
1782	+	setTabAt(tab, i, en);
1783	+	}
1784	+	break;
1785	+	}
1786	+	pred = e;
1787	+	if ((e = e.next) == null)
1788	+	break;
1789	+	}
1790	+	}
1791	+	else if (f instanceof TreeBin) {
1792	+	binCount = 2;
1793	+	TreeBin<K,V> t = (TreeBin<K,V>)f;
1794	+	TreeNode<K,V> r, p;
1795	+	if ((r = t.root) != null &&
1796	+	(p = r.findTreeNode(h, key, null)) != null) {
1797	+	val = remappingFunction.apply(key, p.val);
1798	+	if (val != null)
1799	+	p.val = val;
1800	+	else {
1801	+	delta = -1;
1802	+	if (t.removeTreeNode(p))
1803	+	setTabAt(tab, i, untreeify(t.first));
1804	+	}
1805	+	}
1806	+	}
1807	+	}
1808	+	}
1809	+	if (binCount != 0)
1810	+	break;
1811	+	}
1812	+	}
1813	+	if (delta != 0)
1814	+	addCount((long)delta, binCount);
1815	+	return val;
1816	+	}
1817	+
1818	+	/**
1819	+	* Attempts to compute a mapping for the specified key and its
1820	+	* current mapped value (or {@code null} if there is no current
1821	+	* mapping). The entire method invocation is performed atomically.
1822	+	* Some attempted update operations on this map by other threads
1823	+	* may be blocked while computation is in progress, so the
1824	+	* computation should be short and simple, and must not attempt to
1825	+	* update any other mappings of this Map.
1826	+	*
1827	+	* @param key key with which the specified value is to be associated
1828	+	* @param remappingFunction the function to compute a value
1829	+	* @return the new value associated with the specified key, or null if none
1830	+	* @throws NullPointerException if the specified key or remappingFunction
1831	+	*         is null
1832	+	* @throws IllegalStateException if the computation detectably
1833	+	*         attempts a recursive update to this map that would
1834	+	*         otherwise never complete
1835	+	* @throws RuntimeException or Error if the remappingFunction does so,
1836	+	*         in which case the mapping is unchanged
1837	+	*/
1838	+	public V compute(K key,
1839	+	BiFun<? super K, ? super V, ? extends V> remappingFunction) {
1840	+	if (key == null || remappingFunction == null)
1841	+	throw new NullPointerException();
1842	+	int h = spread(key.hashCode());
1843	+	V val = null;
1844	+	int delta = 0;
1845	+	int binCount = 0;
1846	+	for (Node<K,V>[] tab = table;;) {
1847	+	Node<K,V> f; int n, i, fh;
1848	+	if (tab == null || (n = tab.length) == 0)
1849	+	tab = initTable();
1850	+	else if ((f = tabAt(tab, i = (n - 1) & h)) == null) {
1851	+	Node<K,V> r = new ReservationNode<K,V>();
1852	+	synchronized (r) {
1853	+	if (casTabAt(tab, i, null, r)) {
1854	+	binCount = 1;
1855	+	Node<K,V> node = null;
1856	+	try {
1857	+	if ((val = remappingFunction.apply(key, null)) != null) {
1858	+	delta = 1;
1859	+	node = new Node<K,V>(h, key, val, null);
1860		}
1861		} finally {
1862	<	if (!added)
506	<	setTabAt(tab, i, null);
1862	>	setTabAt(tab, i, node);
1863		}
1864		}
1865		}
1866	<	if (validated)
1866	>	if (binCount != 0)
1867		break;
1868		}
1869	<	else if (e.hash < 0)
1870	<	tab = (Node[])e.key;
515	<	else if (Thread.holdsLock(e))
516	<	throw new IllegalStateException("Recursive map computation");
1869	>	else if ((fh = f.hash) == MOVED)
1870	>	tab = helpTransfer(tab, f);
1871		else {
1872	<	boolean validated = false;
1873	<	boolean checkSize = false;
1874	<	synchronized (e) {
1875	<	if (tabAt(tab, i) == e) {
1876	<	validated = true;
1877	<	for (Node first = e;;) {
1878	<	Object ek, ev, fv;
1879	<	if (e.hash == h &&
1880	<	(ek = e.key) != null &&
1881	<	(ev = e.val) != null &&
1882	<	(k == ek || k.equals(ek))) {
1883	<	if (replace && (fv = f.map(k)) != null)
1884	<	ev = e.val = fv;
1885	<	val = (V)ev;
1886	<	break;
1872	>	synchronized (f) {
1873	>	if (tabAt(tab, i) == f) {
1874	>	if (fh >= 0) {
1875	>	binCount = 1;
1876	>	for (Node<K,V> e = f, pred = null;; ++binCount) {
1877	>	K ek;
1878	>	if (e.hash == h &&
1879	>	((ek = e.key) == key ||
1880	>	(ek != null && key.equals(ek)))) {
1881	>	val = remappingFunction.apply(key, e.val);
1882	>	if (val != null)
1883	>	e.val = val;
1884	>	else {
1885	>	delta = -1;
1886	>	Node<K,V> en = e.next;
1887	>	if (pred != null)
1888	>	pred.next = en;
1889	>	else
1890	>	setTabAt(tab, i, en);
1891	>	}
1892	>	break;
1893	>	}
1894	>	pred = e;
1895	>	if ((e = e.next) == null) {
1896	>	val = remappingFunction.apply(key, null);
1897	>	if (val != null) {
1898	>	delta = 1;
1899	>	pred.next =
1900	>	new Node<K,V>(h, key, val, null);
1901	>	}
1902	>	break;
1903	>	}
1904		}
1905	<	Node last = e;
1906	<	if ((e = e.next) == null) {
1907	<	if ((val = f.map(k)) != null) {
1908	<	last.next = new Node(h, k, val, null);
1909	<	added = true;
1910	<	if (last != first || tab.length <= 64)
1911	<	checkSize = true;
1905	>	}
1906	>	else if (f instanceof TreeBin) {
1907	>	binCount = 1;
1908	>	TreeBin<K,V> t = (TreeBin<K,V>)f;
1909	>	TreeNode<K,V> r, p;
1910	>	if ((r = t.root) != null)
1911	>	p = r.findTreeNode(h, key, null);
1912	>	else
1913	>	p = null;
1914	>	V pv = (p == null) ? null : p.val;
1915	>	val = remappingFunction.apply(key, pv);
1916	>	if (val != null) {
1917	>	if (p != null)
1918	>	p.val = val;
1919	>	else {
1920	>	delta = 1;
1921	>	t.putTreeVal(h, key, val);
1922		}
1923	<	break;
1923	>	}
1924	>	else if (p != null) {
1925	>	delta = -1;
1926	>	if (t.removeTreeNode(p))
1927	>	setTabAt(tab, i, untreeify(t.first));
1928		}
1929		}
1930		}
1931		}
1932	<	if (validated) {
1933	<	if (checkSize && tab.length < MAXIMUM_CAPACITY &&
1934	<	resizing == 0 && counter.sum() >= threshold)
550	<	grow(0);
1932	>	if (binCount != 0) {
1933	>	if (binCount >= TREEIFY_THRESHOLD)
1934	>	treeifyBin(tab, i);
1935		break;
1936		}
1937		}
1938		}
1939	<	if (added)
1940	<	counter.increment();
1939	>	if (delta != 0)
1940	>	addCount((long)delta, binCount);
1941		return val;
1942		}
1943
1944	<	/*
1945	<	* Reclassifies nodes in each bin to new table.  Because we are
1946	<	* using power-of-two expansion, the elements from each bin must
1947	<	* either stay at same index, or move with a power of two
1948	<	* offset. We eliminate unnecessary node creation by catching
1949	<	* cases where old nodes can be reused because their next fields
1950	<	* won't change.  Statistically, at the default threshold, only
1951	<	* about one-sixth of them need cloning when a table doubles. The
1952	<	* nodes they replace will be garbage collectable as soon as they
1953	<	* are no longer referenced by any reader thread that may be in
1954	<	* the midst of concurrently traversing table.
1955	<	*
1956	<	* Transfers are done from the bottom up to preserve iterator
1957	<	* traversability. On each step, the old bin is locked,
1958	<	* moved/copied, and then replaced with a forwarding node.
1959	<	*/
1960	<	private static final void transfer(Node[] tab, Node[] nextTab) {
1961	<	int n = tab.length;
1962	<	int mask = nextTab.length - 1;
1963	<	Node fwd = new Node(MOVED, nextTab, null, null);
1964	<	for (int i = n - 1; i >= 0; --i) {
1965	<	for (Node e;;) {
1966	<	if ((e = tabAt(tab, i)) == null) {
1967	<	if (casTabAt(tab, i, e, fwd))
1968	<	break;
1944	>	/**
1945	>	* If the specified key is not already associated with a
1946	>	* (non-null) value, associates it with the given value.
1947	>	* Otherwise, replaces the value with the results of the given
1948	>	* remapping function, or removes if {@code null}. The entire
1949	>	* method invocation is performed atomically.  Some attempted
1950	>	* update operations on this map by other threads may be blocked
1951	>	* while computation is in progress, so the computation should be
1952	>	* short and simple, and must not attempt to update any other
1953	>	* mappings of this Map.
1954	>	*
1955	>	* @param key key with which the specified value is to be associated
1956	>	* @param value the value to use if absent
1957	>	* @param remappingFunction the function to recompute a value if present
1958	>	* @return the new value associated with the specified key, or null if none
1959	>	* @throws NullPointerException if the specified key or the
1960	>	*         remappingFunction is null
1961	>	* @throws RuntimeException or Error if the remappingFunction does so,
1962	>	*         in which case the mapping is unchanged
1963	>	*/
1964	>	public V merge(K key, V value, BiFun<? super V, ? super V, ? extends V> remappingFunction) {
1965	>	if (key == null || value == null || remappingFunction == null)
1966	>	throw new NullPointerException();
1967	>	int h = spread(key.hashCode());
1968	>	V val = null;
1969	>	int delta = 0;
1970	>	int binCount = 0;
1971	>	for (Node<K,V>[] tab = table;;) {
1972	>	Node<K,V> f; int n, i, fh;
1973	>	if (tab == null || (n = tab.length) == 0)
1974	>	tab = initTable();
1975	>	else if ((f = tabAt(tab, i = (n - 1) & h)) == null) {
1976	>	if (casTabAt(tab, i, null, new Node<K,V>(h, key, value, null))) {
1977	>	delta = 1;
1978	>	val = value;
1979	>	break;
1980		}
1981	<	else {
1982	<	int idx = e.hash & mask;
1983	<	boolean validated = false;
1984	<	synchronized (e) {
1985	<	if (tabAt(tab, i) == e) {
1986	<	validated = true;
1987	<	Node lastRun = e;
1988	<	for (Node p = e.next; p != null; p = p.next) {
1989	<	int j = p.hash & mask;
1990	<	if (j != idx) {
1991	<	idx = j;
1992	<	lastRun = p;
1981	>	}
1982	>	else if ((fh = f.hash) == MOVED)
1983	>	tab = helpTransfer(tab, f);
1984	>	else {
1985	>	synchronized (f) {
1986	>	if (tabAt(tab, i) == f) {
1987	>	if (fh >= 0) {
1988	>	binCount = 1;
1989	>	for (Node<K,V> e = f, pred = null;; ++binCount) {
1990	>	K ek;
1991	>	if (e.hash == h &&
1992	>	((ek = e.key) == key ||
1993	>	(ek != null && key.equals(ek)))) {
1994	>	val = remappingFunction.apply(e.val, value);
1995	>	if (val != null)
1996	>	e.val = val;
1997	>	else {
1998	>	delta = -1;
1999	>	Node<K,V> en = e.next;
2000	>	if (pred != null)
2001	>	pred.next = en;
2002	>	else
2003	>	setTabAt(tab, i, en);
2004	>	}
2005	>	break;
2006	>	}
2007	>	pred = e;
2008	>	if ((e = e.next) == null) {
2009	>	delta = 1;
2010	>	val = value;
2011	>	pred.next =
2012	>	new Node<K,V>(h, key, val, null);
2013	>	break;
2014		}
2015		}
2016	<	relaxedSetTabAt(nextTab, idx, lastRun);
2017	<	for (Node p = e; p != lastRun; p = p.next) {
2018	<	int h = p.hash;
2019	<	int j = h & mask;
2020	<	Node r = relaxedTabAt(nextTab, j);
2021	<	relaxedSetTabAt(nextTab, j,
2022	<	new Node(h, p.key, p.val, r));
2016	>	}
2017	>	else if (f instanceof TreeBin) {
2018	>	binCount = 2;
2019	>	TreeBin<K,V> t = (TreeBin<K,V>)f;
2020	>	TreeNode<K,V> r = t.root;
2021	>	TreeNode<K,V> p = (r == null) ? null :
2022	>	r.findTreeNode(h, key, null);
2023	>	val = (p == null) ? value :
2024	>	remappingFunction.apply(p.val, value);
2025	>	if (val != null) {
2026	>	if (p != null)
2027	>	p.val = val;
2028	>	else {
2029	>	delta = 1;
2030	>	t.putTreeVal(h, key, val);
2031	>	}
2032	>	}
2033	>	else if (p != null) {
2034	>	delta = -1;
2035	>	if (t.removeTreeNode(p))
2036	>	setTabAt(tab, i, untreeify(t.first));
2037		}
608	–	setTabAt(tab, i, fwd);
2038		}
2039		}
2040	<	if (validated)
2041	<	break;
2040	>	}
2041	>	if (binCount != 0) {
2042	>	if (binCount >= TREEIFY_THRESHOLD)
2043	>	treeifyBin(tab, i);
2044	>	break;
2045		}
2046		}
2047		}
2048	+	if (delta != 0)
2049	+	addCount((long)delta, binCount);
2050	+	return val;
2051		}
2052
2053	+	// Hashtable legacy methods
2054	+
2055		/**
2056	<	* If not already resizing, initializes or creates next table and
2057	<	* transfers bins. Rechecks occupancy after a transfer to see if
2058	<	* another resize is already needed because resizings are lagging
2059	<	* additions.
2060	<	*
2061	<	* @param sizeHint overridden capacity target (nonzero only from putAll)
2062	<	* @return current table
2063	<	*/
2064	<	private final Node[] grow(int sizeHint) {
2065	<	if (resizing == 0 &&
2066	<	UNSAFE.compareAndSwapInt(this, resizingOffset, 0, 1)) {
2067	<	try {
2056	>	* Legacy method testing if some key maps into the specified value
2057	>	* in this table.  This method is identical in functionality to
2058	>	* {@link #containsValue(Object)}, and exists solely to ensure
2059	>	* full compatibility with class {@link java.util.Hashtable},
2060	>	* which supported this method prior to introduction of the
2061	>	* Java Collections framework.
2062	>	*
2063	>	* @param  value a value to search for
2064	>	* @return {@code true} if and only if some key maps to the
2065	>	*         {@code value} argument in this table as
2066	>	*         determined by the {@code equals} method;
2067	>	*         {@code false} otherwise
2068	>	* @throws NullPointerException if the specified value is null
2069	>	*/
2070	>	@Deprecated public boolean contains(Object value) {
2071	>	return containsValue(value);
2072	>	}
2073	>
2074	>	/**
2075	>	* Returns an enumeration of the keys in this table.
2076	>	*
2077	>	* @return an enumeration of the keys in this table
2078	>	* @see #keySet()
2079	>	*/
2080	>	public Enumeration<K> keys() {
2081	>	Node<K,V>[] t;
2082	>	int f = (t = table) == null ? 0 : t.length;
2083	>	return new KeyIterator<K,V>(t, f, 0, f, this);
2084	>	}
2085	>
2086	>	/**
2087	>	* Returns an enumeration of the values in this table.
2088	>	*
2089	>	* @return an enumeration of the values in this table
2090	>	* @see #values()
2091	>	*/
2092	>	public Enumeration<V> elements() {
2093	>	Node<K,V>[] t;
2094	>	int f = (t = table) == null ? 0 : t.length;
2095	>	return new ValueIterator<K,V>(t, f, 0, f, this);
2096	>	}
2097	>
2098	>	// ConcurrentHashMapV8-only methods
2099	>
2100	>	/**
2101	>	* Returns the number of mappings. This method should be used
2102	>	* instead of {@link #size} because a ConcurrentHashMapV8 may
2103	>	* contain more mappings than can be represented as an int. The
2104	>	* value returned is an estimate; the actual count may differ if
2105	>	* there are concurrent insertions or removals.
2106	>	*
2107	>	* @return the number of mappings
2108	>	* @since 1.8
2109	>	*/
2110	>	public long mappingCount() {
2111	>	long n = sumCount();
2112	>	return (n < 0L) ? 0L : n; // ignore transient negative values
2113	>	}
2114	>
2115	>	/**
2116	>	* Creates a new {@link Set} backed by a ConcurrentHashMapV8
2117	>	* from the given type to {@code Boolean.TRUE}.
2118	>	*
2119	>	* @return the new set
2120	>	* @since 1.8
2121	>	*/
2122	>	public static <K> KeySetView<K,Boolean> newKeySet() {
2123	>	return new KeySetView<K,Boolean>
2124	>	(new ConcurrentHashMapV8<K,Boolean>(), Boolean.TRUE);
2125	>	}
2126	>
2127	>	/**
2128	>	* Creates a new {@link Set} backed by a ConcurrentHashMapV8
2129	>	* from the given type to {@code Boolean.TRUE}.
2130	>	*
2131	>	* @param initialCapacity The implementation performs internal
2132	>	* sizing to accommodate this many elements.
2133	>	* @return the new set
2134	>	* @throws IllegalArgumentException if the initial capacity of
2135	>	* elements is negative
2136	>	* @since 1.8
2137	>	*/
2138	>	public static <K> KeySetView<K,Boolean> newKeySet(int initialCapacity) {
2139	>	return new KeySetView<K,Boolean>
2140	>	(new ConcurrentHashMapV8<K,Boolean>(initialCapacity), Boolean.TRUE);
2141	>	}
2142	>
2143	>	/**
2144	>	* Returns a {@link Set} view of the keys in this map, using the
2145	>	* given common mapped value for any additions (i.e., {@link
2146	>	* Collection#add} and {@link Collection#addAll(Collection)}).
2147	>	* This is of course only appropriate if it is acceptable to use
2148	>	* the same value for all additions from this view.
2149	>	*
2150	>	* @param mappedValue the mapped value to use for any additions
2151	>	* @return the set view
2152	>	* @throws NullPointerException if the mappedValue is null
2153	>	*/
2154	>	public KeySetView<K,V> keySet(V mappedValue) {
2155	>	if (mappedValue == null)
2156	>	throw new NullPointerException();
2157	>	return new KeySetView<K,V>(this, mappedValue);
2158	>	}
2159	>
2160	>	/* ---------------- Special Nodes -------------- */
2161	>
2162	>	/**
2163	>	* A node inserted at head of bins during transfer operations.
2164	>	*/
2165	>	static final class ForwardingNode<K,V> extends Node<K,V> {
2166	>	final Node<K,V>[] nextTable;
2167	>	ForwardingNode(Node<K,V>[] tab) {
2168	>	super(MOVED, null, null, null);
2169	>	this.nextTable = tab;
2170	>	}
2171	>
2172	>	Node<K,V> find(int h, Object k) {
2173	>	// loop to avoid arbitrarily deep recursion on forwarding nodes
2174	>	outer: for (Node<K,V>[] tab = nextTable;;) {
2175	>	Node<K,V> e; int n;
2176	>	if (k == null || tab == null || (n = tab.length) == 0 ||
2177	>	(e = tabAt(tab, (n - 1) & h)) == null)
2178	>	return null;
2179		for (;;) {
2180	<	int cap, n;
2181	<	Node[] tab = table;
2182	<	if (tab == null) {
2183	<	int c = initCap;
2184	<	if (c < sizeHint)
2185	<	c = sizeHint;
2186	<	if (c == DEFAULT_CAPACITY)
2187	<	cap = c;
640	<	else if (c >= MAXIMUM_CAPACITY)
641	<	cap = MAXIMUM_CAPACITY;
642	<	else {
643	<	cap = MINIMUM_CAPACITY;
644	<	while (cap < c)
645	<	cap <<= 1;
2180	>	int eh; K ek;
2181	>	if ((eh = e.hash) == h &&
2182	>	((ek = e.key) == k || (ek != null && k.equals(ek))))
2183	>	return e;
2184	>	if (eh < 0) {
2185	>	if (e instanceof ForwardingNode) {
2186	>	tab = ((ForwardingNode<K,V>)e).nextTable;
2187	>	continue outer;
2188		}
2189	+	else
2190	+	return e.find(h, k);
2191		}
2192	<	else if ((n = tab.length) < MAXIMUM_CAPACITY &&
2193	<	(sizeHint <= 0 || n < sizeHint))
2194	<	cap = n << 1;
2195	<	else
2196	<	break;
2197	<	threshold = (int)(cap * loadFactor) - THRESHOLD_OFFSET;
2198	<	Node[] nextTab = new Node[cap];
2199	<	if (tab != null)
2200	<	transfer(tab, nextTab);
2201	<	table = nextTab;
2202	<	if (tab == null || cap >= MAXIMUM_CAPACITY ||
2203	<	((sizeHint > 0) ? cap >= sizeHint :
2204	<	counter.sum() < threshold))
2192	>	if ((e = e.next) == null)
2193	>	return null;
2194	>	}
2195	>	}
2196	>	}
2197	>	}
2198	>
2199	>	/**
2200	>	* A place-holder node used in computeIfAbsent and compute
2201	>	*/
2202	>	static final class ReservationNode<K,V> extends Node<K,V> {
2203	>	ReservationNode() {
2204	>	super(RESERVED, null, null, null);
2205	>	}
2206	>
2207	>	Node<K,V> find(int h, Object k) {
2208	>	return null;
2209	>	}
2210	>	}
2211	>
2212	>	/* ---------------- Table Initialization and Resizing -------------- */
2213	>
2214	>	/**
2215	>	* Returns the stamp bits for resizing a table of size n.
2216	>	* Must be negative when shifted left by RESIZE_STAMP_SHIFT.
2217	>	*/
2218	>	static final int resizeStamp(int n) {
2219	>	return Integer.numberOfLeadingZeros(n) | (1 << (RESIZE_STAMP_BITS - 1));
2220	>	}
2221	>
2222	>	/**
2223	>	* Initializes table, using the size recorded in sizeCtl.
2224	>	*/
2225	>	private final Node<K,V>[] initTable() {
2226	>	Node<K,V>[] tab; int sc;
2227	>	while ((tab = table) == null || tab.length == 0) {
2228	>	if ((sc = sizeCtl) < 0)
2229	>	Thread.yield(); // lost initialization race; just spin
2230	>	else if (U.compareAndSwapInt(this, SIZECTL, sc, -1)) {
2231	>	try {
2232	>	if ((tab = table) == null || tab.length == 0) {
2233	>	int n = (sc > 0) ? sc : DEFAULT_CAPACITY;
2234	>	@SuppressWarnings("unchecked")
2235	>	Node<K,V>[] nt = (Node<K,V>[])new Node<?,?>[n];
2236	>	table = tab = nt;
2237	>	sc = n - (n >>> 2);
2238	>	}
2239	>	} finally {
2240	>	sizeCtl = sc;
2241	>	}
2242	>	break;
2243	>	}
2244	>	}
2245	>	return tab;
2246	>	}
2247	>
2248	>	/**
2249	>	* Adds to count, and if table is too small and not already
2250	>	* resizing, initiates transfer. If already resizing, helps
2251	>	* perform transfer if work is available.  Rechecks occupancy
2252	>	* after a transfer to see if another resize is already needed
2253	>	* because resizings are lagging additions.
2254	>	*
2255	>	* @param x the count to add
2256	>	* @param check if <0, don't check resize, if <= 1 only check if uncontended
2257	>	*/
2258	>	private final void addCount(long x, int check) {
2259	>	CounterCell[] as; long b, s;
2260	>	if ((as = counterCells) != null ||
2261	>	!U.compareAndSwapLong(this, BASECOUNT, b = baseCount, s = b + x)) {
2262	>	CounterHashCode hc; CounterCell a; long v; int m;
2263	>	boolean uncontended = true;
2264	>	if ((hc = threadCounterHashCode.get()) == null ||
2265	>	as == null || (m = as.length - 1) < 0 ||
2266	>	(a = as[m & hc.code]) == null ||
2267	>	!(uncontended =
2268	>	U.compareAndSwapLong(a, CELLVALUE, v = a.value, v + x))) {
2269	>	fullAddCount(x, hc, uncontended);
2270	>	return;
2271	>	}
2272	>	if (check <= 1)
2273	>	return;
2274	>	s = sumCount();
2275	>	}
2276	>	if (check >= 0) {
2277	>	Node<K,V>[] tab, nt; int n, sc;
2278	>	while (s >= (long)(sc = sizeCtl) && (tab = table) != null &&
2279	>	(n = tab.length) < MAXIMUM_CAPACITY) {
2280	>	int rs = resizeStamp(n);
2281	>	if (sc < 0) {
2282	>	if ((sc >>> RESIZE_STAMP_SHIFT) != rs || sc == rs + 1 ||
2283	>	sc == rs + MAX_RESIZERS || (nt = nextTable) == null ||
2284	>	transferIndex <= 0)
2285		break;
2286	+	if (U.compareAndSwapInt(this, SIZECTL, sc, sc + 1))
2287	+	transfer(tab, nt);
2288		}
2289	<	} finally {
2290	<	resizing = 0;
2289	>	else if (U.compareAndSwapInt(this, SIZECTL, sc,
2290	>	(rs << RESIZE_STAMP_SHIFT) + 2))
2291	>	transfer(tab, null);
2292	>	s = sumCount();
2293		}
2294		}
2295	<	else if (table == null)
2296	<	Thread.yield(); // lost initialization race; just spin
2295	>	}
2296	>
2297	>	/**
2298	>	* Helps transfer if a resize is in progress.
2299	>	*/
2300	>	final Node<K,V>[] helpTransfer(Node<K,V>[] tab, Node<K,V> f) {
2301	>	Node<K,V>[] nextTab; int sc;
2302	>	if (tab != null && (f instanceof ForwardingNode) &&
2303	>	(nextTab = ((ForwardingNode<K,V>)f).nextTable) != null) {
2304	>	int rs = resizeStamp(tab.length);
2305	>	while (nextTab == nextTable && table == tab &&
2306	>	(sc = sizeCtl) < 0) {
2307	>	if ((sc >>> RESIZE_STAMP_SHIFT) != rs || sc == rs + 1 ||
2308	>	sc == rs + MAX_RESIZERS || transferIndex <= 0)
2309	>	break;
2310	>	if (U.compareAndSwapInt(this, SIZECTL, sc, sc + 1)) {
2311	>	transfer(tab, nextTab);
2312	>	break;
2313	>	}
2314	>	}
2315	>	return nextTab;
2316	>	}
2317		return table;
2318		}
2319
2320		/**
2321	<	* Implementation for putAll and constructor with Map
2322	<	* argument. Tries to first override initial capacity or grow
2323	<	* based on map size to pre-allocate table space.
2321	>	* Tries to presize table to accommodate the given number of elements.
2322	>	*
2323	>	* @param size number of elements (doesn't need to be perfectly accurate)
2324		*/
2325	<	private final void internalPutAll(Map<? extends K, ? extends V> m) {
2326	<	int s = m.size();
2327	<	grow((s >= (MAXIMUM_CAPACITY >>> 1)) ? s : s + (s >>> 1));
2328	<	for (Map.Entry<? extends K, ? extends V> e : m.entrySet()) {
2329	<	Object k = e.getKey();
2330	<	Object v = e.getValue();
2331	<	if (k == null || v == null)
2332	<	throw new NullPointerException();
2333	<	internalPut(k, v, true);
2325	>	private final void tryPresize(int size) {
2326	>	int c = (size >= (MAXIMUM_CAPACITY >>> 1)) ? MAXIMUM_CAPACITY :
2327	>	tableSizeFor(size + (size >>> 1) + 1);
2328	>	int sc;
2329	>	while ((sc = sizeCtl) >= 0) {
2330	>	Node<K,V>[] tab = table; int n;
2331	>	if (tab == null || (n = tab.length) == 0) {
2332	>	n = (sc > c) ? sc : c;
2333	>	if (U.compareAndSwapInt(this, SIZECTL, sc, -1)) {
2334	>	try {
2335	>	if (table == tab) {
2336	>	@SuppressWarnings("unchecked")
2337	>	Node<K,V>[] nt = (Node<K,V>[])new Node<?,?>[n];
2338	>	table = nt;
2339	>	sc = n - (n >>> 2);
2340	>	}
2341	>	} finally {
2342	>	sizeCtl = sc;
2343	>	}
2344	>	}
2345	>	}
2346	>	else if (c <= sc || n >= MAXIMUM_CAPACITY)
2347	>	break;
2348	>	else if (tab == table) {
2349	>	int rs = resizeStamp(n);
2350	>	if (sc < 0) {
2351	>	Node<K,V>[] nt;
2352	>	if ((sc >>> RESIZE_STAMP_SHIFT) != rs || sc == rs + 1 ||
2353	>	sc == rs + MAX_RESIZERS || (nt = nextTable) == null ||
2354	>	transferIndex <= 0)
2355	>	break;
2356	>	if (U.compareAndSwapInt(this, SIZECTL, sc, sc + 1))
2357	>	transfer(tab, nt);
2358	>	}
2359	>	else if (U.compareAndSwapInt(this, SIZECTL, sc,
2360	>	(rs << RESIZE_STAMP_SHIFT) + 2))
2361	>	transfer(tab, null);
2362	>	}
2363		}
2364		}
2365
2366		/**
2367	<	* Implementation for clear. Steps through each bin, removing all nodes.
2367	>	* Moves and/or copies the nodes in each bin to new table. See
2368	>	* above for explanation.
2369		*/
2370	<	private final void internalClear() {
2371	<	long deletions = 0L;
2372	<	int i = 0;
2373	<	Node[] tab = table;
2374	<	while (tab != null && i < tab.length) {
2375	<	Node e = tabAt(tab, i);
2376	<	if (e == null)
2377	<	++i;
2378	<	else if (e.hash < 0)
2379	<	tab = (Node[])e.key;
2370	>	private final void transfer(Node<K,V>[] tab, Node<K,V>[] nextTab) {
2371	>	int n = tab.length, stride;
2372	>	if ((stride = (NCPU > 1) ? (n >>> 3) / NCPU : n) < MIN_TRANSFER_STRIDE)
2373	>	stride = MIN_TRANSFER_STRIDE; // subdivide range
2374	>	if (nextTab == null) {            // initiating
2375	>	try {
2376	>	@SuppressWarnings("unchecked")
2377	>	Node<K,V>[] nt = (Node<K,V>[])new Node<?,?>[n << 1];
2378	>	nextTab = nt;
2379	>	} catch (Throwable ex) {      // try to cope with OOME
2380	>	sizeCtl = Integer.MAX_VALUE;
2381	>	return;
2382	>	}
2383	>	nextTable = nextTab;
2384	>	transferIndex = n;
2385	>	}
2386	>	int nextn = nextTab.length;
2387	>	ForwardingNode<K,V> fwd = new ForwardingNode<K,V>(nextTab);
2388	>	boolean advance = true;
2389	>	boolean finishing = false; // to ensure sweep before committing nextTab
2390	>	for (int i = 0, bound = 0;;) {
2391	>	Node<K,V> f; int fh;
2392	>	while (advance) {
2393	>	int nextIndex, nextBound;
2394	>	if (--i >= bound || finishing)
2395	>	advance = false;
2396	>	else if ((nextIndex = transferIndex) <= 0) {
2397	>	i = -1;
2398	>	advance = false;
2399	>	}
2400	>	else if (U.compareAndSwapInt
2401	>	(this, TRANSFERINDEX, nextIndex,
2402	>	nextBound = (nextIndex > stride ?
2403	>	nextIndex - stride : 0))) {
2404	>	bound = nextBound;
2405	>	i = nextIndex - 1;
2406	>	advance = false;
2407	>	}
2408	>	}
2409	>	if (i < 0 || i >= n || i + n >= nextn) {
2410	>	int sc;
2411	>	if (finishing) {
2412	>	nextTable = null;
2413	>	table = nextTab;
2414	>	sizeCtl = (n << 1) - (n >>> 1);
2415	>	return;
2416	>	}
2417	>	if (U.compareAndSwapInt(this, SIZECTL, sc = sizeCtl, sc - 1)) {
2418	>	if ((sc - 2) != resizeStamp(n) << RESIZE_STAMP_SHIFT)
2419	>	return;
2420	>	finishing = advance = true;
2421	>	i = n; // recheck before commit
2422	>	}
2423	>	}
2424	>	else if ((f = tabAt(tab, i)) == null)
2425	>	advance = casTabAt(tab, i, null, fwd);
2426	>	else if ((fh = f.hash) == MOVED)
2427	>	advance = true; // already processed
2428		else {
2429	<	boolean validated = false;
2430	<	synchronized (e) {
2431	<	if (tabAt(tab, i) == e) {
2432	<	validated = true;
2433	<	do {
2434	<	if (e.val != null) {
2435	<	e.val = null;
2436	<	++deletions;
2429	>	synchronized (f) {
2430	>	if (tabAt(tab, i) == f) {
2431	>	Node<K,V> ln, hn;
2432	>	if (fh >= 0) {
2433	>	int runBit = fh & n;
2434	>	Node<K,V> lastRun = f;
2435	>	for (Node<K,V> p = f.next; p != null; p = p.next) {
2436	>	int b = p.hash & n;
2437	>	if (b != runBit) {
2438	>	runBit = b;
2439	>	lastRun = p;
2440	>	}
2441	>	}
2442	>	if (runBit == 0) {
2443	>	ln = lastRun;
2444	>	hn = null;
2445	>	}
2446	>	else {
2447	>	hn = lastRun;
2448	>	ln = null;
2449		}
2450	<	} while ((e = e.next) != null);
2451	<	setTabAt(tab, i, null);
2450	>	for (Node<K,V> p = f; p != lastRun; p = p.next) {
2451	>	int ph = p.hash; K pk = p.key; V pv = p.val;
2452	>	if ((ph & n) == 0)
2453	>	ln = new Node<K,V>(ph, pk, pv, ln);
2454	>	else
2455	>	hn = new Node<K,V>(ph, pk, pv, hn);
2456	>	}
2457	>	setTabAt(nextTab, i, ln);
2458	>	setTabAt(nextTab, i + n, hn);
2459	>	setTabAt(tab, i, fwd);
2460	>	advance = true;
2461	>	}
2462	>	else if (f instanceof TreeBin) {
2463	>	TreeBin<K,V> t = (TreeBin<K,V>)f;
2464	>	TreeNode<K,V> lo = null, loTail = null;
2465	>	TreeNode<K,V> hi = null, hiTail = null;
2466	>	int lc = 0, hc = 0;
2467	>	for (Node<K,V> e = t.first; e != null; e = e.next) {
2468	>	int h = e.hash;
2469	>	TreeNode<K,V> p = new TreeNode<K,V>
2470	>	(h, e.key, e.val, null, null);
2471	>	if ((h & n) == 0) {
2472	>	if ((p.prev = loTail) == null)
2473	>	lo = p;
2474	>	else
2475	>	loTail.next = p;
2476	>	loTail = p;
2477	>	++lc;
2478	>	}
2479	>	else {
2480	>	if ((p.prev = hiTail) == null)
2481	>	hi = p;
2482	>	else
2483	>	hiTail.next = p;
2484	>	hiTail = p;
2485	>	++hc;
2486	>	}
2487	>	}
2488	>	ln = (lc <= UNTREEIFY_THRESHOLD) ? untreeify(lo) :
2489	>	(hc != 0) ? new TreeBin<K,V>(lo) : t;
2490	>	hn = (hc <= UNTREEIFY_THRESHOLD) ? untreeify(hi) :
2491	>	(lc != 0) ? new TreeBin<K,V>(hi) : t;
2492	>	setTabAt(nextTab, i, ln);
2493	>	setTabAt(nextTab, i + n, hn);
2494	>	setTabAt(tab, i, fwd);
2495	>	advance = true;
2496	>	}
2497		}
2498		}
2499	<	if (validated) {
2500	<	++i;
2501	<	if (deletions > THRESHOLD_OFFSET) { // bound lag in counts
2502	<	counter.add(-deletions);
2503	<	deletions = 0L;
2499	>	}
2500	>	}
2501	>	}
2502	>
2503	>	/* ---------------- Conversion from/to TreeBins -------------- */
2504	>
2505	>	/**
2506	>	* Replaces all linked nodes in bin at given index unless table is
2507	>	* too small, in which case resizes instead.
2508	>	*/
2509	>	private final void treeifyBin(Node<K,V>[] tab, int index) {
2510	>	Node<K,V> b; int n, sc;
2511	>	if (tab != null) {
2512	>	if ((n = tab.length) < MIN_TREEIFY_CAPACITY)
2513	>	tryPresize(n << 1);
2514	>	else if ((b = tabAt(tab, index)) != null && b.hash >= 0) {
2515	>	synchronized (b) {
2516	>	if (tabAt(tab, index) == b) {
2517	>	TreeNode<K,V> hd = null, tl = null;
2518	>	for (Node<K,V> e = b; e != null; e = e.next) {
2519	>	TreeNode<K,V> p =
2520	>	new TreeNode<K,V>(e.hash, e.key, e.val,
2521	>	null, null);
2522	>	if ((p.prev = tl) == null)
2523	>	hd = p;
2524	>	else
2525	>	tl.next = p;
2526	>	tl = p;
2527	>	}
2528	>	setTabAt(tab, index, new TreeBin<K,V>(hd));
2529		}
2530		}
2531		}
2532		}
725	–	if (deletions != 0L)
726	–	counter.add(-deletions);
2533		}
2534
2535		/**
2536	<	* Base class for key, value, and entry iterators, plus internal
731	<	* implementations of public traversal-based methods, to avoid
732	<	* duplicating traversal code.
2536	>	* Returns a list on non-TreeNodes replacing those in given list.
2537		*/
2538	<	class HashIterator {
2539	<	private Node next;          // the next entry to return
2540	<	private Node[] tab;         // current table; updated if resized
2541	<	private Node lastReturned;  // the last entry returned, for remove
2542	<	private Object nextVal;     // cached value of next
2543	<	private int index;          // index of bin to use next
2544	<	private int baseIndex;      // current index of initial table
2545	<	private final int baseSize; // initial table size
2538	>	static <K,V> Node<K,V> untreeify(Node<K,V> b) {
2539	>	Node<K,V> hd = null, tl = null;
2540	>	for (Node<K,V> q = b; q != null; q = q.next) {
2541	>	Node<K,V> p = new Node<K,V>(q.hash, q.key, q.val, null);
2542	>	if (tl == null)
2543	>	hd = p;
2544	>	else
2545	>	tl.next = p;
2546	>	tl = p;
2547	>	}
2548	>	return hd;
2549	>	}
2550
2551	<	HashIterator() {
2552	<	Node[] t = tab = table;
2553	<	if (t == null)
2554	<	baseSize = 0;
2555	<	else {
2556	<	baseSize = t.length;
2557	<	advance(null);
2558	<	}
2551	>	/* ---------------- TreeNodes -------------- */
2552	>
2553	>	/**
2554	>	* Nodes for use in TreeBins
2555	>	*/
2556	>	static final class TreeNode<K,V> extends Node<K,V> {
2557	>	TreeNode<K,V> parent;  // red-black tree links
2558	>	TreeNode<K,V> left;
2559	>	TreeNode<K,V> right;
2560	>	TreeNode<K,V> prev;    // needed to unlink next upon deletion
2561	>	boolean red;
2562	>
2563	>	TreeNode(int hash, K key, V val, Node<K,V> next,
2564	>	TreeNode<K,V> parent) {
2565	>	super(hash, key, val, next);
2566	>	this.parent = parent;
2567		}
2568
2569	<	public final boolean hasNext()         { return next != null; }
2570	<	public final boolean hasMoreElements() { return next != null; }
2569	>	Node<K,V> find(int h, Object k) {
2570	>	return findTreeNode(h, k, null);
2571	>	}
2572
2573		/**
2574	<	* Advances next.  Normally, iteration proceeds bin-by-bin
2575	<	* traversing lists.  However, if the table has been resized,
759	<	* then all future steps must traverse both the bin at the
760	<	* current index as well as at (index + baseSize); and so on
761	<	* for further resizings. To paranoically cope with potential
762	<	* (improper) sharing of iterators across threads, table reads
763	<	* are bounds-checked.
2574	>	* Returns the TreeNode (or null if not found) for the given key
2575	>	* starting at given root.
2576		*/
2577	<	final void advance(Node e) {
2578	<	for (;;) {
2579	<	Node[] t; int i; // for bounds checks
2580	<	if (e != null) {
2581	<	Object ek = e.key, ev = e.val;
2582	<	if (ev != null && ek != null) {
2583	<	nextVal = ev;
2584	<	next = e;
2585	<	break;
2586	<	}
2587	<	e = e.next;
2588	<	}
2589	<	else if (baseIndex < baseSize && (t = tab) != null &&
2590	<	t.length > (i = index) && i >= 0) {
2591	<	if ((e = tabAt(t, i)) != null && e.hash < 0) {
2592	<	tab = (Node[])e.key;
2593	<	e = null;
2594	<	}
2595	<	else if (i + baseSize < t.length)
2596	<	index += baseSize;    // visit forwarded upper slots
2577	>	final TreeNode<K,V> findTreeNode(int h, Object k, Class<?> kc) {
2578	>	if (k != null) {
2579	>	TreeNode<K,V> p = this;
2580	>	do {
2581	>	int ph, dir; K pk; TreeNode<K,V> q;
2582	>	TreeNode<K,V> pl = p.left, pr = p.right;
2583	>	if ((ph = p.hash) > h)
2584	>	p = pl;
2585	>	else if (ph < h)
2586	>	p = pr;
2587	>	else if ((pk = p.key) == k || (pk != null && k.equals(pk)))
2588	>	return p;
2589	>	else if (pl == null)
2590	>	p = pr;
2591	>	else if (pr == null)
2592	>	p = pl;
2593	>	else if ((kc != null ||
2594	>	(kc = comparableClassFor(k)) != null) &&
2595	>	(dir = compareComparables(kc, k, pk)) != 0)
2596	>	p = (dir < 0) ? pl : pr;
2597	>	else if ((q = pr.findTreeNode(h, k, kc)) != null)
2598	>	return q;
2599		else
2600	<	index = ++baseIndex;
2600	>	p = pl;
2601	>	} while (p != null);
2602	>	}
2603	>	return null;
2604	>	}
2605	>	}
2606	>
2607	>	/* ---------------- TreeBins -------------- */
2608	>
2609	>	/**
2610	>	* TreeNodes used at the heads of bins. TreeBins do not hold user
2611	>	* keys or values, but instead point to list of TreeNodes and
2612	>	* their root. They also maintain a parasitic read-write lock
2613	>	* forcing writers (who hold bin lock) to wait for readers (who do
2614	>	* not) to complete before tree restructuring operations.
2615	>	*/
2616	>	static final class TreeBin<K,V> extends Node<K,V> {
2617	>	TreeNode<K,V> root;
2618	>	volatile TreeNode<K,V> first;
2619	>	volatile Thread waiter;
2620	>	volatile int lockState;
2621	>	// values for lockState
2622	>	static final int WRITER = 1; // set while holding write lock
2623	>	static final int WAITER = 2; // set when waiting for write lock
2624	>	static final int READER = 4; // increment value for setting read lock
2625	>
2626	>	/**
2627	>	* Tie-breaking utility for ordering insertions when equal
2628	>	* hashCodes and non-comparable. We don't require a total
2629	>	* order, just a consistent insertion rule to maintain
2630	>	* equivalence across rebalancings. Tie-breaking further than
2631	>	* necessary simplifies testing a bit.
2632	>	*/
2633	>	static int tieBreakOrder(Object a, Object b) {
2634	>	int d;
2635	>	if (a == null || b == null ||
2636	>	(d = a.getClass().getName().
2637	>	compareTo(b.getClass().getName())) == 0)
2638	>	d = (System.identityHashCode(a) <= System.identityHashCode(b) ?
2639	>	-1 : 1);
2640	>	return d;
2641	>	}
2642	>
2643	>	/**
2644	>	* Creates bin with initial set of nodes headed by b.
2645	>	*/
2646	>	TreeBin(TreeNode<K,V> b) {
2647	>	super(TREEBIN, null, null, null);
2648	>	this.first = b;
2649	>	TreeNode<K,V> r = null;
2650	>	for (TreeNode<K,V> x = b, next; x != null; x = next) {
2651	>	next = (TreeNode<K,V>)x.next;
2652	>	x.left = x.right = null;
2653	>	if (r == null) {
2654	>	x.parent = null;
2655	>	x.red = false;
2656	>	r = x;
2657		}
2658		else {
2659	<	next = null;
2660	<	break;
2659	>	K k = x.key;
2660	>	int h = x.hash;
2661	>	Class<?> kc = null;
2662	>	for (TreeNode<K,V> p = r;;) {
2663	>	int dir, ph;
2664	>	K pk = p.key;
2665	>	if ((ph = p.hash) > h)
2666	>	dir = -1;
2667	>	else if (ph < h)
2668	>	dir = 1;
2669	>	else if ((kc == null &&
2670	>	(kc = comparableClassFor(k)) == null) ||
2671	>	(dir = compareComparables(kc, k, pk)) == 0)
2672	>	dir = tieBreakOrder(k, pk);
2673	>	TreeNode<K,V> xp = p;
2674	>	if ((p = (dir <= 0) ? p.left : p.right) == null) {
2675	>	x.parent = xp;
2676	>	if (dir <= 0)
2677	>	xp.left = x;
2678	>	else
2679	>	xp.right = x;
2680	>	r = balanceInsertion(r, x);
2681	>	break;
2682	>	}
2683	>	}
2684		}
2685		}
2686	+	this.root = r;
2687	+	assert checkInvariants(root);
2688		}
2689
2690	<	final Object nextKey() {
2691	<	Node e = next;
2692	<	if (e == null)
2693	<	throw new NoSuchElementException();
2694	<	Object k = e.key;
2695	<	advance((lastReturned = e).next);
801	<	return k;
2690	>	/**
2691	>	* Acquires write lock for tree restructuring.
2692	>	*/
2693	>	private final void lockRoot() {
2694	>	if (!U.compareAndSwapInt(this, LOCKSTATE, 0, WRITER))
2695	>	contendedLock(); // offload to separate method
2696		}
2697
2698	<	final Object nextValue() {
2699	<	Node e = next;
2700	<	if (e == null)
2701	<	throw new NoSuchElementException();
2702	<	Object v = nextVal;
809	<	advance((lastReturned = e).next);
810	<	return v;
2698	>	/**
2699	>	* Releases write lock for tree restructuring.
2700	>	*/
2701	>	private final void unlockRoot() {
2702	>	lockState = 0;
2703		}
2704
2705	<	final WriteThroughEntry nextEntry() {
2706	<	Node e = next;
2707	<	if (e == null)
2708	<	throw new NoSuchElementException();
2709	<	WriteThroughEntry entry =
2710	<	new WriteThroughEntry(e.key, nextVal);
2711	<	advance((lastReturned = e).next);
2712	<	return entry;
2705	>	/**
2706	>	* Possibly blocks awaiting root lock.
2707	>	*/
2708	>	private final void contendedLock() {
2709	>	boolean waiting = false;
2710	>	for (int s;;) {
2711	>	if (((s = lockState) & ~WAITER) == 0) {
2712	>	if (U.compareAndSwapInt(this, LOCKSTATE, s, WRITER)) {
2713	>	if (waiting)
2714	>	waiter = null;
2715	>	return;
2716	>	}
2717	>	}
2718	>	else if ((s & WAITER) == 0) {
2719	>	if (U.compareAndSwapInt(this, LOCKSTATE, s, s | WAITER)) {
2720	>	waiting = true;
2721	>	waiter = Thread.currentThread();
2722	>	}
2723	>	}
2724	>	else if (waiting)
2725	>	LockSupport.park(this);
2726	>	}
2727		}
2728
2729	<	public final void remove() {
2730	<	if (lastReturned == null)
2731	<	throw new IllegalStateException();
2732	<	ConcurrentHashMapV8.this.remove(lastReturned.key);
2733	<	lastReturned = null;
2729	>	/**
2730	>	* Returns matching node or null if none. Tries to search
2731	>	* using tree comparisons from root, but continues linear
2732	>	* search when lock not available.
2733	>	*/
2734	>	final Node<K,V> find(int h, Object k) {
2735	>	if (k != null) {
2736	>	for (Node<K,V> e = first; e != null; ) {
2737	>	int s; K ek;
2738	>	if (((s = lockState) & (WAITER|WRITER)) != 0) {
2739	>	if (e.hash == h &&
2740	>	((ek = e.key) == k || (ek != null && k.equals(ek))))
2741	>	return e;
2742	>	e = e.next;
2743	>	}
2744	>	else if (U.compareAndSwapInt(this, LOCKSTATE, s,
2745	>	s + READER)) {
2746	>	TreeNode<K,V> r, p;
2747	>	try {
2748	>	p = ((r = root) == null ? null :
2749	>	r.findTreeNode(h, k, null));
2750	>	} finally {
2751	>	Thread w;
2752	>	int ls;
2753	>	do {} while (!U.compareAndSwapInt
2754	>	(this, LOCKSTATE,
2755	>	ls = lockState, ls - READER));
2756	>	if (ls == (READER|WAITER) && (w = waiter) != null)
2757	>	LockSupport.unpark(w);
2758	>	}
2759	>	return p;
2760	>	}
2761	>	}
2762	>	}
2763	>	return null;
2764		}
2765
2766	<	/** Helper for serialization */
2767	<	final void writeEntries(java.io.ObjectOutputStream s)
2768	<	throws java.io.IOException {
2769	<	Node e;
2770	<	while ((e = next) != null) {
2771	<	s.writeObject(e.key);
2772	<	s.writeObject(nextVal);
2773	<	advance(e.next);
2766	>	/**
2767	>	* Finds or adds a node.
2768	>	* @return null if added
2769	>	*/
2770	>	final TreeNode<K,V> putTreeVal(int h, K k, V v) {
2771	>	Class<?> kc = null;
2772	>	boolean searched = false;
2773	>	for (TreeNode<K,V> p = root;;) {
2774	>	int dir, ph; K pk;
2775	>	if (p == null) {
2776	>	first = root = new TreeNode<K,V>(h, k, v, null, null);
2777	>	break;
2778	>	}
2779	>	else if ((ph = p.hash) > h)
2780	>	dir = -1;
2781	>	else if (ph < h)
2782	>	dir = 1;
2783	>	else if ((pk = p.key) == k || (pk != null && k.equals(pk)))
2784	>	return p;
2785	>	else if ((kc == null &&
2786	>	(kc = comparableClassFor(k)) == null) ||
2787	>	(dir = compareComparables(kc, k, pk)) == 0) {
2788	>	if (!searched) {
2789	>	TreeNode<K,V> q, ch;
2790	>	searched = true;
2791	>	if (((ch = p.left) != null &&
2792	>	(q = ch.findTreeNode(h, k, kc)) != null) ||
2793	>	((ch = p.right) != null &&
2794	>	(q = ch.findTreeNode(h, k, kc)) != null))
2795	>	return q;
2796	>	}
2797	>	dir = tieBreakOrder(k, pk);
2798	>	}
2799	>
2800	>	TreeNode<K,V> xp = p;
2801	>	if ((p = (dir <= 0) ? p.left : p.right) == null) {
2802	>	TreeNode<K,V> x, f = first;
2803	>	first = x = new TreeNode<K,V>(h, k, v, f, xp);
2804	>	if (f != null)
2805	>	f.prev = x;
2806	>	if (dir <= 0)
2807	>	xp.left = x;
2808	>	else
2809	>	xp.right = x;
2810	>	if (!xp.red)
2811	>	x.red = true;
2812	>	else {
2813	>	lockRoot();
2814	>	try {
2815	>	root = balanceInsertion(root, x);
2816	>	} finally {
2817	>	unlockRoot();
2818	>	}
2819	>	}
2820	>	break;
2821	>	}
2822		}
2823	+	assert checkInvariants(root);
2824	+	return null;
2825		}
2826
2827	<	/** Helper for containsValue */
2828	<	final boolean containsVal(Object value) {
2829	<	if (value != null) {
2830	<	Node e;
2831	<	while ((e = next) != null) {
2832	<	Object v = nextVal;
2833	<	if (value == v || value.equals(v))
2834	<	return true;
2835	<	advance(e.next);
2827	>	/**
2828	>	* Removes the given node, that must be present before this
2829	>	* call.  This is messier than typical red-black deletion code
2830	>	* because we cannot swap the contents of an interior node
2831	>	* with a leaf successor that is pinned by "next" pointers
2832	>	* that are accessible independently of lock. So instead we
2833	>	* swap the tree linkages.
2834	>	*
2835	>	* @return true if now too small, so should be untreeified
2836	>	*/
2837	>	final boolean removeTreeNode(TreeNode<K,V> p) {
2838	>	TreeNode<K,V> next = (TreeNode<K,V>)p.next;
2839	>	TreeNode<K,V> pred = p.prev;  // unlink traversal pointers
2840	>	TreeNode<K,V> r, rl;
2841	>	if (pred == null)
2842	>	first = next;
2843	>	else
2844	>	pred.next = next;
2845	>	if (next != null)
2846	>	next.prev = pred;
2847	>	if (first == null) {
2848	>	root = null;
2849	>	return true;
2850	>	}
2851	>	if ((r = root) == null || r.right == null || // too small
2852	>	(rl = r.left) == null || rl.left == null)
2853	>	return true;
2854	>	lockRoot();
2855	>	try {
2856	>	TreeNode<K,V> replacement;
2857	>	TreeNode<K,V> pl = p.left;
2858	>	TreeNode<K,V> pr = p.right;
2859	>	if (pl != null && pr != null) {
2860	>	TreeNode<K,V> s = pr, sl;
2861	>	while ((sl = s.left) != null) // find successor
2862	>	s = sl;
2863	>	boolean c = s.red; s.red = p.red; p.red = c; // swap colors
2864	>	TreeNode<K,V> sr = s.right;
2865	>	TreeNode<K,V> pp = p.parent;
2866	>	if (s == pr) { // p was s's direct parent
2867	>	p.parent = s;
2868	>	s.right = p;
2869	>	}
2870	>	else {
2871	>	TreeNode<K,V> sp = s.parent;
2872	>	if ((p.parent = sp) != null) {
2873	>	if (s == sp.left)
2874	>	sp.left = p;
2875	>	else
2876	>	sp.right = p;
2877	>	}
2878	>	if ((s.right = pr) != null)
2879	>	pr.parent = s;
2880	>	}
2881	>	p.left = null;
2882	>	if ((p.right = sr) != null)
2883	>	sr.parent = p;
2884	>	if ((s.left = pl) != null)
2885	>	pl.parent = s;
2886	>	if ((s.parent = pp) == null)
2887	>	r = s;
2888	>	else if (p == pp.left)
2889	>	pp.left = s;
2890	>	else
2891	>	pp.right = s;
2892	>	if (sr != null)
2893	>	replacement = sr;
2894	>	else
2895	>	replacement = p;
2896		}
2897	+	else if (pl != null)
2898	+	replacement = pl;
2899	+	else if (pr != null)
2900	+	replacement = pr;
2901	+	else
2902	+	replacement = p;
2903	+	if (replacement != p) {
2904	+	TreeNode<K,V> pp = replacement.parent = p.parent;
2905	+	if (pp == null)
2906	+	r = replacement;
2907	+	else if (p == pp.left)
2908	+	pp.left = replacement;
2909	+	else
2910	+	pp.right = replacement;
2911	+	p.left = p.right = p.parent = null;
2912	+	}
2913	+
2914	+	root = (p.red) ? r : balanceDeletion(r, replacement);
2915	+
2916	+	if (p == replacement) {  // detach pointers
2917	+	TreeNode<K,V> pp;
2918	+	if ((pp = p.parent) != null) {
2919	+	if (p == pp.left)
2920	+	pp.left = null;
2921	+	else if (p == pp.right)
2922	+	pp.right = null;
2923	+	p.parent = null;
2924	+	}
2925	+	}
2926	+	} finally {
2927	+	unlockRoot();
2928		}
2929	+	assert checkInvariants(root);
2930		return false;
2931		}
2932
2933	<	/** Helper for Map.hashCode */
2934	<	final int mapHashCode() {
2935	<	int h = 0;
2936	<	Node e;
2937	<	while ((e = next) != null) {
2938	<	h += e.key.hashCode() ^ nextVal.hashCode();
2939	<	advance(e.next);
2933	>	/* ------------------------------------------------------------ */
2934	>	// Red-black tree methods, all adapted from CLR
2935	>
2936	>	static <K,V> TreeNode<K,V> rotateLeft(TreeNode<K,V> root,
2937	>	TreeNode<K,V> p) {
2938	>	TreeNode<K,V> r, pp, rl;
2939	>	if (p != null && (r = p.right) != null) {
2940	>	if ((rl = p.right = r.left) != null)
2941	>	rl.parent = p;
2942	>	if ((pp = r.parent = p.parent) == null)
2943	>	(root = r).red = false;
2944	>	else if (pp.left == p)
2945	>	pp.left = r;
2946	>	else
2947	>	pp.right = r;
2948	>	r.left = p;
2949	>	p.parent = r;
2950		}
2951	<	return h;
2951	>	return root;
2952		}
2953
2954	<	/** Helper for Map.toString */
2955	<	final String mapToString() {
2956	<	Node e = next;
2957	<	if (e == null)
2958	<	return "{}";
2959	<	StringBuilder sb = new StringBuilder();
2960	<	sb.append('{');
2961	<	for (;;) {
2962	<	sb.append(e.key   == this ? "(this Map)" : e.key);
2963	<	sb.append('=');
876	<	sb.append(nextVal == this ? "(this Map)" : nextVal);
877	<	advance(e.next);
878	<	if ((e = next) != null)
879	<	sb.append(',').append(' ');
2954	>	static <K,V> TreeNode<K,V> rotateRight(TreeNode<K,V> root,
2955	>	TreeNode<K,V> p) {
2956	>	TreeNode<K,V> l, pp, lr;
2957	>	if (p != null && (l = p.left) != null) {
2958	>	if ((lr = p.left = l.right) != null)
2959	>	lr.parent = p;
2960	>	if ((pp = l.parent = p.parent) == null)
2961	>	(root = l).red = false;
2962	>	else if (pp.right == p)
2963	>	pp.right = l;
2964		else
2965	<	return sb.append('}').toString();
2965	>	pp.left = l;
2966	>	l.right = p;
2967	>	p.parent = l;
2968	>	}
2969	>	return root;
2970	>	}
2971	>
2972	>	static <K,V> TreeNode<K,V> balanceInsertion(TreeNode<K,V> root,
2973	>	TreeNode<K,V> x) {
2974	>	x.red = true;
2975	>	for (TreeNode<K,V> xp, xpp, xppl, xppr;;) {
2976	>	if ((xp = x.parent) == null) {
2977	>	x.red = false;
2978	>	return x;
2979	>	}
2980	>	else if (!xp.red || (xpp = xp.parent) == null)
2981	>	return root;
2982	>	if (xp == (xppl = xpp.left)) {
2983	>	if ((xppr = xpp.right) != null && xppr.red) {
2984	>	xppr.red = false;
2985	>	xp.red = false;
2986	>	xpp.red = true;
2987	>	x = xpp;
2988	>	}
2989	>	else {
2990	>	if (x == xp.right) {
2991	>	root = rotateLeft(root, x = xp);
2992	>	xpp = (xp = x.parent) == null ? null : xp.parent;
2993	>	}
2994	>	if (xp != null) {
2995	>	xp.red = false;
2996	>	if (xpp != null) {
2997	>	xpp.red = true;
2998	>	root = rotateRight(root, xpp);
2999	>	}
3000	>	}
3001	>	}
3002	>	}
3003	>	else {
3004	>	if (xppl != null && xppl.red) {
3005	>	xppl.red = false;
3006	>	xp.red = false;
3007	>	xpp.red = true;
3008	>	x = xpp;
3009	>	}
3010	>	else {
3011	>	if (x == xp.left) {
3012	>	root = rotateRight(root, x = xp);
3013	>	xpp = (xp = x.parent) == null ? null : xp.parent;
3014	>	}
3015	>	if (xp != null) {
3016	>	xp.red = false;
3017	>	if (xpp != null) {
3018	>	xpp.red = true;
3019	>	root = rotateLeft(root, xpp);
3020	>	}
3021	>	}
3022	>	}
3023	>	}
3024	>	}
3025	>	}
3026	>
3027	>	static <K,V> TreeNode<K,V> balanceDeletion(TreeNode<K,V> root,
3028	>	TreeNode<K,V> x) {
3029	>	for (TreeNode<K,V> xp, xpl, xpr;;) {
3030	>	if (x == null || x == root)
3031	>	return root;
3032	>	else if ((xp = x.parent) == null) {
3033	>	x.red = false;
3034	>	return x;
3035	>	}
3036	>	else if (x.red) {
3037	>	x.red = false;
3038	>	return root;
3039	>	}
3040	>	else if ((xpl = xp.left) == x) {
3041	>	if ((xpr = xp.right) != null && xpr.red) {
3042	>	xpr.red = false;
3043	>	xp.red = true;
3044	>	root = rotateLeft(root, xp);
3045	>	xpr = (xp = x.parent) == null ? null : xp.right;
3046	>	}
3047	>	if (xpr == null)
3048	>	x = xp;
3049	>	else {
3050	>	TreeNode<K,V> sl = xpr.left, sr = xpr.right;
3051	>	if ((sr == null || !sr.red) &&
3052	>	(sl == null || !sl.red)) {
3053	>	xpr.red = true;
3054	>	x = xp;
3055	>	}
3056	>	else {
3057	>	if (sr == null || !sr.red) {
3058	>	if (sl != null)
3059	>	sl.red = false;
3060	>	xpr.red = true;
3061	>	root = rotateRight(root, xpr);
3062	>	xpr = (xp = x.parent) == null ?
3063	>	null : xp.right;
3064	>	}
3065	>	if (xpr != null) {
3066	>	xpr.red = (xp == null) ? false : xp.red;
3067	>	if ((sr = xpr.right) != null)
3068	>	sr.red = false;
3069	>	}
3070	>	if (xp != null) {
3071	>	xp.red = false;
3072	>	root = rotateLeft(root, xp);
3073	>	}
3074	>	x = root;
3075	>	}
3076	>	}
3077	>	}
3078	>	else { // symmetric
3079	>	if (xpl != null && xpl.red) {
3080	>	xpl.red = false;
3081	>	xp.red = true;
3082	>	root = rotateRight(root, xp);
3083	>	xpl = (xp = x.parent) == null ? null : xp.left;
3084	>	}
3085	>	if (xpl == null)
3086	>	x = xp;
3087	>	else {
3088	>	TreeNode<K,V> sl = xpl.left, sr = xpl.right;
3089	>	if ((sl == null || !sl.red) &&
3090	>	(sr == null || !sr.red)) {
3091	>	xpl.red = true;
3092	>	x = xp;
3093	>	}
3094	>	else {
3095	>	if (sl == null || !sl.red) {
3096	>	if (sr != null)
3097	>	sr.red = false;
3098	>	xpl.red = true;
3099	>	root = rotateLeft(root, xpl);
3100	>	xpl = (xp = x.parent) == null ?
3101	>	null : xp.left;
3102	>	}
3103	>	if (xpl != null) {
3104	>	xpl.red = (xp == null) ? false : xp.red;
3105	>	if ((sl = xpl.left) != null)
3106	>	sl.red = false;
3107	>	}
3108	>	if (xp != null) {
3109	>	xp.red = false;
3110	>	root = rotateRight(root, xp);
3111	>	}
3112	>	x = root;
3113	>	}
3114	>	}
3115	>	}
3116	>	}
3117	>	}
3118	>
3119	>	/**
3120	>	* Recursive invariant check
3121	>	*/
3122	>	static <K,V> boolean checkInvariants(TreeNode<K,V> t) {
3123	>	TreeNode<K,V> tp = t.parent, tl = t.left, tr = t.right,
3124	>	tb = t.prev, tn = (TreeNode<K,V>)t.next;
3125	>	if (tb != null && tb.next != t)
3126	>	return false;
3127	>	if (tn != null && tn.prev != t)
3128	>	return false;
3129	>	if (tp != null && t != tp.left && t != tp.right)
3130	>	return false;
3131	>	if (tl != null && (tl.parent != t || tl.hash > t.hash))
3132	>	return false;
3133	>	if (tr != null && (tr.parent != t || tr.hash < t.hash))
3134	>	return false;
3135	>	if (t.red && tl != null && tl.red && tr != null && tr.red)
3136	>	return false;
3137	>	if (tl != null && !checkInvariants(tl))
3138	>	return false;
3139	>	if (tr != null && !checkInvariants(tr))
3140	>	return false;
3141	>	return true;
3142	>	}
3143	>
3144	>	private static final sun.misc.Unsafe U;
3145	>	private static final long LOCKSTATE;
3146	>	static {
3147	>	try {
3148	>	U = getUnsafe();
3149	>	Class<?> k = TreeBin.class;
3150	>	LOCKSTATE = U.objectFieldOffset
3151	>	(k.getDeclaredField("lockState"));
3152	>	} catch (Exception e) {
3153	>	throw new Error(e);
3154		}
3155		}
3156		}
3157
3158	<	/* ---------------- Public operations -------------- */
3158	>	/* ----------------Table Traversal -------------- */
3159
3160		/**
3161	<	* Creates a new, empty map with the specified initial
3162	<	* capacity, load factor and concurrency level.
3163	<	*
3164	<	* @param initialCapacity the initial capacity. The implementation
3165	<	* performs internal sizing to accommodate this many elements.
3166	<	* @param loadFactor  the load factor threshold, used to control resizing.
3167	<	* Resizing may be performed when the average number of elements per
3168	<	* bin exceeds this threshold.
3169	<	* @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.
903	<	*/
904	<	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();
3161	>	* Records the table, its length, and current traversal index for a
3162	>	* traverser that must process a region of a forwarded table before
3163	>	* proceeding with current table.
3164	>	*/
3165	>	static final class TableStack<K,V> {
3166	>	int length;
3167	>	int index;
3168	>	Node<K,V>[] tab;
3169	>	TableStack<K,V> next;
3170		}
3171
3172		/**
3173	<	* Creates a new, empty map with the specified initial capacity
3174	<	* and load factor and with the default concurrencyLevel (16).
3175	<	*
3176	<	* @param initialCapacity The implementation performs internal
3177	<	* sizing to accommodate this many elements.
3178	<	* @param loadFactor  the load factor threshold, used to control resizing.
3179	<	* Resizing may be performed when the average number of elements per
3180	<	* bin exceeds this threshold.
3181	<	* @throws IllegalArgumentException if the initial capacity of
3182	<	* elements is negative or the load factor is nonpositive
3183	<	*
3184	<	* @since 1.6
3185	<	*/
3186	<	public ConcurrentHashMapV8(int initialCapacity, float loadFactor) {
3187	<	this(initialCapacity, loadFactor, DEFAULT_CONCURRENCY_LEVEL);
3173	>	* Encapsulates traversal for methods such as containsValue; also
3174	>	* serves as a base class for other iterators and spliterators.
3175	>	*
3176	>	* Method advance visits once each still-valid node that was
3177	>	* reachable upon iterator construction. It might miss some that
3178	>	* were added to a bin after the bin was visited, which is OK wrt
3179	>	* consistency guarantees. Maintaining this property in the face
3180	>	* of possible ongoing resizes requires a fair amount of
3181	>	* bookkeeping state that is difficult to optimize away amidst
3182	>	* volatile accesses.  Even so, traversal maintains reasonable
3183	>	* throughput.
3184	>	*
3185	>	* Normally, iteration proceeds bin-by-bin traversing lists.
3186	>	* However, if the table has been resized, then all future steps
3187	>	* must traverse both the bin at the current index as well as at
3188	>	* (index + baseSize); and so on for further resizings. To
3189	>	* paranoically cope with potential sharing by users of iterators
3190	>	* across threads, iteration terminates if a bounds checks fails
3191	>	* for a table read.
3192	>	*/
3193	>	static class Traverser<K,V> {
3194	>	Node<K,V>[] tab;        // current table; updated if resized
3195	>	Node<K,V> next;         // the next entry to use
3196	>	TableStack<K,V> stack, spare; // to save/restore on ForwardingNodes
3197	>	int index;              // index of bin to use next
3198	>	int baseIndex;          // current index of initial table
3199	>	int baseLimit;          // index bound for initial table
3200	>	final int baseSize;     // initial table size
3201	>
3202	>	Traverser(Node<K,V>[] tab, int size, int index, int limit) {
3203	>	this.tab = tab;
3204	>	this.baseSize = size;
3205	>	this.baseIndex = this.index = index;
3206	>	this.baseLimit = limit;
3207	>	this.next = null;
3208	>	}
3209	>
3210	>	/**
3211	>	* Advances if possible, returning next valid node, or null if none.
3212	>	*/
3213	>	final Node<K,V> advance() {
3214	>	Node<K,V> e;
3215	>	if ((e = next) != null)
3216	>	e = e.next;
3217	>	for (;;) {
3218	>	Node<K,V>[] t; int i, n;  // must use locals in checks
3219	>	if (e != null)
3220	>	return next = e;
3221	>	if (baseIndex >= baseLimit || (t = tab) == null ||
3222	>	(n = t.length) <= (i = index) || i < 0)
3223	>	return next = null;
3224	>	if ((e = tabAt(t, i)) != null && e.hash < 0) {
3225	>	if (e instanceof ForwardingNode) {
3226	>	tab = ((ForwardingNode<K,V>)e).nextTable;
3227	>	e = null;
3228	>	pushState(t, i, n);
3229	>	continue;
3230	>	}
3231	>	else if (e instanceof TreeBin)
3232	>	e = ((TreeBin<K,V>)e).first;
3233	>	else
3234	>	e = null;
3235	>	}
3236	>	if (stack != null)
3237	>	recoverState(n);
3238	>	else if ((index = i + baseSize) >= n)
3239	>	index = ++baseIndex; // visit upper slots if present
3240	>	}
3241	>	}
3242	>
3243	>	/**
3244	>	* Saves traversal state upon encountering a forwarding node.
3245	>	*/
3246	>	private void pushState(Node<K,V>[] t, int i, int n) {
3247	>	TableStack<K,V> s = spare;  // reuse if possible
3248	>	if (s != null)
3249	>	spare = s.next;
3250	>	else
3251	>	s = new TableStack<K,V>();
3252	>	s.tab = t;
3253	>	s.length = n;
3254	>	s.index = i;
3255	>	s.next = stack;
3256	>	stack = s;
3257	>	}
3258	>
3259	>	/**
3260	>	* Possibly pops traversal state.
3261	>	*
3262	>	* @param n length of current table
3263	>	*/
3264	>	private void recoverState(int n) {
3265	>	TableStack<K,V> s; int len;
3266	>	while ((s = stack) != null && (index += (len = s.length)) >= n) {
3267	>	n = len;
3268	>	index = s.index;
3269	>	tab = s.tab;
3270	>	s.tab = null;
3271	>	TableStack<K,V> next = s.next;
3272	>	s.next = spare; // save for reuse
3273	>	stack = next;
3274	>	spare = s;
3275	>	}
3276	>	if (s == null && (index += baseSize) >= n)
3277	>	index = ++baseIndex;
3278	>	}
3279		}
3280
3281		/**
3282	<	* Creates a new, empty map with the specified initial capacity,
3283	<	* and with default load factor (0.75) and concurrencyLevel (16).
934	<	*
935	<	* @param initialCapacity the initial capacity. The implementation
936	<	* performs internal sizing to accommodate this many elements.
937	<	* @throws IllegalArgumentException if the initial capacity of
938	<	* elements is negative.
3282	>	* Base of key, value, and entry Iterators. Adds fields to
3283	>	* Traverser to support iterator.remove.
3284		*/
3285	<	public ConcurrentHashMapV8(int initialCapacity) {
3286	<	this(initialCapacity, DEFAULT_LOAD_FACTOR, DEFAULT_CONCURRENCY_LEVEL);
3285	>	static class BaseIterator<K,V> extends Traverser<K,V> {
3286	>	final ConcurrentHashMapV8<K,V> map;
3287	>	Node<K,V> lastReturned;
3288	>	BaseIterator(Node<K,V>[] tab, int size, int index, int limit,
3289	>	ConcurrentHashMapV8<K,V> map) {
3290	>	super(tab, size, index, limit);
3291	>	this.map = map;
3292	>	advance();
3293	>	}
3294	>
3295	>	public final boolean hasNext() { return next != null; }
3296	>	public final boolean hasMoreElements() { return next != null; }
3297	>
3298	>	public final void remove() {
3299	>	Node<K,V> p;
3300	>	if ((p = lastReturned) == null)
3301	>	throw new IllegalStateException();
3302	>	lastReturned = null;
3303	>	map.replaceNode(p.key, null, null);
3304	>	}
3305	>	}
3306	>
3307	>	static final class KeyIterator<K,V> extends BaseIterator<K,V>
3308	>	implements Iterator<K>, Enumeration<K> {
3309	>	KeyIterator(Node<K,V>[] tab, int index, int size, int limit,
3310	>	ConcurrentHashMapV8<K,V> map) {
3311	>	super(tab, index, size, limit, map);
3312	>	}
3313	>
3314	>	public final K next() {
3315	>	Node<K,V> p;
3316	>	if ((p = next) == null)
3317	>	throw new NoSuchElementException();
3318	>	K k = p.key;
3319	>	lastReturned = p;
3320	>	advance();
3321	>	return k;
3322	>	}
3323	>
3324	>	public final K nextElement() { return next(); }
3325	>	}
3326	>
3327	>	static final class ValueIterator<K,V> extends BaseIterator<K,V>
3328	>	implements Iterator<V>, Enumeration<V> {
3329	>	ValueIterator(Node<K,V>[] tab, int index, int size, int limit,
3330	>	ConcurrentHashMapV8<K,V> map) {
3331	>	super(tab, index, size, limit, map);
3332	>	}
3333	>
3334	>	public final V next() {
3335	>	Node<K,V> p;
3336	>	if ((p = next) == null)
3337	>	throw new NoSuchElementException();
3338	>	V v = p.val;
3339	>	lastReturned = p;
3340	>	advance();
3341	>	return v;
3342	>	}
3343	>
3344	>	public final V nextElement() { return next(); }
3345	>	}
3346	>
3347	>	static final class EntryIterator<K,V> extends BaseIterator<K,V>
3348	>	implements Iterator<Map.Entry<K,V>> {
3349	>	EntryIterator(Node<K,V>[] tab, int index, int size, int limit,
3350	>	ConcurrentHashMapV8<K,V> map) {
3351	>	super(tab, index, size, limit, map);
3352	>	}
3353	>
3354	>	public final Map.Entry<K,V> next() {
3355	>	Node<K,V> p;
3356	>	if ((p = next) == null)
3357	>	throw new NoSuchElementException();
3358	>	K k = p.key;
3359	>	V v = p.val;
3360	>	lastReturned = p;
3361	>	advance();
3362	>	return new MapEntry<K,V>(k, v, map);
3363	>	}
3364		}
3365
3366		/**
3367	<	* Creates a new, empty map with a default initial capacity (16),
946	<	* load factor (0.75) and concurrencyLevel (16).
3367	>	* Exported Entry for EntryIterator
3368		*/
3369	<	public ConcurrentHashMapV8() {
3370	<	this(DEFAULT_CAPACITY, DEFAULT_LOAD_FACTOR, DEFAULT_CONCURRENCY_LEVEL);
3369	>	static final class MapEntry<K,V> implements Map.Entry<K,V> {
3370	>	final K key; // non-null
3371	>	V val;       // non-null
3372	>	final ConcurrentHashMapV8<K,V> map;
3373	>	MapEntry(K key, V val, ConcurrentHashMapV8<K,V> map) {
3374	>	this.key = key;
3375	>	this.val = val;
3376	>	this.map = map;
3377	>	}
3378	>	public K getKey()        { return key; }
3379	>	public V getValue()      { return val; }
3380	>	public int hashCode()    { return key.hashCode() ^ val.hashCode(); }
3381	>	public String toString() { return key + "=" + val; }
3382	>
3383	>	public boolean equals(Object o) {
3384	>	Object k, v; Map.Entry<?,?> e;
3385	>	return ((o instanceof Map.Entry) &&
3386	>	(k = (e = (Map.Entry<?,?>)o).getKey()) != null &&
3387	>	(v = e.getValue()) != null &&
3388	>	(k == key || k.equals(key)) &&
3389	>	(v == val || v.equals(val)));
3390	>	}
3391	>
3392	>	/**
3393	>	* Sets our entry's value and writes through to the map. The
3394	>	* value to return is somewhat arbitrary here. Since we do not
3395	>	* necessarily track asynchronous changes, the most recent
3396	>	* "previous" value could be different from what we return (or
3397	>	* could even have been removed, in which case the put will
3398	>	* re-establish). We do not and cannot guarantee more.
3399	>	*/
3400	>	public V setValue(V value) {
3401	>	if (value == null) throw new NullPointerException();
3402	>	V v = val;
3403	>	val = value;
3404	>	map.put(key, value);
3405	>	return v;
3406	>	}
3407	>	}
3408	>
3409	>	static final class KeySpliterator<K,V> extends Traverser<K,V>
3410	>	implements ConcurrentHashMapSpliterator<K> {
3411	>	long est;               // size estimate
3412	>	KeySpliterator(Node<K,V>[] tab, int size, int index, int limit,
3413	>	long est) {
3414	>	super(tab, size, index, limit);
3415	>	this.est = est;
3416	>	}
3417	>
3418	>	public ConcurrentHashMapSpliterator<K> trySplit() {
3419	>	int i, f, h;
3420	>	return (h = ((i = baseIndex) + (f = baseLimit)) >>> 1) <= i ? null :
3421	>	new KeySpliterator<K,V>(tab, baseSize, baseLimit = h,
3422	>	f, est >>>= 1);
3423	>	}
3424	>
3425	>	public void forEachRemaining(Action<? super K> action) {
3426	>	if (action == null) throw new NullPointerException();
3427	>	for (Node<K,V> p; (p = advance()) != null;)
3428	>	action.apply(p.key);
3429	>	}
3430	>
3431	>	public boolean tryAdvance(Action<? super K> action) {
3432	>	if (action == null) throw new NullPointerException();
3433	>	Node<K,V> p;
3434	>	if ((p = advance()) == null)
3435	>	return false;
3436	>	action.apply(p.key);
3437	>	return true;
3438	>	}
3439	>
3440	>	public long estimateSize() { return est; }
3441	>
3442	>	}
3443	>
3444	>	static final class ValueSpliterator<K,V> extends Traverser<K,V>
3445	>	implements ConcurrentHashMapSpliterator<V> {
3446	>	long est;               // size estimate
3447	>	ValueSpliterator(Node<K,V>[] tab, int size, int index, int limit,
3448	>	long est) {
3449	>	super(tab, size, index, limit);
3450	>	this.est = est;
3451	>	}
3452	>
3453	>	public ConcurrentHashMapSpliterator<V> trySplit() {
3454	>	int i, f, h;
3455	>	return (h = ((i = baseIndex) + (f = baseLimit)) >>> 1) <= i ? null :
3456	>	new ValueSpliterator<K,V>(tab, baseSize, baseLimit = h,
3457	>	f, est >>>= 1);
3458	>	}
3459	>
3460	>	public void forEachRemaining(Action<? super V> action) {
3461	>	if (action == null) throw new NullPointerException();
3462	>	for (Node<K,V> p; (p = advance()) != null;)
3463	>	action.apply(p.val);
3464	>	}
3465	>
3466	>	public boolean tryAdvance(Action<? super V> action) {
3467	>	if (action == null) throw new NullPointerException();
3468	>	Node<K,V> p;
3469	>	if ((p = advance()) == null)
3470	>	return false;
3471	>	action.apply(p.val);
3472	>	return true;
3473	>	}
3474	>
3475	>	public long estimateSize() { return est; }
3476	>
3477		}
3478
3479	+	static final class EntrySpliterator<K,V> extends Traverser<K,V>
3480	+	implements ConcurrentHashMapSpliterator<Map.Entry<K,V>> {
3481	+	final ConcurrentHashMapV8<K,V> map; // To export MapEntry
3482	+	long est;               // size estimate
3483	+	EntrySpliterator(Node<K,V>[] tab, int size, int index, int limit,
3484	+	long est, ConcurrentHashMapV8<K,V> map) {
3485	+	super(tab, size, index, limit);
3486	+	this.map = map;
3487	+	this.est = est;
3488	+	}
3489	+
3490	+	public ConcurrentHashMapSpliterator<Map.Entry<K,V>> trySplit() {
3491	+	int i, f, h;
3492	+	return (h = ((i = baseIndex) + (f = baseLimit)) >>> 1) <= i ? null :
3493	+	new EntrySpliterator<K,V>(tab, baseSize, baseLimit = h,
3494	+	f, est >>>= 1, map);
3495	+	}
3496	+
3497	+	public void forEachRemaining(Action<? super Map.Entry<K,V>> action) {
3498	+	if (action == null) throw new NullPointerException();
3499	+	for (Node<K,V> p; (p = advance()) != null; )
3500	+	action.apply(new MapEntry<K,V>(p.key, p.val, map));
3501	+	}
3502	+
3503	+	public boolean tryAdvance(Action<? super Map.Entry<K,V>> action) {
3504	+	if (action == null) throw new NullPointerException();
3505	+	Node<K,V> p;
3506	+	if ((p = advance()) == null)
3507	+	return false;
3508	+	action.apply(new MapEntry<K,V>(p.key, p.val, map));
3509	+	return true;
3510	+	}
3511	+
3512	+	public long estimateSize() { return est; }
3513	+
3514	+	}
3515	+
3516	+	// Parallel bulk operations
3517	+
3518		/**
3519	<	* Creates a new map with the same mappings as the given map.
3520	<	* The map is created with a capacity of 1.5 times the number
3521	<	* of mappings in the given map or 16 (whichever is greater),
3522	<	* and a default load factor (0.75) and concurrencyLevel (16).
3523	<	*
3524	<	* @param m the map
3519	>	* Computes initial batch value for bulk tasks. The returned value
3520	>	* is approximately exp2 of the number of times (minus one) to
3521	>	* split task by two before executing leaf action. This value is
3522	>	* faster to compute and more convenient to use as a guide to
3523	>	* splitting than is the depth, since it is used while dividing by
3524	>	* two anyway.
3525		*/
3526	<	public ConcurrentHashMapV8(Map<? extends K, ? extends V> m) {
3527	<	this(DEFAULT_CAPACITY, DEFAULT_LOAD_FACTOR, DEFAULT_CONCURRENCY_LEVEL);
3528	<	if (m == null)
3529	<	throw new NullPointerException();
3530	<	internalPutAll(m);
3526	>	final int batchFor(long b) {
3527	>	long n;
3528	>	if (b == Long.MAX_VALUE || (n = sumCount()) <= 1L || n < b)
3529	>	return 0;
3530	>	int sp = ForkJoinPool.getCommonPoolParallelism() << 2; // slack of 4
3531	>	return (b <= 0L || (n /= b) >= sp) ? sp : (int)n;
3532		}
3533
3534		/**
3535	<	* Returns {@code true} if this map contains no key-value mappings.
3535	>	* Performs the given action for each (key, value).
3536		*
3537	<	* @return {@code true} if this map contains no key-value mappings
3537	>	* @param parallelismThreshold the (estimated) number of elements
3538	>	* needed for this operation to be executed in parallel
3539	>	* @param action the action
3540	>	* @since 1.8
3541		*/
3542	<	public boolean isEmpty() {
3543	<	return counter.sum() <= 0L; // ignore transient negative values
3542	>	public void forEach(long parallelismThreshold,
3543	>	BiAction<? super K,? super V> action) {
3544	>	if (action == null) throw new NullPointerException();
3545	>	new ForEachMappingTask<K,V>
3546	>	(null, batchFor(parallelismThreshold), 0, 0, table,
3547	>	action).invoke();
3548		}
3549
3550		/**
3551	<	* Returns the number of key-value mappings in this map.  If the
3552	<	* map contains more than {@code Integer.MAX_VALUE} elements, returns
979	<	* {@code Integer.MAX_VALUE}.
3551	>	* Performs the given action for each non-null transformation
3552	>	* of each (key, value).
3553		*
3554	<	* @return the number of key-value mappings in this map
3554	>	* @param parallelismThreshold the (estimated) number of elements
3555	>	* needed for this operation to be executed in parallel
3556	>	* @param transformer a function returning the transformation
3557	>	* for an element, or null if there is no transformation (in
3558	>	* which case the action is not applied)
3559	>	* @param action the action
3560	>	* @since 1.8
3561		*/
3562	<	public int size() {
3563	<	long n = counter.sum();
3564	<	return ((n >>> 31) == 0) ? (int)n : (n < 0L) ? 0 : Integer.MAX_VALUE;
3562	>	public <U> void forEach(long parallelismThreshold,
3563	>	BiFun<? super K, ? super V, ? extends U> transformer,
3564	>	Action<? super U> action) {
3565	>	if (transformer == null || action == null)
3566	>	throw new NullPointerException();
3567	>	new ForEachTransformedMappingTask<K,V,U>
3568	>	(null, batchFor(parallelismThreshold), 0, 0, table,
3569	>	transformer, action).invoke();
3570		}
3571
3572		/**
3573	<	* Returns the value to which the specified key is mapped,
3574	<	* or {@code null} if this map contains no mapping for the key.
3575	<	*
3576	<	* <p>More formally, if this map contains a mapping from a key
3577	<	* {@code k} to a value {@code v} such that {@code key.equals(k)},
3578	<	* then this method returns {@code v}; otherwise it returns
3579	<	* {@code null}.  (There can be at most one such mapping.)
3580	<	*
3581	<	* @throws NullPointerException if the specified key is null
3582	<	*/
3583	<	@SuppressWarnings("unchecked")
3584	<	public V get(Object key) {
3585	<	if (key == null)
3573	>	* Returns a non-null result from applying the given search
3574	>	* function on each (key, value), or null if none.  Upon
3575	>	* success, further element processing is suppressed and the
3576	>	* results of any other parallel invocations of the search
3577	>	* function are ignored.
3578	>	*
3579	>	* @param parallelismThreshold the (estimated) number of elements
3580	>	* needed for this operation to be executed in parallel
3581	>	* @param searchFunction a function returning a non-null
3582	>	* result on success, else null
3583	>	* @return a non-null result from applying the given search
3584	>	* function on each (key, value), or null if none
3585	>	* @since 1.8
3586	>	*/
3587	>	public <U> U search(long parallelismThreshold,
3588	>	BiFun<? super K, ? super V, ? extends U> searchFunction) {
3589	>	if (searchFunction == null) throw new NullPointerException();
3590	>	return new SearchMappingsTask<K,V,U>
3591	>	(null, batchFor(parallelismThreshold), 0, 0, table,
3592	>	searchFunction, new AtomicReference<U>()).invoke();
3593	>	}
3594	>
3595	>	/**
3596	>	* Returns the result of accumulating the given transformation
3597	>	* of all (key, value) pairs using the given reducer to
3598	>	* combine values, or null if none.
3599	>	*
3600	>	* @param parallelismThreshold the (estimated) number of elements
3601	>	* needed for this operation to be executed in parallel
3602	>	* @param transformer a function returning the transformation
3603	>	* for an element, or null if there is no transformation (in
3604	>	* which case it is not combined)
3605	>	* @param reducer a commutative associative combining function
3606	>	* @return the result of accumulating the given transformation
3607	>	* of all (key, value) pairs
3608	>	* @since 1.8
3609	>	*/
3610	>	public <U> U reduce(long parallelismThreshold,
3611	>	BiFun<? super K, ? super V, ? extends U> transformer,
3612	>	BiFun<? super U, ? super U, ? extends U> reducer) {
3613	>	if (transformer == null || reducer == null)
3614		throw new NullPointerException();
3615	<	return (V)internalGet(key);
3615	>	return new MapReduceMappingsTask<K,V,U>
3616	>	(null, batchFor(parallelismThreshold), 0, 0, table,
3617	>	null, transformer, reducer).invoke();
3618		}
3619
3620		/**
3621	<	* Tests if the specified object is a key in this table.
3622	<	*
3623	<	* @param  key   possible key
3624	<	* @return {@code true} if and only if the specified object
3625	<	*         is a key in this table, as determined by the
3626	<	*         {@code equals} method; {@code false} otherwise.
3627	<	* @throws NullPointerException if the specified key is null
3628	<	*/
3629	<	public boolean containsKey(Object key) {
3630	<	if (key == null)
3621	>	* Returns the result of accumulating the given transformation
3622	>	* of all (key, value) pairs using the given reducer to
3623	>	* combine values, and the given basis as an identity value.
3624	>	*
3625	>	* @param parallelismThreshold the (estimated) number of elements
3626	>	* needed for this operation to be executed in parallel
3627	>	* @param transformer a function returning the transformation
3628	>	* for an element
3629	>	* @param basis the identity (initial default value) for the reduction
3630	>	* @param reducer a commutative associative combining function
3631	>	* @return the result of accumulating the given transformation
3632	>	* of all (key, value) pairs
3633	>	* @since 1.8
3634	>	*/
3635	>	public double reduceToDouble(long parallelismThreshold,
3636	>	ObjectByObjectToDouble<? super K, ? super V> transformer,
3637	>	double basis,
3638	>	DoubleByDoubleToDouble reducer) {
3639	>	if (transformer == null || reducer == null)
3640		throw new NullPointerException();
3641	<	return internalGet(key) != null;
3641	>	return new MapReduceMappingsToDoubleTask<K,V>
3642	>	(null, batchFor(parallelismThreshold), 0, 0, table,
3643	>	null, transformer, basis, reducer).invoke();
3644		}
3645
3646		/**
3647	<	* Returns {@code true} if this map maps one or more keys to the
3648	<	* specified value. Note: This method requires a full internal
3649	<	* traversal of the hash table, and so is much slower than
3650	<	* method {@code containsKey}.
3651	<	*
3652	<	* @param value value whose presence in this map is to be tested
3653	<	* @return {@code true} if this map maps one or more keys to the
3654	<	*         specified value
3655	<	* @throws NullPointerException if the specified value is null
3656	<	*/
3657	<	public boolean containsValue(Object value) {
3658	<	if (value == null)
3647	>	* Returns the result of accumulating the given transformation
3648	>	* of all (key, value) pairs using the given reducer to
3649	>	* combine values, and the given basis as an identity value.
3650	>	*
3651	>	* @param parallelismThreshold the (estimated) number of elements
3652	>	* needed for this operation to be executed in parallel
3653	>	* @param transformer a function returning the transformation
3654	>	* for an element
3655	>	* @param basis the identity (initial default value) for the reduction
3656	>	* @param reducer a commutative associative combining function
3657	>	* @return the result of accumulating the given transformation
3658	>	* of all (key, value) pairs
3659	>	* @since 1.8
3660	>	*/
3661	>	public long reduceToLong(long parallelismThreshold,
3662	>	ObjectByObjectToLong<? super K, ? super V> transformer,
3663	>	long basis,
3664	>	LongByLongToLong reducer) {
3665	>	if (transformer == null || reducer == null)
3666		throw new NullPointerException();
3667	<	return new HashIterator().containsVal(value);
3667	>	return new MapReduceMappingsToLongTask<K,V>
3668	>	(null, batchFor(parallelismThreshold), 0, 0, table,
3669	>	null, transformer, basis, reducer).invoke();
3670		}
3671
3672		/**
3673	<	* Legacy method testing if some key maps into the specified value
3674	<	* in this table.  This method is identical in functionality to
3675	<	* {@link #containsValue}, and exists solely to ensure
3676	<	* full compatibility with class {@link java.util.Hashtable},
3677	<	* which supported this method prior to introduction of the
3678	<	* Java Collections framework.
3679	<	*
3680	<	* @param  value a value to search for
3681	<	* @return {@code true} if and only if some key maps to the
3682	<	*         {@code value} argument in this table as
3683	<	*         determined by the {@code equals} method;
3684	<	*         {@code false} otherwise
3685	<	* @throws NullPointerException if the specified value is null
3686	<	*/
3687	<	public boolean contains(Object value) {
3688	<	return containsValue(value);
3673	>	* Returns the result of accumulating the given transformation
3674	>	* of all (key, value) pairs using the given reducer to
3675	>	* combine values, and the given basis as an identity value.
3676	>	*
3677	>	* @param parallelismThreshold the (estimated) number of elements
3678	>	* needed for this operation to be executed in parallel
3679	>	* @param transformer a function returning the transformation
3680	>	* for an element
3681	>	* @param basis the identity (initial default value) for the reduction
3682	>	* @param reducer a commutative associative combining function
3683	>	* @return the result of accumulating the given transformation
3684	>	* of all (key, value) pairs
3685	>	* @since 1.8
3686	>	*/
3687	>	public int reduceToInt(long parallelismThreshold,
3688	>	ObjectByObjectToInt<? super K, ? super V> transformer,
3689	>	int basis,
3690	>	IntByIntToInt reducer) {
3691	>	if (transformer == null || reducer == null)
3692	>	throw new NullPointerException();
3693	>	return new MapReduceMappingsToIntTask<K,V>
3694	>	(null, batchFor(parallelismThreshold), 0, 0, table,
3695	>	null, transformer, basis, reducer).invoke();
3696		}
3697
3698		/**
3699	<	* Maps the specified key to the specified value in this table.
1059	<	* Neither the key nor the value can be null.
1060	<	*
1061	<	* <p> The value can be retrieved by calling the {@code get} method
1062	<	* with a key that is equal to the original key.
3699	>	* Performs the given action for each key.
3700		*
3701	<	* @param key key with which the specified value is to be associated
3702	<	* @param value value to be associated with the specified key
3703	<	* @return the previous value associated with {@code key}, or
3704	<	*         {@code null} if there was no mapping for {@code key}
1068	<	* @throws NullPointerException if the specified key or value is null
3701	>	* @param parallelismThreshold the (estimated) number of elements
3702	>	* needed for this operation to be executed in parallel
3703	>	* @param action the action
3704	>	* @since 1.8
3705		*/
3706	<	@SuppressWarnings("unchecked")
3707	<	public V put(K key, V value) {
3708	<	if (key == null || value == null)
3709	<	throw new NullPointerException();
3710	<	return (V)internalPut(key, value, true);
3706	>	public void forEachKey(long parallelismThreshold,
3707	>	Action<? super K> action) {
3708	>	if (action == null) throw new NullPointerException();
3709	>	new ForEachKeyTask<K,V>
3710	>	(null, batchFor(parallelismThreshold), 0, 0, table,
3711	>	action).invoke();
3712		}
3713
3714		/**
3715	<	* {@inheritDoc}
3715	>	* Performs the given action for each non-null transformation
3716	>	* of each key.
3717		*
3718	<	* @return the previous value associated with the specified key,
3719	<	*         or {@code null} if there was no mapping for the key
3720	<	* @throws NullPointerException if the specified key or value is null
3718	>	* @param parallelismThreshold the (estimated) number of elements
3719	>	* needed for this operation to be executed in parallel
3720	>	* @param transformer a function returning the transformation
3721	>	* for an element, or null if there is no transformation (in
3722	>	* which case the action is not applied)
3723	>	* @param action the action
3724	>	* @since 1.8
3725		*/
3726	<	@SuppressWarnings("unchecked")
3727	<	public V putIfAbsent(K key, V value) {
3728	<	if (key == null || value == null)
3726	>	public <U> void forEachKey(long parallelismThreshold,
3727	>	Fun<? super K, ? extends U> transformer,
3728	>	Action<? super U> action) {
3729	>	if (transformer == null || action == null)
3730	>	throw new NullPointerException();
3731	>	new ForEachTransformedKeyTask<K,V,U>
3732	>	(null, batchFor(parallelismThreshold), 0, 0, table,
3733	>	transformer, action).invoke();
3734	>	}
3735	>
3736	>	/**
3737	>	* Returns a non-null result from applying the given search
3738	>	* function on each key, or null if none. Upon success,
3739	>	* further element processing is suppressed and the results of
3740	>	* any other parallel invocations of the search function are
3741	>	* ignored.
3742	>	*
3743	>	* @param parallelismThreshold the (estimated) number of elements
3744	>	* needed for this operation to be executed in parallel
3745	>	* @param searchFunction a function returning a non-null
3746	>	* result on success, else null
3747	>	* @return a non-null result from applying the given search
3748	>	* function on each key, or null if none
3749	>	* @since 1.8
3750	>	*/
3751	>	public <U> U searchKeys(long parallelismThreshold,
3752	>	Fun<? super K, ? extends U> searchFunction) {
3753	>	if (searchFunction == null) throw new NullPointerException();
3754	>	return new SearchKeysTask<K,V,U>
3755	>	(null, batchFor(parallelismThreshold), 0, 0, table,
3756	>	searchFunction, new AtomicReference<U>()).invoke();
3757	>	}
3758	>
3759	>	/**
3760	>	* Returns the result of accumulating all keys using the given
3761	>	* reducer to combine values, or null if none.
3762	>	*
3763	>	* @param parallelismThreshold the (estimated) number of elements
3764	>	* needed for this operation to be executed in parallel
3765	>	* @param reducer a commutative associative combining function
3766	>	* @return the result of accumulating all keys using the given
3767	>	* reducer to combine values, or null if none
3768	>	* @since 1.8
3769	>	*/
3770	>	public K reduceKeys(long parallelismThreshold,
3771	>	BiFun<? super K, ? super K, ? extends K> reducer) {
3772	>	if (reducer == null) throw new NullPointerException();
3773	>	return new ReduceKeysTask<K,V>
3774	>	(null, batchFor(parallelismThreshold), 0, 0, table,
3775	>	null, reducer).invoke();
3776	>	}
3777	>
3778	>	/**
3779	>	* Returns the result of accumulating the given transformation
3780	>	* of all keys using the given reducer to combine values, or
3781	>	* null if none.
3782	>	*
3783	>	* @param parallelismThreshold the (estimated) number of elements
3784	>	* needed for this operation to be executed in parallel
3785	>	* @param transformer a function returning the transformation
3786	>	* for an element, or null if there is no transformation (in
3787	>	* which case it is not combined)
3788	>	* @param reducer a commutative associative combining function
3789	>	* @return the result of accumulating the given transformation
3790	>	* of all keys
3791	>	* @since 1.8
3792	>	*/
3793	>	public <U> U reduceKeys(long parallelismThreshold,
3794	>	Fun<? super K, ? extends U> transformer,
3795	>	BiFun<? super U, ? super U, ? extends U> reducer) {
3796	>	if (transformer == null || reducer == null)
3797		throw new NullPointerException();
3798	<	return (V)internalPut(key, value, false);
3798	>	return new MapReduceKeysTask<K,V,U>
3799	>	(null, batchFor(parallelismThreshold), 0, 0, table,
3800	>	null, transformer, reducer).invoke();
3801		}
3802
3803		/**
3804	<	* Copies all of the mappings from the specified map to this one.
3805	<	* These mappings replace any mappings that this map had for any of the
3806	<	* keys currently in the specified map.
3807	<	*
3808	<	* @param m mappings to be stored in this map
3809	<	*/
3810	<	public void putAll(Map<? extends K, ? extends V> m) {
3811	<	if (m == null)
3804	>	* Returns the result of accumulating the given transformation
3805	>	* of all keys using the given reducer to combine values, and
3806	>	* the given basis as an identity value.
3807	>	*
3808	>	* @param parallelismThreshold the (estimated) number of elements
3809	>	* needed for this operation to be executed in parallel
3810	>	* @param transformer a function returning the transformation
3811	>	* for an element
3812	>	* @param basis the identity (initial default value) for the reduction
3813	>	* @param reducer a commutative associative combining function
3814	>	* @return the result of accumulating the given transformation
3815	>	* of all keys
3816	>	* @since 1.8
3817	>	*/
3818	>	public double reduceKeysToDouble(long parallelismThreshold,
3819	>	ObjectToDouble<? super K> transformer,
3820	>	double basis,
3821	>	DoubleByDoubleToDouble reducer) {
3822	>	if (transformer == null || reducer == null)
3823		throw new NullPointerException();
3824	<	internalPutAll(m);
3824	>	return new MapReduceKeysToDoubleTask<K,V>
3825	>	(null, batchFor(parallelismThreshold), 0, 0, table,
3826	>	null, transformer, basis, reducer).invoke();
3827		}
3828
3829		/**
3830	<	* If the specified key is not already associated with a value,
3831	<	* computes its value using the given mappingFunction, and if
3832	<	* non-null, enters it into the map.  This is equivalent to
3833	<	*
3834	<	* <pre>
3835	<	*   if (map.containsKey(key))
3836	<	*       return map.get(key);
3837	<	*   value = mappingFunction.map(key);
3838	<	*   if (value != null)
3839	<	*      map.put(key, value);
3840	<	*   return value;
3841	<	* </pre>
3842	<	*
3843	<	* except that the action is performed atomically.  Some attempted
3844	<	* update operations on this map by other threads may be blocked
3845	<	* while computation is in progress, so the computation should be
3846	<	* short and simple, and must not attempt to update any other
3847	<	* mappings of this Map. The most appropriate usage is to
3848	<	* construct a new object serving as an initial mapped value, or
1124	<	* memoized result, as in:
1125	<	* <pre>{@code
1126	<	* map.computeIfAbsent(key, new MappingFunction<K, V>() {
1127	<	*   public V map(K k) { return new Value(f(k)); }};
1128	<	* }</pre>
1129	<	*
1130	<	* @param key key with which the specified value is to be associated
1131	<	* @param mappingFunction the function to compute a value
1132	<	* @return the current (existing or computed) value associated with
1133	<	*         the specified key, or {@code null} if the computation
1134	<	*         returned {@code null}.
1135	<	* @throws NullPointerException if the specified key or mappingFunction
1136	<	*         is null,
1137	<	* @throws IllegalStateException if the computation detectably
1138	<	*         attempts a recursive update to this map that would
1139	<	*         otherwise never complete.
1140	<	* @throws RuntimeException or Error if the mappingFunction does so,
1141	<	*         in which case the mapping is left unestablished.
1142	<	*/
1143	<	public V computeIfAbsent(K key, MappingFunction<? super K, ? extends V> mappingFunction) {
1144	<	if (key == null || mappingFunction == null)
3830	>	* Returns the result of accumulating the given transformation
3831	>	* of all keys using the given reducer to combine values, and
3832	>	* the given basis as an identity value.
3833	>	*
3834	>	* @param parallelismThreshold the (estimated) number of elements
3835	>	* needed for this operation to be executed in parallel
3836	>	* @param transformer a function returning the transformation
3837	>	* for an element
3838	>	* @param basis the identity (initial default value) for the reduction
3839	>	* @param reducer a commutative associative combining function
3840	>	* @return the result of accumulating the given transformation
3841	>	* of all keys
3842	>	* @since 1.8
3843	>	*/
3844	>	public long reduceKeysToLong(long parallelismThreshold,
3845	>	ObjectToLong<? super K> transformer,
3846	>	long basis,
3847	>	LongByLongToLong reducer) {
3848	>	if (transformer == null || reducer == null)
3849		throw new NullPointerException();
3850	<	return internalCompute(key, mappingFunction, false);
3850	>	return new MapReduceKeysToLongTask<K,V>
3851	>	(null, batchFor(parallelismThreshold), 0, 0, table,
3852	>	null, transformer, basis, reducer).invoke();
3853		}
3854
3855		/**
3856	<	* Computes the value associated with the given key using the given
3857	<	* mappingFunction, and if non-null, enters it into the map.  This
3858	<	* is equivalent to
3859	<	*
3860	<	* <pre>
3861	<	*   value = mappingFunction.map(key);
3862	<	*   if (value != null)
3863	<	*      map.put(key, value);
3864	<	*   else
3865	<	*      value = map.get(key);
3866	<	*   return value;
3867	<	* </pre>
3868	<	*
3869	<	* except that the action is performed atomically.  Some attempted
3870	<	* update operations on this map by other threads may be blocked
3871	<	* while computation is in progress, so the computation should be
3872	<	* short and simple, and must not attempt to update any other
3873	<	* mappings of this Map.
3874	<	*
1169	<	* @param key key with which the specified value is to be associated
1170	<	* @param mappingFunction the function to compute a value
1171	<	* @return the current value associated with
1172	<	*         the specified key, or {@code null} if the computation
1173	<	*         returned {@code null} and the value was not otherwise present.
1174	<	* @throws NullPointerException if the specified key or mappingFunction
1175	<	*         is null,
1176	<	* @throws IllegalStateException if the computation detectably
1177	<	*         attempts a recursive update to this map that would
1178	<	*         otherwise never complete.
1179	<	* @throws RuntimeException or Error if the mappingFunction does so,
1180	<	*         in which case the mapping is unchanged.
1181	<	*/
1182	<	public V compute(K key, MappingFunction<? super K, ? extends V> mappingFunction) {
1183	<	if (key == null || mappingFunction == null)
3856	>	* Returns the result of accumulating the given transformation
3857	>	* of all keys using the given reducer to combine values, and
3858	>	* the given basis as an identity value.
3859	>	*
3860	>	* @param parallelismThreshold the (estimated) number of elements
3861	>	* needed for this operation to be executed in parallel
3862	>	* @param transformer a function returning the transformation
3863	>	* for an element
3864	>	* @param basis the identity (initial default value) for the reduction
3865	>	* @param reducer a commutative associative combining function
3866	>	* @return the result of accumulating the given transformation
3867	>	* of all keys
3868	>	* @since 1.8
3869	>	*/
3870	>	public int reduceKeysToInt(long parallelismThreshold,
3871	>	ObjectToInt<? super K> transformer,
3872	>	int basis,
3873	>	IntByIntToInt reducer) {
3874	>	if (transformer == null || reducer == null)
3875		throw new NullPointerException();
3876	<	return internalCompute(key, mappingFunction, true);
3876	>	return new MapReduceKeysToIntTask<K,V>
3877	>	(null, batchFor(parallelismThreshold), 0, 0, table,
3878	>	null, transformer, basis, reducer).invoke();
3879		}
3880
3881		/**
3882	<	* Removes the key (and its corresponding value) from this map.
1190	<	* This method does nothing if the key is not in the map.
3882	>	* Performs the given action for each value.
3883		*
3884	<	* @param  key the key that needs to be removed
3885	<	* @return the previous value associated with {@code key}, or
3886	<	*         {@code null} if there was no mapping for {@code key}
3887	<	* @throws NullPointerException if the specified key is null
3884	>	* @param parallelismThreshold the (estimated) number of elements
3885	>	* needed for this operation to be executed in parallel
3886	>	* @param action the action
3887	>	* @since 1.8
3888		*/
3889	<	@SuppressWarnings("unchecked")
3890	<	public V remove(Object key) {
3891	<	if (key == null)
3889	>	public void forEachValue(long parallelismThreshold,
3890	>	Action<? super V> action) {
3891	>	if (action == null)
3892		throw new NullPointerException();
3893	<	return (V)internalReplace(key, null, null);
3893	>	new ForEachValueTask<K,V>
3894	>	(null, batchFor(parallelismThreshold), 0, 0, table,
3895	>	action).invoke();
3896		}
3897
3898		/**
3899	<	* {@inheritDoc}
3900	<	*
3901	<	* @throws NullPointerException if the specified key is null
3902	<	*/
3903	<	public boolean remove(Object key, Object value) {
3904	<	if (key == null)
3899	>	* Performs the given action for each non-null transformation
3900	>	* of each value.
3901	>	*
3902	>	* @param parallelismThreshold the (estimated) number of elements
3903	>	* needed for this operation to be executed in parallel
3904	>	* @param transformer a function returning the transformation
3905	>	* for an element, or null if there is no transformation (in
3906	>	* which case the action is not applied)
3907	>	* @param action the action
3908	>	* @since 1.8
3909	>	*/
3910	>	public <U> void forEachValue(long parallelismThreshold,
3911	>	Fun<? super V, ? extends U> transformer,
3912	>	Action<? super U> action) {
3913	>	if (transformer == null || action == null)
3914		throw new NullPointerException();
3915	<	if (value == null)
3916	<	return false;
3917	<	return internalReplace(key, null, value) != null;
3915	>	new ForEachTransformedValueTask<K,V,U>
3916	>	(null, batchFor(parallelismThreshold), 0, 0, table,
3917	>	transformer, action).invoke();
3918	>	}
3919	>
3920	>	/**
3921	>	* Returns a non-null result from applying the given search
3922	>	* function on each value, or null if none.  Upon success,
3923	>	* further element processing is suppressed and the results of
3924	>	* any other parallel invocations of the search function are
3925	>	* ignored.
3926	>	*
3927	>	* @param parallelismThreshold the (estimated) number of elements
3928	>	* needed for this operation to be executed in parallel
3929	>	* @param searchFunction a function returning a non-null
3930	>	* result on success, else null
3931	>	* @return a non-null result from applying the given search
3932	>	* function on each value, or null if none
3933	>	* @since 1.8
3934	>	*/
3935	>	public <U> U searchValues(long parallelismThreshold,
3936	>	Fun<? super V, ? extends U> searchFunction) {
3937	>	if (searchFunction == null) throw new NullPointerException();
3938	>	return new SearchValuesTask<K,V,U>
3939	>	(null, batchFor(parallelismThreshold), 0, 0, table,
3940	>	searchFunction, new AtomicReference<U>()).invoke();
3941	>	}
3942	>
3943	>	/**
3944	>	* Returns the result of accumulating all values using the
3945	>	* given reducer to combine values, or null if none.
3946	>	*
3947	>	* @param parallelismThreshold the (estimated) number of elements
3948	>	* needed for this operation to be executed in parallel
3949	>	* @param reducer a commutative associative combining function
3950	>	* @return the result of accumulating all values
3951	>	* @since 1.8
3952	>	*/
3953	>	public V reduceValues(long parallelismThreshold,
3954	>	BiFun<? super V, ? super V, ? extends V> reducer) {
3955	>	if (reducer == null) throw new NullPointerException();
3956	>	return new ReduceValuesTask<K,V>
3957	>	(null, batchFor(parallelismThreshold), 0, 0, table,
3958	>	null, reducer).invoke();
3959	>	}
3960	>
3961	>	/**
3962	>	* Returns the result of accumulating the given transformation
3963	>	* of all values using the given reducer to combine values, or
3964	>	* null if none.
3965	>	*
3966	>	* @param parallelismThreshold the (estimated) number of elements
3967	>	* needed for this operation to be executed in parallel
3968	>	* @param transformer a function returning the transformation
3969	>	* for an element, or null if there is no transformation (in
3970	>	* which case it is not combined)
3971	>	* @param reducer a commutative associative combining function
3972	>	* @return the result of accumulating the given transformation
3973	>	* of all values
3974	>	* @since 1.8
3975	>	*/
3976	>	public <U> U reduceValues(long parallelismThreshold,
3977	>	Fun<? super V, ? extends U> transformer,
3978	>	BiFun<? super U, ? super U, ? extends U> reducer) {
3979	>	if (transformer == null || reducer == null)
3980	>	throw new NullPointerException();
3981	>	return new MapReduceValuesTask<K,V,U>
3982	>	(null, batchFor(parallelismThreshold), 0, 0, table,
3983	>	null, transformer, reducer).invoke();
3984		}
3985
3986		/**
3987	<	* {@inheritDoc}
3988	<	*
3989	<	* @throws NullPointerException if any of the arguments are null
3990	<	*/
3991	<	public boolean replace(K key, V oldValue, V newValue) {
3992	<	if (key == null || oldValue == null || newValue == null)
3987	>	* Returns the result of accumulating the given transformation
3988	>	* of all values using the given reducer to combine values,
3989	>	* and the given basis as an identity value.
3990	>	*
3991	>	* @param parallelismThreshold the (estimated) number of elements
3992	>	* needed for this operation to be executed in parallel
3993	>	* @param transformer a function returning the transformation
3994	>	* for an element
3995	>	* @param basis the identity (initial default value) for the reduction
3996	>	* @param reducer a commutative associative combining function
3997	>	* @return the result of accumulating the given transformation
3998	>	* of all values
3999	>	* @since 1.8
4000	>	*/
4001	>	public double reduceValuesToDouble(long parallelismThreshold,
4002	>	ObjectToDouble<? super V> transformer,
4003	>	double basis,
4004	>	DoubleByDoubleToDouble reducer) {
4005	>	if (transformer == null || reducer == null)
4006		throw new NullPointerException();
4007	<	return internalReplace(key, newValue, oldValue) != null;
4007	>	return new MapReduceValuesToDoubleTask<K,V>
4008	>	(null, batchFor(parallelismThreshold), 0, 0, table,
4009	>	null, transformer, basis, reducer).invoke();
4010		}
4011
4012		/**
4013	<	* {@inheritDoc}
4014	<	*
4015	<	* @return the previous value associated with the specified key,
4016	<	*         or {@code null} if there was no mapping for the key
4017	<	* @throws NullPointerException if the specified key or value is null
4018	<	*/
4019	<	@SuppressWarnings("unchecked")
4020	<	public V replace(K key, V value) {
4021	<	if (key == null || value == null)
4013	>	* Returns the result of accumulating the given transformation
4014	>	* of all values using the given reducer to combine values,
4015	>	* and the given basis as an identity value.
4016	>	*
4017	>	* @param parallelismThreshold the (estimated) number of elements
4018	>	* needed for this operation to be executed in parallel
4019	>	* @param transformer a function returning the transformation
4020	>	* for an element
4021	>	* @param basis the identity (initial default value) for the reduction
4022	>	* @param reducer a commutative associative combining function
4023	>	* @return the result of accumulating the given transformation
4024	>	* of all values
4025	>	* @since 1.8
4026	>	*/
4027	>	public long reduceValuesToLong(long parallelismThreshold,
4028	>	ObjectToLong<? super V> transformer,
4029	>	long basis,
4030	>	LongByLongToLong reducer) {
4031	>	if (transformer == null || reducer == null)
4032		throw new NullPointerException();
4033	<	return (V)internalReplace(key, value, null);
4033	>	return new MapReduceValuesToLongTask<K,V>
4034	>	(null, batchFor(parallelismThreshold), 0, 0, table,
4035	>	null, transformer, basis, reducer).invoke();
4036		}
4037
4038		/**
4039	<	* Removes all of the mappings from this map.
4040	<	*/
4041	<	public void clear() {
4042	<	internalClear();
4039	>	* Returns the result of accumulating the given transformation
4040	>	* of all values using the given reducer to combine values,
4041	>	* and the given basis as an identity value.
4042	>	*
4043	>	* @param parallelismThreshold the (estimated) number of elements
4044	>	* needed for this operation to be executed in parallel
4045	>	* @param transformer a function returning the transformation
4046	>	* for an element
4047	>	* @param basis the identity (initial default value) for the reduction
4048	>	* @param reducer a commutative associative combining function
4049	>	* @return the result of accumulating the given transformation
4050	>	* of all values
4051	>	* @since 1.8
4052	>	*/
4053	>	public int reduceValuesToInt(long parallelismThreshold,
4054	>	ObjectToInt<? super V> transformer,
4055	>	int basis,
4056	>	IntByIntToInt reducer) {
4057	>	if (transformer == null || reducer == null)
4058	>	throw new NullPointerException();
4059	>	return new MapReduceValuesToIntTask<K,V>
4060	>	(null, batchFor(parallelismThreshold), 0, 0, table,
4061	>	null, transformer, basis, reducer).invoke();
4062		}
4063
4064		/**
4065	<	* Returns a {@link Set} view of the keys contained in this map.
1251	<	* The set is backed by the map, so changes to the map are
1252	<	* reflected in the set, and vice-versa.  The set supports element
1253	<	* removal, which removes the corresponding mapping from this map,
1254	<	* via the {@code Iterator.remove}, {@code Set.remove},
1255	<	* {@code removeAll}, {@code retainAll}, and {@code clear}
1256	<	* operations.  It does not support the {@code add} or
1257	<	* {@code addAll} operations.
4065	>	* Performs the given action for each entry.
4066		*
4067	<	* <p>The view's {@code iterator} is a "weakly consistent" iterator
4068	<	* that will never throw {@link ConcurrentModificationException},
4069	<	* and guarantees to traverse elements as they existed upon
4070	<	* construction of the iterator, and may (but is not guaranteed to)
1263	<	* reflect any modifications subsequent to construction.
4067	>	* @param parallelismThreshold the (estimated) number of elements
4068	>	* needed for this operation to be executed in parallel
4069	>	* @param action the action
4070	>	* @since 1.8
4071		*/
4072	<	public Set<K> keySet() {
4073	<	Set<K> ks = keySet;
4074	<	return (ks != null) ? ks : (keySet = new KeySet());
4072	>	public void forEachEntry(long parallelismThreshold,
4073	>	Action<? super Map.Entry<K,V>> action) {
4074	>	if (action == null) throw new NullPointerException();
4075	>	new ForEachEntryTask<K,V>(null, batchFor(parallelismThreshold), 0, 0, table,
4076	>	action).invoke();
4077		}
4078
4079		/**
4080	<	* Returns a {@link Collection} view of the values contained in this map.
4081	<	* The collection is backed by the map, so changes to the map are
1273	<	* reflected in the collection, and vice-versa.  The collection
1274	<	* supports element removal, which removes the corresponding
1275	<	* mapping from this map, via the {@code Iterator.remove},
1276	<	* {@code Collection.remove}, {@code removeAll},
1277	<	* {@code retainAll}, and {@code clear} operations.  It does not
1278	<	* support the {@code add} or {@code addAll} operations.
4080	>	* Performs the given action for each non-null transformation
4081	>	* of each entry.
4082		*
4083	<	* <p>The view's {@code iterator} is a "weakly consistent" iterator
4084	<	* that will never throw {@link ConcurrentModificationException},
4085	<	* and guarantees to traverse elements as they existed upon
4086	<	* construction of the iterator, and may (but is not guaranteed to)
4087	<	* reflect any modifications subsequent to construction.
4083	>	* @param parallelismThreshold the (estimated) number of elements
4084	>	* needed for this operation to be executed in parallel
4085	>	* @param transformer a function returning the transformation
4086	>	* for an element, or null if there is no transformation (in
4087	>	* which case the action is not applied)
4088	>	* @param action the action
4089	>	* @since 1.8
4090		*/
4091	<	public Collection<V> values() {
4092	<	Collection<V> vs = values;
4093	<	return (vs != null) ? vs : (values = new Values());
4091	>	public <U> void forEachEntry(long parallelismThreshold,
4092	>	Fun<Map.Entry<K,V>, ? extends U> transformer,
4093	>	Action<? super U> action) {
4094	>	if (transformer == null || action == null)
4095	>	throw new NullPointerException();
4096	>	new ForEachTransformedEntryTask<K,V,U>
4097	>	(null, batchFor(parallelismThreshold), 0, 0, table,
4098	>	transformer, action).invoke();
4099	>	}
4100	>
4101	>	/**
4102	>	* Returns a non-null result from applying the given search
4103	>	* function on each entry, or null if none.  Upon success,
4104	>	* further element processing is suppressed and the results of
4105	>	* any other parallel invocations of the search function are
4106	>	* ignored.
4107	>	*
4108	>	* @param parallelismThreshold the (estimated) number of elements
4109	>	* needed for this operation to be executed in parallel
4110	>	* @param searchFunction a function returning a non-null
4111	>	* result on success, else null
4112	>	* @return a non-null result from applying the given search
4113	>	* function on each entry, or null if none
4114	>	* @since 1.8
4115	>	*/
4116	>	public <U> U searchEntries(long parallelismThreshold,
4117	>	Fun<Map.Entry<K,V>, ? extends U> searchFunction) {
4118	>	if (searchFunction == null) throw new NullPointerException();
4119	>	return new SearchEntriesTask<K,V,U>
4120	>	(null, batchFor(parallelismThreshold), 0, 0, table,
4121	>	searchFunction, new AtomicReference<U>()).invoke();
4122	>	}
4123	>
4124	>	/**
4125	>	* Returns the result of accumulating all entries using the
4126	>	* given reducer to combine values, or null if none.
4127	>	*
4128	>	* @param parallelismThreshold the (estimated) number of elements
4129	>	* needed for this operation to be executed in parallel
4130	>	* @param reducer a commutative associative combining function
4131	>	* @return the result of accumulating all entries
4132	>	* @since 1.8
4133	>	*/
4134	>	public Map.Entry<K,V> reduceEntries(long parallelismThreshold,
4135	>	BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer) {
4136	>	if (reducer == null) throw new NullPointerException();
4137	>	return new ReduceEntriesTask<K,V>
4138	>	(null, batchFor(parallelismThreshold), 0, 0, table,
4139	>	null, reducer).invoke();
4140	>	}
4141	>
4142	>	/**
4143	>	* Returns the result of accumulating the given transformation
4144	>	* of all entries using the given reducer to combine values,
4145	>	* or null if none.
4146	>	*
4147	>	* @param parallelismThreshold the (estimated) number of elements
4148	>	* needed for this operation to be executed in parallel
4149	>	* @param transformer a function returning the transformation
4150	>	* for an element, or null if there is no transformation (in
4151	>	* which case it is not combined)
4152	>	* @param reducer a commutative associative combining function
4153	>	* @return the result of accumulating the given transformation
4154	>	* of all entries
4155	>	* @since 1.8
4156	>	*/
4157	>	public <U> U reduceEntries(long parallelismThreshold,
4158	>	Fun<Map.Entry<K,V>, ? extends U> transformer,
4159	>	BiFun<? super U, ? super U, ? extends U> reducer) {
4160	>	if (transformer == null || reducer == null)
4161	>	throw new NullPointerException();
4162	>	return new MapReduceEntriesTask<K,V,U>
4163	>	(null, batchFor(parallelismThreshold), 0, 0, table,
4164	>	null, transformer, reducer).invoke();
4165		}
4166
4167		/**
4168	<	* Returns a {@link Set} view of the mappings contained in this map.
4169	<	* The set is backed by the map, so changes to the map are
4170	<	* reflected in the set, and vice-versa.  The set supports element
4171	<	* removal, which removes the corresponding mapping from the map,
4172	<	* via the {@code Iterator.remove}, {@code Set.remove},
4173	<	* {@code removeAll}, {@code retainAll}, and {@code clear}
4174	<	* operations.  It does not support the {@code add} or
4175	<	* {@code addAll} operations.
4176	<	*
4177	<	* <p>The view's {@code iterator} is a "weakly consistent" iterator
4178	<	* that will never throw {@link ConcurrentModificationException},
4179	<	* and guarantees to traverse elements as they existed upon
4180	<	* construction of the iterator, and may (but is not guaranteed to)
4181	<	* reflect any modifications subsequent to construction.
4182	<	*/
4183	<	public Set<Map.Entry<K,V>> entrySet() {
4184	<	Set<Map.Entry<K,V>> es = entrySet;
4185	<	return (es != null) ? es : (entrySet = new EntrySet());
4168	>	* Returns the result of accumulating the given transformation
4169	>	* of all entries using the given reducer to combine values,
4170	>	* and the given basis as an identity value.
4171	>	*
4172	>	* @param parallelismThreshold the (estimated) number of elements
4173	>	* needed for this operation to be executed in parallel
4174	>	* @param transformer a function returning the transformation
4175	>	* for an element
4176	>	* @param basis the identity (initial default value) for the reduction
4177	>	* @param reducer a commutative associative combining function
4178	>	* @return the result of accumulating the given transformation
4179	>	* of all entries
4180	>	* @since 1.8
4181	>	*/
4182	>	public double reduceEntriesToDouble(long parallelismThreshold,
4183	>	ObjectToDouble<Map.Entry<K,V>> transformer,
4184	>	double basis,
4185	>	DoubleByDoubleToDouble reducer) {
4186	>	if (transformer == null || reducer == null)
4187	>	throw new NullPointerException();
4188	>	return new MapReduceEntriesToDoubleTask<K,V>
4189	>	(null, batchFor(parallelismThreshold), 0, 0, table,
4190	>	null, transformer, basis, reducer).invoke();
4191		}
4192
4193		/**
4194	<	* Returns an enumeration of the keys in this table.
4195	<	*
4196	<	* @return an enumeration of the keys in this table
4197	<	* @see #keySet()
4198	<	*/
4199	<	public Enumeration<K> keys() {
4200	<	return new KeyIterator();
4194	>	* Returns the result of accumulating the given transformation
4195	>	* of all entries using the given reducer to combine values,
4196	>	* and the given basis as an identity value.
4197	>	*
4198	>	* @param parallelismThreshold the (estimated) number of elements
4199	>	* needed for this operation to be executed in parallel
4200	>	* @param transformer a function returning the transformation
4201	>	* for an element
4202	>	* @param basis the identity (initial default value) for the reduction
4203	>	* @param reducer a commutative associative combining function
4204	>	* @return the result of accumulating the given transformation
4205	>	* of all entries
4206	>	* @since 1.8
4207	>	*/
4208	>	public long reduceEntriesToLong(long parallelismThreshold,
4209	>	ObjectToLong<Map.Entry<K,V>> transformer,
4210	>	long basis,
4211	>	LongByLongToLong reducer) {
4212	>	if (transformer == null || reducer == null)
4213	>	throw new NullPointerException();
4214	>	return new MapReduceEntriesToLongTask<K,V>
4215	>	(null, batchFor(parallelismThreshold), 0, 0, table,
4216	>	null, transformer, basis, reducer).invoke();
4217		}
4218
4219		/**
4220	<	* Returns an enumeration of the values in this table.
4221	<	*
4222	<	* @return an enumeration of the values in this table
4223	<	* @see #values()
4224	<	*/
4225	<	public Enumeration<V> elements() {
4226	<	return new ValueIterator();
4220	>	* Returns the result of accumulating the given transformation
4221	>	* of all entries using the given reducer to combine values,
4222	>	* and the given basis as an identity value.
4223	>	*
4224	>	* @param parallelismThreshold the (estimated) number of elements
4225	>	* needed for this operation to be executed in parallel
4226	>	* @param transformer a function returning the transformation
4227	>	* for an element
4228	>	* @param basis the identity (initial default value) for the reduction
4229	>	* @param reducer a commutative associative combining function
4230	>	* @return the result of accumulating the given transformation
4231	>	* of all entries
4232	>	* @since 1.8
4233	>	*/
4234	>	public int reduceEntriesToInt(long parallelismThreshold,
4235	>	ObjectToInt<Map.Entry<K,V>> transformer,
4236	>	int basis,
4237	>	IntByIntToInt reducer) {
4238	>	if (transformer == null || reducer == null)
4239	>	throw new NullPointerException();
4240	>	return new MapReduceEntriesToIntTask<K,V>
4241	>	(null, batchFor(parallelismThreshold), 0, 0, table,
4242	>	null, transformer, basis, reducer).invoke();
4243		}
4244
4245	+
4246	+	/* ----------------Views -------------- */
4247	+
4248		/**
4249	<	* Returns the hash code value for this {@link Map}, i.e.,
1334	<	* the sum of, for each key-value pair in the map,
1335	<	* {@code key.hashCode() ^ value.hashCode()}.
1336	<	*
1337	<	* @return the hash code value for this map
4249	>	* Base class for views.
4250		*/
4251	<	public int hashCode() {
4252	<	return new HashIterator().mapHashCode();
4251	>	abstract static class CollectionView<K,V,E>
4252	>	implements Collection<E>, java.io.Serializable {
4253	>	private static final long serialVersionUID = 7249069246763182397L;
4254	>	final ConcurrentHashMapV8<K,V> map;
4255	>	CollectionView(ConcurrentHashMapV8<K,V> map)  { this.map = map; }
4256	>
4257	>	/**
4258	>	* Returns the map backing this view.
4259	>	*
4260	>	* @return the map backing this view
4261	>	*/
4262	>	public ConcurrentHashMapV8<K,V> getMap() { return map; }
4263	>
4264	>	/**
4265	>	* Removes all of the elements from this view, by removing all
4266	>	* the mappings from the map backing this view.
4267	>	*/
4268	>	public final void clear()      { map.clear(); }
4269	>	public final int size()        { return map.size(); }
4270	>	public final boolean isEmpty() { return map.isEmpty(); }
4271	>
4272	>	// implementations below rely on concrete classes supplying these
4273	>	// abstract methods
4274	>	/**
4275	>	* Returns a "weakly consistent" iterator that will never
4276	>	* throw {@link ConcurrentModificationException}, and
4277	>	* guarantees to traverse elements as they existed upon
4278	>	* construction of the iterator, and may (but is not
4279	>	* guaranteed to) reflect any modifications subsequent to
4280	>	* construction.
4281	>	*/
4282	>	public abstract Iterator<E> iterator();
4283	>	public abstract boolean contains(Object o);
4284	>	public abstract boolean remove(Object o);
4285	>
4286	>	private static final String oomeMsg = "Required array size too large";
4287	>
4288	>	public final Object[] toArray() {
4289	>	long sz = map.mappingCount();
4290	>	if (sz > MAX_ARRAY_SIZE)
4291	>	throw new OutOfMemoryError(oomeMsg);
4292	>	int n = (int)sz;
4293	>	Object[] r = new Object[n];
4294	>	int i = 0;
4295	>	for (E e : this) {
4296	>	if (i == n) {
4297	>	if (n >= MAX_ARRAY_SIZE)
4298	>	throw new OutOfMemoryError(oomeMsg);
4299	>	if (n >= MAX_ARRAY_SIZE - (MAX_ARRAY_SIZE >>> 1) - 1)
4300	>	n = MAX_ARRAY_SIZE;
4301	>	else
4302	>	n += (n >>> 1) + 1;
4303	>	r = Arrays.copyOf(r, n);
4304	>	}
4305	>	r[i++] = e;
4306	>	}
4307	>	return (i == n) ? r : Arrays.copyOf(r, i);
4308	>	}
4309	>
4310	>	@SuppressWarnings("unchecked")
4311	>	public final <T> T[] toArray(T[] a) {
4312	>	long sz = map.mappingCount();
4313	>	if (sz > MAX_ARRAY_SIZE)
4314	>	throw new OutOfMemoryError(oomeMsg);
4315	>	int m = (int)sz;
4316	>	T[] r = (a.length >= m) ? a :
4317	>	(T[])java.lang.reflect.Array
4318	>	.newInstance(a.getClass().getComponentType(), m);
4319	>	int n = r.length;
4320	>	int i = 0;
4321	>	for (E e : this) {
4322	>	if (i == n) {
4323	>	if (n >= MAX_ARRAY_SIZE)
4324	>	throw new OutOfMemoryError(oomeMsg);
4325	>	if (n >= MAX_ARRAY_SIZE - (MAX_ARRAY_SIZE >>> 1) - 1)
4326	>	n = MAX_ARRAY_SIZE;
4327	>	else
4328	>	n += (n >>> 1) + 1;
4329	>	r = Arrays.copyOf(r, n);
4330	>	}
4331	>	r[i++] = (T)e;
4332	>	}
4333	>	if (a == r && i < n) {
4334	>	r[i] = null; // null-terminate
4335	>	return r;
4336	>	}
4337	>	return (i == n) ? r : Arrays.copyOf(r, i);
4338	>	}
4339	>
4340	>	/**
4341	>	* Returns a string representation of this collection.
4342	>	* The string representation consists of the string representations
4343	>	* of the collection's elements in the order they are returned by
4344	>	* its iterator, enclosed in square brackets ({@code "[]"}).
4345	>	* Adjacent elements are separated by the characters {@code ", "}
4346	>	* (comma and space).  Elements are converted to strings as by
4347	>	* {@link String#valueOf(Object)}.
4348	>	*
4349	>	* @return a string representation of this collection
4350	>	*/
4351	>	public final String toString() {
4352	>	StringBuilder sb = new StringBuilder();
4353	>	sb.append('[');
4354	>	Iterator<E> it = iterator();
4355	>	if (it.hasNext()) {
4356	>	for (;;) {
4357	>	Object e = it.next();
4358	>	sb.append(e == this ? "(this Collection)" : e);
4359	>	if (!it.hasNext())
4360	>	break;
4361	>	sb.append(',').append(' ');
4362	>	}
4363	>	}
4364	>	return sb.append(']').toString();
4365	>	}
4366	>
4367	>	public final boolean containsAll(Collection<?> c) {
4368	>	if (c != this) {
4369	>	for (Object e : c) {
4370	>	if (e == null || !contains(e))
4371	>	return false;
4372	>	}
4373	>	}
4374	>	return true;
4375	>	}
4376	>
4377	>	public final boolean removeAll(Collection<?> c) {
4378	>	boolean modified = false;
4379	>	for (Iterator<E> it = iterator(); it.hasNext();) {
4380	>	if (c.contains(it.next())) {
4381	>	it.remove();
4382	>	modified = true;
4383	>	}
4384	>	}
4385	>	return modified;
4386	>	}
4387	>
4388	>	public final boolean retainAll(Collection<?> c) {
4389	>	boolean modified = false;
4390	>	for (Iterator<E> it = iterator(); it.hasNext();) {
4391	>	if (!c.contains(it.next())) {
4392	>	it.remove();
4393	>	modified = true;
4394	>	}
4395	>	}
4396	>	return modified;
4397	>	}
4398	>
4399		}
4400
4401		/**
4402	<	* Returns a string representation of this map.  The string
4403	<	* representation consists of a list of key-value mappings (in no
4404	<	* particular order) enclosed in braces ("{@code {}}").  Adjacent
4405	<	* mappings are separated by the characters {@code ", "} (comma
4406	<	* and space).  Each key-value mapping is rendered as the key
4407	<	* followed by an equals sign ("{@code =}") followed by the
4408	<	* associated value.
4402	>	* A view of a ConcurrentHashMapV8 as a {@link Set} of keys, in
4403	>	* which additions may optionally be enabled by mapping to a
4404	>	* common value.  This class cannot be directly instantiated.
4405	>	* See {@link #keySet() keySet()},
4406	>	* {@link #keySet(Object) keySet(V)},
4407	>	* {@link #newKeySet() newKeySet()},
4408	>	* {@link #newKeySet(int) newKeySet(int)}.
4409		*
4410	<	* @return a string representation of this map
4410	>	* @since 1.8
4411		*/
4412	<	public String toString() {
4413	<	return new HashIterator().mapToString();
4412	>	public static class KeySetView<K,V> extends CollectionView<K,V,K>
4413	>	implements Set<K>, java.io.Serializable {
4414	>	private static final long serialVersionUID = 7249069246763182397L;
4415	>	private final V value;
4416	>	KeySetView(ConcurrentHashMapV8<K,V> map, V value) {  // non-public
4417	>	super(map);
4418	>	this.value = value;
4419	>	}
4420	>
4421	>	/**
4422	>	* Returns the default mapped value for additions,
4423	>	* or {@code null} if additions are not supported.
4424	>	*
4425	>	* @return the default mapped value for additions, or {@code null}
4426	>	* if not supported
4427	>	*/
4428	>	public V getMappedValue() { return value; }
4429	>
4430	>	/**
4431	>	* {@inheritDoc}
4432	>	* @throws NullPointerException if the specified key is null
4433	>	*/
4434	>	public boolean contains(Object o) { return map.containsKey(o); }
4435	>
4436	>	/**
4437	>	* Removes the key from this map view, by removing the key (and its
4438	>	* corresponding value) from the backing map.  This method does
4439	>	* nothing if the key is not in the map.
4440	>	*
4441	>	* @param  o the key to be removed from the backing map
4442	>	* @return {@code true} if the backing map contained the specified key
4443	>	* @throws NullPointerException if the specified key is null
4444	>	*/
4445	>	public boolean remove(Object o) { return map.remove(o) != null; }
4446	>
4447	>	/**
4448	>	* @return an iterator over the keys of the backing map
4449	>	*/
4450	>	public Iterator<K> iterator() {
4451	>	Node<K,V>[] t;
4452	>	ConcurrentHashMapV8<K,V> m = map;
4453	>	int f = (t = m.table) == null ? 0 : t.length;
4454	>	return new KeyIterator<K,V>(t, f, 0, f, m);
4455	>	}
4456	>
4457	>	/**
4458	>	* Adds the specified key to this set view by mapping the key to
4459	>	* the default mapped value in the backing map, if defined.
4460	>	*
4461	>	* @param e key to be added
4462	>	* @return {@code true} if this set changed as a result of the call
4463	>	* @throws NullPointerException if the specified key is null
4464	>	* @throws UnsupportedOperationException if no default mapped value
4465	>	* for additions was provided
4466	>	*/
4467	>	public boolean add(K e) {
4468	>	V v;
4469	>	if ((v = value) == null)
4470	>	throw new UnsupportedOperationException();
4471	>	return map.putVal(e, v, true) == null;
4472	>	}
4473	>
4474	>	/**
4475	>	* Adds all of the elements in the specified collection to this set,
4476	>	* as if by calling {@link #add} on each one.
4477	>	*
4478	>	* @param c the elements to be inserted into this set
4479	>	* @return {@code true} if this set changed as a result of the call
4480	>	* @throws NullPointerException if the collection or any of its
4481	>	* elements are {@code null}
4482	>	* @throws UnsupportedOperationException if no default mapped value
4483	>	* for additions was provided
4484	>	*/
4485	>	public boolean addAll(Collection<? extends K> c) {
4486	>	boolean added = false;
4487	>	V v;
4488	>	if ((v = value) == null)
4489	>	throw new UnsupportedOperationException();
4490	>	for (K e : c) {
4491	>	if (map.putVal(e, v, true) == null)
4492	>	added = true;
4493	>	}
4494	>	return added;
4495	>	}
4496	>
4497	>	public int hashCode() {
4498	>	int h = 0;
4499	>	for (K e : this)
4500	>	h += e.hashCode();
4501	>	return h;
4502	>	}
4503	>
4504	>	public boolean equals(Object o) {
4505	>	Set<?> c;
4506	>	return ((o instanceof Set) &&
4507	>	((c = (Set<?>)o) == this ||
4508	>	(containsAll(c) && c.containsAll(this))));
4509	>	}
4510	>
4511	>	public ConcurrentHashMapSpliterator<K> spliterator() {
4512	>	Node<K,V>[] t;
4513	>	ConcurrentHashMapV8<K,V> m = map;
4514	>	long n = m.sumCount();
4515	>	int f = (t = m.table) == null ? 0 : t.length;
4516	>	return new KeySpliterator<K,V>(t, f, 0, f, n < 0L ? 0L : n);
4517	>	}
4518	>
4519	>	public void forEach(Action<? super K> action) {
4520	>	if (action == null) throw new NullPointerException();
4521	>	Node<K,V>[] t;
4522	>	if ((t = map.table) != null) {
4523	>	Traverser<K,V> it = new Traverser<K,V>(t, t.length, 0, t.length);
4524	>	for (Node<K,V> p; (p = it.advance()) != null; )
4525	>	action.apply(p.key);
4526	>	}
4527	>	}
4528		}
4529
4530		/**
4531	<	* Compares the specified object with this map for equality.
4532	<	* Returns {@code true} if the given object is a map with the same
4533	<	* mappings as this map.  This operation may return misleading
1362	<	* results if either map is concurrently modified during execution
1363	<	* of this method.
1364	<	*
1365	<	* @param o object to be compared for equality with this map
1366	<	* @return {@code true} if the specified object is equal to this map
4531	>	* A view of a ConcurrentHashMapV8 as a {@link Collection} of
4532	>	* values, in which additions are disabled. This class cannot be
4533	>	* directly instantiated. See {@link #values()}.
4534		*/
4535	<	public boolean equals(Object o) {
4536	<	if (o == this)
4537	<	return true;
4538	<	if (!(o instanceof Map))
4539	<	return false;
4540	<	Map<?,?> m = (Map<?,?>) o;
4541	<	try {
4542	<	for (Map.Entry<K,V> e : this.entrySet())
4543	<	if (! e.getValue().equals(m.get(e.getKey())))
4544	<	return false;
4545	<	for (Map.Entry<?,?> e : m.entrySet()) {
4546	<	Object k = e.getKey();
4547	<	Object v = e.getValue();
4548	<	if (k == null || v == null || !v.equals(get(k)))
4549	<	return false;
4535	>	static final class ValuesView<K,V> extends CollectionView<K,V,V>
4536	>	implements Collection<V>, java.io.Serializable {
4537	>	private static final long serialVersionUID = 2249069246763182397L;
4538	>	ValuesView(ConcurrentHashMapV8<K,V> map) { super(map); }
4539	>	public final boolean contains(Object o) {
4540	>	return map.containsValue(o);
4541	>	}
4542	>
4543	>	public final boolean remove(Object o) {
4544	>	if (o != null) {
4545	>	for (Iterator<V> it = iterator(); it.hasNext();) {
4546	>	if (o.equals(it.next())) {
4547	>	it.remove();
4548	>	return true;
4549	>	}
4550	>	}
4551		}
1384	–	return true;
1385	–	} catch (ClassCastException unused) {
1386	–	return false;
1387	–	} catch (NullPointerException unused) {
4552		return false;
4553		}
4554	+
4555	+	public final Iterator<V> iterator() {
4556	+	ConcurrentHashMapV8<K,V> m = map;
4557	+	Node<K,V>[] t;
4558	+	int f = (t = m.table) == null ? 0 : t.length;
4559	+	return new ValueIterator<K,V>(t, f, 0, f, m);
4560	+	}
4561	+
4562	+	public final boolean add(V e) {
4563	+	throw new UnsupportedOperationException();
4564	+	}
4565	+	public final boolean addAll(Collection<? extends V> c) {
4566	+	throw new UnsupportedOperationException();
4567	+	}
4568	+
4569	+	public ConcurrentHashMapSpliterator<V> spliterator() {
4570	+	Node<K,V>[] t;
4571	+	ConcurrentHashMapV8<K,V> m = map;
4572	+	long n = m.sumCount();
4573	+	int f = (t = m.table) == null ? 0 : t.length;
4574	+	return new ValueSpliterator<K,V>(t, f, 0, f, n < 0L ? 0L : n);
4575	+	}
4576	+
4577	+	public void forEach(Action<? super V> action) {
4578	+	if (action == null) throw new NullPointerException();
4579	+	Node<K,V>[] t;
4580	+	if ((t = map.table) != null) {
4581	+	Traverser<K,V> it = new Traverser<K,V>(t, t.length, 0, t.length);
4582	+	for (Node<K,V> p; (p = it.advance()) != null; )
4583	+	action.apply(p.val);
4584	+	}
4585	+	}
4586		}
4587
4588		/**
4589	<	* Custom Entry class used by EntryIterator.next(), that relays
4590	<	* setValue changes to the underlying map.
4589	>	* A view of a ConcurrentHashMapV8 as a {@link Set} of (key, value)
4590	>	* entries.  This class cannot be directly instantiated. See
4591	>	* {@link #entrySet()}.
4592		*/
4593	<	final class WriteThroughEntry extends AbstractMap.SimpleEntry<K,V> {
4594	<	@SuppressWarnings("unchecked")
4595	<	WriteThroughEntry(Object k, Object v) {
4596	<	super((K)k, (V)v);
4593	>	static final class EntrySetView<K,V> extends CollectionView<K,V,Map.Entry<K,V>>
4594	>	implements Set<Map.Entry<K,V>>, java.io.Serializable {
4595	>	private static final long serialVersionUID = 2249069246763182397L;
4596	>	EntrySetView(ConcurrentHashMapV8<K,V> map) { super(map); }
4597	>
4598	>	public boolean contains(Object o) {
4599	>	Object k, v, r; Map.Entry<?,?> e;
4600	>	return ((o instanceof Map.Entry) &&
4601	>	(k = (e = (Map.Entry<?,?>)o).getKey()) != null &&
4602	>	(r = map.get(k)) != null &&
4603	>	(v = e.getValue()) != null &&
4604	>	(v == r || v.equals(r)));
4605	>	}
4606	>
4607	>	public boolean remove(Object o) {
4608	>	Object k, v; Map.Entry<?,?> e;
4609	>	return ((o instanceof Map.Entry) &&
4610	>	(k = (e = (Map.Entry<?,?>)o).getKey()) != null &&
4611	>	(v = e.getValue()) != null &&
4612	>	map.remove(k, v));
4613		}
4614
4615		/**
4616	<	* Sets our entry's value and writes through to the map. The
1404	<	* value to return is somewhat arbitrary here. Since a
1405	<	* WriteThroughEntry does not necessarily track asynchronous
1406	<	* changes, the most recent "previous" value could be
1407	<	* different from what we return (or could even have been
1408	<	* removed in which case the put will re-establish). We do not
1409	<	* and cannot guarantee more.
4616	>	* @return an iterator over the entries of the backing map
4617		*/
4618	<	public V setValue(V value) {
4619	<	if (value == null) throw new NullPointerException();
4620	<	V v = super.setValue(value);
4621	<	ConcurrentHashMapV8.this.put(getKey(), value);
4622	<	return v;
4618	>	public Iterator<Map.Entry<K,V>> iterator() {
4619	>	ConcurrentHashMapV8<K,V> m = map;
4620	>	Node<K,V>[] t;
4621	>	int f = (t = m.table) == null ? 0 : t.length;
4622	>	return new EntryIterator<K,V>(t, f, 0, f, m);
4623	>	}
4624	>
4625	>	public boolean add(Entry<K,V> e) {
4626	>	return map.putVal(e.getKey(), e.getValue(), false) == null;
4627	>	}
4628	>
4629	>	public boolean addAll(Collection<? extends Entry<K,V>> c) {
4630	>	boolean added = false;
4631	>	for (Entry<K,V> e : c) {
4632	>	if (add(e))
4633	>	added = true;
4634	>	}
4635	>	return added;
4636		}
4637	+
4638	+	public final int hashCode() {
4639	+	int h = 0;
4640	+	Node<K,V>[] t;
4641	+	if ((t = map.table) != null) {
4642	+	Traverser<K,V> it = new Traverser<K,V>(t, t.length, 0, t.length);
4643	+	for (Node<K,V> p; (p = it.advance()) != null; ) {
4644	+	h += p.hashCode();
4645	+	}
4646	+	}
4647	+	return h;
4648	+	}
4649	+
4650	+	public final boolean equals(Object o) {
4651	+	Set<?> c;
4652	+	return ((o instanceof Set) &&
4653	+	((c = (Set<?>)o) == this ||
4654	+	(containsAll(c) && c.containsAll(this))));
4655	+	}
4656	+
4657	+	public ConcurrentHashMapSpliterator<Map.Entry<K,V>> spliterator() {
4658	+	Node<K,V>[] t;
4659	+	ConcurrentHashMapV8<K,V> m = map;
4660	+	long n = m.sumCount();
4661	+	int f = (t = m.table) == null ? 0 : t.length;
4662	+	return new EntrySpliterator<K,V>(t, f, 0, f, n < 0L ? 0L : n, m);
4663	+	}
4664	+
4665	+	public void forEach(Action<? super Map.Entry<K,V>> action) {
4666	+	if (action == null) throw new NullPointerException();
4667	+	Node<K,V>[] t;
4668	+	if ((t = map.table) != null) {
4669	+	Traverser<K,V> it = new Traverser<K,V>(t, t.length, 0, t.length);
4670	+	for (Node<K,V> p; (p = it.advance()) != null; )
4671	+	action.apply(new MapEntry<K,V>(p.key, p.val, map));
4672	+	}
4673	+	}
4674	+
4675		}
4676
4677	<	final class KeyIterator extends HashIterator
4678	<	implements Iterator<K>, Enumeration<K> {
4679	<	@SuppressWarnings("unchecked")
4680	<	public final K next()        { return (K)super.nextKey(); }
4681	<	@SuppressWarnings("unchecked")
4682	<	public final K nextElement() { return (K)super.nextKey(); }
4677	>	// -------------------------------------------------------
4678	>
4679	>	/**
4680	>	* Base class for bulk tasks. Repeats some fields and code from
4681	>	* class Traverser, because we need to subclass CountedCompleter.
4682	>	*/
4683	>	abstract static class BulkTask<K,V,R> extends CountedCompleter<R> {
4684	>	Node<K,V>[] tab;        // same as Traverser
4685	>	Node<K,V> next;
4686	>	int index;
4687	>	int baseIndex;
4688	>	int baseLimit;
4689	>	final int baseSize;
4690	>	int batch;              // split control
4691	>
4692	>	BulkTask(BulkTask<K,V,?> par, int b, int i, int f, Node<K,V>[] t) {
4693	>	super(par);
4694	>	this.batch = b;
4695	>	this.index = this.baseIndex = i;
4696	>	if ((this.tab = t) == null)
4697	>	this.baseSize = this.baseLimit = 0;
4698	>	else if (par == null)
4699	>	this.baseSize = this.baseLimit = t.length;
4700	>	else {
4701	>	this.baseLimit = f;
4702	>	this.baseSize = par.baseSize;
4703	>	}
4704	>	}
4705	>
4706	>	/**
4707	>	* Same as Traverser version
4708	>	*/
4709	>	final Node<K,V> advance() {
4710	>	Node<K,V> e;
4711	>	if ((e = next) != null)
4712	>	e = e.next;
4713	>	for (;;) {
4714	>	Node<K,V>[] t; int i, n; K ek;  // must use locals in checks
4715	>	if (e != null)
4716	>	return next = e;
4717	>	if (baseIndex >= baseLimit || (t = tab) == null ||
4718	>	(n = t.length) <= (i = index) || i < 0)
4719	>	return next = null;
4720	>	if ((e = tabAt(t, index)) != null && e.hash < 0) {
4721	>	if (e instanceof ForwardingNode) {
4722	>	tab = ((ForwardingNode<K,V>)e).nextTable;
4723	>	e = null;
4724	>	continue;
4725	>	}
4726	>	else if (e instanceof TreeBin)
4727	>	e = ((TreeBin<K,V>)e).first;
4728	>	else
4729	>	e = null;
4730	>	}
4731	>	if ((index += baseSize) >= n)
4732	>	index = ++baseIndex;    // visit upper slots if present
4733	>	}
4734	>	}
4735		}
4736
4737	<	final class ValueIterator extends HashIterator
4738	<	implements Iterator<V>, Enumeration<V> {
4739	<	@SuppressWarnings("unchecked")
4740	<	public final V next()        { return (V)super.nextValue(); }
4741	<	@SuppressWarnings("unchecked")
4742	<	public final V nextElement() { return (V)super.nextValue(); }
4737	>	/*
4738	>	* Task classes. Coded in a regular but ugly format/style to
4739	>	* simplify checks that each variant differs in the right way from
4740	>	* others. The null screenings exist because compilers cannot tell
4741	>	* that we've already null-checked task arguments, so we force
4742	>	* simplest hoisted bypass to help avoid convoluted traps.
4743	>	*/
4744	>	@SuppressWarnings("serial")
4745	>	static final class ForEachKeyTask<K,V>
4746	>	extends BulkTask<K,V,Void> {
4747	>	final Action<? super K> action;
4748	>	ForEachKeyTask
4749	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
4750	>	Action<? super K> action) {
4751	>	super(p, b, i, f, t);
4752	>	this.action = action;
4753	>	}
4754	>	public final void compute() {
4755	>	final Action<? super K> action;
4756	>	if ((action = this.action) != null) {
4757	>	for (int i = baseIndex, f, h; batch > 0 &&
4758	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
4759	>	addToPendingCount(1);
4760	>	new ForEachKeyTask<K,V>
4761	>	(this, batch >>>= 1, baseLimit = h, f, tab,
4762	>	action).fork();
4763	>	}
4764	>	for (Node<K,V> p; (p = advance()) != null;)
4765	>	action.apply(p.key);
4766	>	propagateCompletion();
4767	>	}
4768	>	}
4769		}
4770
4771	<	final class EntryIterator extends HashIterator
4772	<	implements Iterator<Entry<K,V>> {
4773	<	public final Map.Entry<K,V> next() { return super.nextEntry(); }
4771	>	@SuppressWarnings("serial")
4772	>	static final class ForEachValueTask<K,V>
4773	>	extends BulkTask<K,V,Void> {
4774	>	final Action<? super V> action;
4775	>	ForEachValueTask
4776	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
4777	>	Action<? super V> action) {
4778	>	super(p, b, i, f, t);
4779	>	this.action = action;
4780	>	}
4781	>	public final void compute() {
4782	>	final Action<? super V> action;
4783	>	if ((action = this.action) != null) {
4784	>	for (int i = baseIndex, f, h; batch > 0 &&
4785	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
4786	>	addToPendingCount(1);
4787	>	new ForEachValueTask<K,V>
4788	>	(this, batch >>>= 1, baseLimit = h, f, tab,
4789	>	action).fork();
4790	>	}
4791	>	for (Node<K,V> p; (p = advance()) != null;)
4792	>	action.apply(p.val);
4793	>	propagateCompletion();
4794	>	}
4795	>	}
4796		}
4797
4798	<	final class KeySet extends AbstractSet<K> {
4799	<	public int size() {
4800	<	return ConcurrentHashMapV8.this.size();
4798	>	@SuppressWarnings("serial")
4799	>	static final class ForEachEntryTask<K,V>
4800	>	extends BulkTask<K,V,Void> {
4801	>	final Action<? super Entry<K,V>> action;
4802	>	ForEachEntryTask
4803	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
4804	>	Action<? super Entry<K,V>> action) {
4805	>	super(p, b, i, f, t);
4806	>	this.action = action;
4807	>	}
4808	>	public final void compute() {
4809	>	final Action<? super Entry<K,V>> action;
4810	>	if ((action = this.action) != null) {
4811	>	for (int i = baseIndex, f, h; batch > 0 &&
4812	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
4813	>	addToPendingCount(1);
4814	>	new ForEachEntryTask<K,V>
4815	>	(this, batch >>>= 1, baseLimit = h, f, tab,
4816	>	action).fork();
4817	>	}
4818	>	for (Node<K,V> p; (p = advance()) != null; )
4819	>	action.apply(p);
4820	>	propagateCompletion();
4821	>	}
4822		}
4823	<	public boolean isEmpty() {
4824	<	return ConcurrentHashMapV8.this.isEmpty();
4823	>	}
4824	>
4825	>	@SuppressWarnings("serial")
4826	>	static final class ForEachMappingTask<K,V>
4827	>	extends BulkTask<K,V,Void> {
4828	>	final BiAction<? super K, ? super V> action;
4829	>	ForEachMappingTask
4830	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
4831	>	BiAction<? super K,? super V> action) {
4832	>	super(p, b, i, f, t);
4833	>	this.action = action;
4834	>	}
4835	>	public final void compute() {
4836	>	final BiAction<? super K, ? super V> action;
4837	>	if ((action = this.action) != null) {
4838	>	for (int i = baseIndex, f, h; batch > 0 &&
4839	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
4840	>	addToPendingCount(1);
4841	>	new ForEachMappingTask<K,V>
4842	>	(this, batch >>>= 1, baseLimit = h, f, tab,
4843	>	action).fork();
4844	>	}
4845	>	for (Node<K,V> p; (p = advance()) != null; )
4846	>	action.apply(p.key, p.val);
4847	>	propagateCompletion();
4848	>	}
4849		}
4850	<	public void clear() {
4851	<	ConcurrentHashMapV8.this.clear();
4850	>	}
4851	>
4852	>	@SuppressWarnings("serial")
4853	>	static final class ForEachTransformedKeyTask<K,V,U>
4854	>	extends BulkTask<K,V,Void> {
4855	>	final Fun<? super K, ? extends U> transformer;
4856	>	final Action<? super U> action;
4857	>	ForEachTransformedKeyTask
4858	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
4859	>	Fun<? super K, ? extends U> transformer, Action<? super U> action) {
4860	>	super(p, b, i, f, t);
4861	>	this.transformer = transformer; this.action = action;
4862	>	}
4863	>	public final void compute() {
4864	>	final Fun<? super K, ? extends U> transformer;
4865	>	final Action<? super U> action;
4866	>	if ((transformer = this.transformer) != null &&
4867	>	(action = this.action) != null) {
4868	>	for (int i = baseIndex, f, h; batch > 0 &&
4869	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
4870	>	addToPendingCount(1);
4871	>	new ForEachTransformedKeyTask<K,V,U>
4872	>	(this, batch >>>= 1, baseLimit = h, f, tab,
4873	>	transformer, action).fork();
4874	>	}
4875	>	for (Node<K,V> p; (p = advance()) != null; ) {
4876	>	U u;
4877	>	if ((u = transformer.apply(p.key)) != null)
4878	>	action.apply(u);
4879	>	}
4880	>	propagateCompletion();
4881	>	}
4882		}
4883	<	public Iterator<K> iterator() {
4884	<	return new KeyIterator();
4883	>	}
4884	>
4885	>	@SuppressWarnings("serial")
4886	>	static final class ForEachTransformedValueTask<K,V,U>
4887	>	extends BulkTask<K,V,Void> {
4888	>	final Fun<? super V, ? extends U> transformer;
4889	>	final Action<? super U> action;
4890	>	ForEachTransformedValueTask
4891	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
4892	>	Fun<? super V, ? extends U> transformer, Action<? super U> action) {
4893	>	super(p, b, i, f, t);
4894	>	this.transformer = transformer; this.action = action;
4895	>	}
4896	>	public final void compute() {
4897	>	final Fun<? super V, ? extends U> transformer;
4898	>	final Action<? super U> action;
4899	>	if ((transformer = this.transformer) != null &&
4900	>	(action = this.action) != null) {
4901	>	for (int i = baseIndex, f, h; batch > 0 &&
4902	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
4903	>	addToPendingCount(1);
4904	>	new ForEachTransformedValueTask<K,V,U>
4905	>	(this, batch >>>= 1, baseLimit = h, f, tab,
4906	>	transformer, action).fork();
4907	>	}
4908	>	for (Node<K,V> p; (p = advance()) != null; ) {
4909	>	U u;
4910	>	if ((u = transformer.apply(p.val)) != null)
4911	>	action.apply(u);
4912	>	}
4913	>	propagateCompletion();
4914	>	}
4915		}
4916	<	public boolean contains(Object o) {
4917	<	return ConcurrentHashMapV8.this.containsKey(o);
4916	>	}
4917	>
4918	>	@SuppressWarnings("serial")
4919	>	static final class ForEachTransformedEntryTask<K,V,U>
4920	>	extends BulkTask<K,V,Void> {
4921	>	final Fun<Map.Entry<K,V>, ? extends U> transformer;
4922	>	final Action<? super U> action;
4923	>	ForEachTransformedEntryTask
4924	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
4925	>	Fun<Map.Entry<K,V>, ? extends U> transformer, Action<? super U> action) {
4926	>	super(p, b, i, f, t);
4927	>	this.transformer = transformer; this.action = action;
4928	>	}
4929	>	public final void compute() {
4930	>	final Fun<Map.Entry<K,V>, ? extends U> transformer;
4931	>	final Action<? super U> action;
4932	>	if ((transformer = this.transformer) != null &&
4933	>	(action = this.action) != null) {
4934	>	for (int i = baseIndex, f, h; batch > 0 &&
4935	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
4936	>	addToPendingCount(1);
4937	>	new ForEachTransformedEntryTask<K,V,U>
4938	>	(this, batch >>>= 1, baseLimit = h, f, tab,
4939	>	transformer, action).fork();
4940	>	}
4941	>	for (Node<K,V> p; (p = advance()) != null; ) {
4942	>	U u;
4943	>	if ((u = transformer.apply(p)) != null)
4944	>	action.apply(u);
4945	>	}
4946	>	propagateCompletion();
4947	>	}
4948		}
4949	<	public boolean remove(Object o) {
4950	<	return ConcurrentHashMapV8.this.remove(o) != null;
4949	>	}
4950	>
4951	>	@SuppressWarnings("serial")
4952	>	static final class ForEachTransformedMappingTask<K,V,U>
4953	>	extends BulkTask<K,V,Void> {
4954	>	final BiFun<? super K, ? super V, ? extends U> transformer;
4955	>	final Action<? super U> action;
4956	>	ForEachTransformedMappingTask
4957	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
4958	>	BiFun<? super K, ? super V, ? extends U> transformer,
4959	>	Action<? super U> action) {
4960	>	super(p, b, i, f, t);
4961	>	this.transformer = transformer; this.action = action;
4962	>	}
4963	>	public final void compute() {
4964	>	final BiFun<? super K, ? super V, ? extends U> transformer;
4965	>	final Action<? super U> action;
4966	>	if ((transformer = this.transformer) != null &&
4967	>	(action = this.action) != null) {
4968	>	for (int i = baseIndex, f, h; batch > 0 &&
4969	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
4970	>	addToPendingCount(1);
4971	>	new ForEachTransformedMappingTask<K,V,U>
4972	>	(this, batch >>>= 1, baseLimit = h, f, tab,
4973	>	transformer, action).fork();
4974	>	}
4975	>	for (Node<K,V> p; (p = advance()) != null; ) {
4976	>	U u;
4977	>	if ((u = transformer.apply(p.key, p.val)) != null)
4978	>	action.apply(u);
4979	>	}
4980	>	propagateCompletion();
4981	>	}
4982		}
4983		}
4984
4985	<	final class Values extends AbstractCollection<V> {
4986	<	public int size() {
4987	<	return ConcurrentHashMapV8.this.size();
4985	>	@SuppressWarnings("serial")
4986	>	static final class SearchKeysTask<K,V,U>
4987	>	extends BulkTask<K,V,U> {
4988	>	final Fun<? super K, ? extends U> searchFunction;
4989	>	final AtomicReference<U> result;
4990	>	SearchKeysTask
4991	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
4992	>	Fun<? super K, ? extends U> searchFunction,
4993	>	AtomicReference<U> result) {
4994	>	super(p, b, i, f, t);
4995	>	this.searchFunction = searchFunction; this.result = result;
4996	>	}
4997	>	public final U getRawResult() { return result.get(); }
4998	>	public final void compute() {
4999	>	final Fun<? super K, ? extends U> searchFunction;
5000	>	final AtomicReference<U> result;
5001	>	if ((searchFunction = this.searchFunction) != null &&
5002	>	(result = this.result) != null) {
5003	>	for (int i = baseIndex, f, h; batch > 0 &&
5004	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
5005	>	if (result.get() != null)
5006	>	return;
5007	>	addToPendingCount(1);
5008	>	new SearchKeysTask<K,V,U>
5009	>	(this, batch >>>= 1, baseLimit = h, f, tab,
5010	>	searchFunction, result).fork();
5011	>	}
5012	>	while (result.get() == null) {
5013	>	U u;
5014	>	Node<K,V> p;
5015	>	if ((p = advance()) == null) {
5016	>	propagateCompletion();
5017	>	break;
5018	>	}
5019	>	if ((u = searchFunction.apply(p.key)) != null) {
5020	>	if (result.compareAndSet(null, u))
5021	>	quietlyCompleteRoot();
5022	>	break;
5023	>	}
5024	>	}
5025	>	}
5026		}
5027	<	public boolean isEmpty() {
5028	<	return ConcurrentHashMapV8.this.isEmpty();
5027	>	}
5028	>
5029	>	@SuppressWarnings("serial")
5030	>	static final class SearchValuesTask<K,V,U>
5031	>	extends BulkTask<K,V,U> {
5032	>	final Fun<? super V, ? extends U> searchFunction;
5033	>	final AtomicReference<U> result;
5034	>	SearchValuesTask
5035	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
5036	>	Fun<? super V, ? extends U> searchFunction,
5037	>	AtomicReference<U> result) {
5038	>	super(p, b, i, f, t);
5039	>	this.searchFunction = searchFunction; this.result = result;
5040	>	}
5041	>	public final U getRawResult() { return result.get(); }
5042	>	public final void compute() {
5043	>	final Fun<? super V, ? extends U> searchFunction;
5044	>	final AtomicReference<U> result;
5045	>	if ((searchFunction = this.searchFunction) != null &&
5046	>	(result = this.result) != null) {
5047	>	for (int i = baseIndex, f, h; batch > 0 &&
5048	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
5049	>	if (result.get() != null)
5050	>	return;
5051	>	addToPendingCount(1);
5052	>	new SearchValuesTask<K,V,U>
5053	>	(this, batch >>>= 1, baseLimit = h, f, tab,
5054	>	searchFunction, result).fork();
5055	>	}
5056	>	while (result.get() == null) {
5057	>	U u;
5058	>	Node<K,V> p;
5059	>	if ((p = advance()) == null) {
5060	>	propagateCompletion();
5061	>	break;
5062	>	}
5063	>	if ((u = searchFunction.apply(p.val)) != null) {
5064	>	if (result.compareAndSet(null, u))
5065	>	quietlyCompleteRoot();
5066	>	break;
5067	>	}
5068	>	}
5069	>	}
5070		}
5071	<	public void clear() {
5072	<	ConcurrentHashMapV8.this.clear();
5071	>	}
5072	>
5073	>	@SuppressWarnings("serial")
5074	>	static final class SearchEntriesTask<K,V,U>
5075	>	extends BulkTask<K,V,U> {
5076	>	final Fun<Entry<K,V>, ? extends U> searchFunction;
5077	>	final AtomicReference<U> result;
5078	>	SearchEntriesTask
5079	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
5080	>	Fun<Entry<K,V>, ? extends U> searchFunction,
5081	>	AtomicReference<U> result) {
5082	>	super(p, b, i, f, t);
5083	>	this.searchFunction = searchFunction; this.result = result;
5084	>	}
5085	>	public final U getRawResult() { return result.get(); }
5086	>	public final void compute() {
5087	>	final Fun<Entry<K,V>, ? extends U> searchFunction;
5088	>	final AtomicReference<U> result;
5089	>	if ((searchFunction = this.searchFunction) != null &&
5090	>	(result = this.result) != null) {
5091	>	for (int i = baseIndex, f, h; batch > 0 &&
5092	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
5093	>	if (result.get() != null)
5094	>	return;
5095	>	addToPendingCount(1);
5096	>	new SearchEntriesTask<K,V,U>
5097	>	(this, batch >>>= 1, baseLimit = h, f, tab,
5098	>	searchFunction, result).fork();
5099	>	}
5100	>	while (result.get() == null) {
5101	>	U u;
5102	>	Node<K,V> p;
5103	>	if ((p = advance()) == null) {
5104	>	propagateCompletion();
5105	>	break;
5106	>	}
5107	>	if ((u = searchFunction.apply(p)) != null) {
5108	>	if (result.compareAndSet(null, u))
5109	>	quietlyCompleteRoot();
5110	>	return;
5111	>	}
5112	>	}
5113	>	}
5114		}
5115	<	public Iterator<V> iterator() {
5116	<	return new ValueIterator();
5115	>	}
5116	>
5117	>	@SuppressWarnings("serial")
5118	>	static final class SearchMappingsTask<K,V,U>
5119	>	extends BulkTask<K,V,U> {
5120	>	final BiFun<? super K, ? super V, ? extends U> searchFunction;
5121	>	final AtomicReference<U> result;
5122	>	SearchMappingsTask
5123	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
5124	>	BiFun<? super K, ? super V, ? extends U> searchFunction,
5125	>	AtomicReference<U> result) {
5126	>	super(p, b, i, f, t);
5127	>	this.searchFunction = searchFunction; this.result = result;
5128	>	}
5129	>	public final U getRawResult() { return result.get(); }
5130	>	public final void compute() {
5131	>	final BiFun<? super K, ? super V, ? extends U> searchFunction;
5132	>	final AtomicReference<U> result;
5133	>	if ((searchFunction = this.searchFunction) != null &&
5134	>	(result = this.result) != null) {
5135	>	for (int i = baseIndex, f, h; batch > 0 &&
5136	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
5137	>	if (result.get() != null)
5138	>	return;
5139	>	addToPendingCount(1);
5140	>	new SearchMappingsTask<K,V,U>
5141	>	(this, batch >>>= 1, baseLimit = h, f, tab,
5142	>	searchFunction, result).fork();
5143	>	}
5144	>	while (result.get() == null) {
5145	>	U u;
5146	>	Node<K,V> p;
5147	>	if ((p = advance()) == null) {
5148	>	propagateCompletion();
5149	>	break;
5150	>	}
5151	>	if ((u = searchFunction.apply(p.key, p.val)) != null) {
5152	>	if (result.compareAndSet(null, u))
5153	>	quietlyCompleteRoot();
5154	>	break;
5155	>	}
5156	>	}
5157	>	}
5158		}
5159	<	public boolean contains(Object o) {
5160	<	return ConcurrentHashMapV8.this.containsValue(o);
5159	>	}
5160	>
5161	>	@SuppressWarnings("serial")
5162	>	static final class ReduceKeysTask<K,V>
5163	>	extends BulkTask<K,V,K> {
5164	>	final BiFun<? super K, ? super K, ? extends K> reducer;
5165	>	K result;
5166	>	ReduceKeysTask<K,V> rights, nextRight;
5167	>	ReduceKeysTask
5168	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
5169	>	ReduceKeysTask<K,V> nextRight,
5170	>	BiFun<? super K, ? super K, ? extends K> reducer) {
5171	>	super(p, b, i, f, t); this.nextRight = nextRight;
5172	>	this.reducer = reducer;
5173	>	}
5174	>	public final K getRawResult() { return result; }
5175	>	public final void compute() {
5176	>	final BiFun<? super K, ? super K, ? extends K> reducer;
5177	>	if ((reducer = this.reducer) != null) {
5178	>	for (int i = baseIndex, f, h; batch > 0 &&
5179	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
5180	>	addToPendingCount(1);
5181	>	(rights = new ReduceKeysTask<K,V>
5182	>	(this, batch >>>= 1, baseLimit = h, f, tab,
5183	>	rights, reducer)).fork();
5184	>	}
5185	>	K r = null;
5186	>	for (Node<K,V> p; (p = advance()) != null; ) {
5187	>	K u = p.key;
5188	>	r = (r == null) ? u : u == null ? r : reducer.apply(r, u);
5189	>	}
5190	>	result = r;
5191	>	CountedCompleter<?> c;
5192	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5193	>	@SuppressWarnings("unchecked") ReduceKeysTask<K,V>
5194	>	t = (ReduceKeysTask<K,V>)c,
5195	>	s = t.rights;
5196	>	while (s != null) {
5197	>	K tr, sr;
5198	>	if ((sr = s.result) != null)
5199	>	t.result = (((tr = t.result) == null) ? sr :
5200	>	reducer.apply(tr, sr));
5201	>	s = t.rights = s.nextRight;
5202	>	}
5203	>	}
5204	>	}
5205		}
5206		}
5207
5208	<	final class EntrySet extends AbstractSet<Map.Entry<K,V>> {
5209	<	public int size() {
5210	<	return ConcurrentHashMapV8.this.size();
5208	>	@SuppressWarnings("serial")
5209	>	static final class ReduceValuesTask<K,V>
5210	>	extends BulkTask<K,V,V> {
5211	>	final BiFun<? super V, ? super V, ? extends V> reducer;
5212	>	V result;
5213	>	ReduceValuesTask<K,V> rights, nextRight;
5214	>	ReduceValuesTask
5215	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
5216	>	ReduceValuesTask<K,V> nextRight,
5217	>	BiFun<? super V, ? super V, ? extends V> reducer) {
5218	>	super(p, b, i, f, t); this.nextRight = nextRight;
5219	>	this.reducer = reducer;
5220	>	}
5221	>	public final V getRawResult() { return result; }
5222	>	public final void compute() {
5223	>	final BiFun<? super V, ? super V, ? extends V> reducer;
5224	>	if ((reducer = this.reducer) != null) {
5225	>	for (int i = baseIndex, f, h; batch > 0 &&
5226	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
5227	>	addToPendingCount(1);
5228	>	(rights = new ReduceValuesTask<K,V>
5229	>	(this, batch >>>= 1, baseLimit = h, f, tab,
5230	>	rights, reducer)).fork();
5231	>	}
5232	>	V r = null;
5233	>	for (Node<K,V> p; (p = advance()) != null; ) {
5234	>	V v = p.val;
5235	>	r = (r == null) ? v : reducer.apply(r, v);
5236	>	}
5237	>	result = r;
5238	>	CountedCompleter<?> c;
5239	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5240	>	@SuppressWarnings("unchecked") ReduceValuesTask<K,V>
5241	>	t = (ReduceValuesTask<K,V>)c,
5242	>	s = t.rights;
5243	>	while (s != null) {
5244	>	V tr, sr;
5245	>	if ((sr = s.result) != null)
5246	>	t.result = (((tr = t.result) == null) ? sr :
5247	>	reducer.apply(tr, sr));
5248	>	s = t.rights = s.nextRight;
5249	>	}
5250	>	}
5251	>	}
5252		}
5253	<	public boolean isEmpty() {
5254	<	return ConcurrentHashMapV8.this.isEmpty();
5253	>	}
5254	>
5255	>	@SuppressWarnings("serial")
5256	>	static final class ReduceEntriesTask<K,V>
5257	>	extends BulkTask<K,V,Map.Entry<K,V>> {
5258	>	final BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer;
5259	>	Map.Entry<K,V> result;
5260	>	ReduceEntriesTask<K,V> rights, nextRight;
5261	>	ReduceEntriesTask
5262	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
5263	>	ReduceEntriesTask<K,V> nextRight,
5264	>	BiFun<Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer) {
5265	>	super(p, b, i, f, t); this.nextRight = nextRight;
5266	>	this.reducer = reducer;
5267	>	}
5268	>	public final Map.Entry<K,V> getRawResult() { return result; }
5269	>	public final void compute() {
5270	>	final BiFun<Map.Entry<K,V>, Map.Entry<K,V>, ? extends Map.Entry<K,V>> reducer;
5271	>	if ((reducer = this.reducer) != null) {
5272	>	for (int i = baseIndex, f, h; batch > 0 &&
5273	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
5274	>	addToPendingCount(1);
5275	>	(rights = new ReduceEntriesTask<K,V>
5276	>	(this, batch >>>= 1, baseLimit = h, f, tab,
5277	>	rights, reducer)).fork();
5278	>	}
5279	>	Map.Entry<K,V> r = null;
5280	>	for (Node<K,V> p; (p = advance()) != null; )
5281	>	r = (r == null) ? p : reducer.apply(r, p);
5282	>	result = r;
5283	>	CountedCompleter<?> c;
5284	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5285	>	@SuppressWarnings("unchecked") ReduceEntriesTask<K,V>
5286	>	t = (ReduceEntriesTask<K,V>)c,
5287	>	s = t.rights;
5288	>	while (s != null) {
5289	>	Map.Entry<K,V> tr, sr;
5290	>	if ((sr = s.result) != null)
5291	>	t.result = (((tr = t.result) == null) ? sr :
5292	>	reducer.apply(tr, sr));
5293	>	s = t.rights = s.nextRight;
5294	>	}
5295	>	}
5296	>	}
5297		}
5298	<	public void clear() {
5299	<	ConcurrentHashMapV8.this.clear();
5298	>	}
5299	>
5300	>	@SuppressWarnings("serial")
5301	>	static final class MapReduceKeysTask<K,V,U>
5302	>	extends BulkTask<K,V,U> {
5303	>	final Fun<? super K, ? extends U> transformer;
5304	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5305	>	U result;
5306	>	MapReduceKeysTask<K,V,U> rights, nextRight;
5307	>	MapReduceKeysTask
5308	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
5309	>	MapReduceKeysTask<K,V,U> nextRight,
5310	>	Fun<? super K, ? extends U> transformer,
5311	>	BiFun<? super U, ? super U, ? extends U> reducer) {
5312	>	super(p, b, i, f, t); this.nextRight = nextRight;
5313	>	this.transformer = transformer;
5314	>	this.reducer = reducer;
5315	>	}
5316	>	public final U getRawResult() { return result; }
5317	>	public final void compute() {
5318	>	final Fun<? super K, ? extends U> transformer;
5319	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5320	>	if ((transformer = this.transformer) != null &&
5321	>	(reducer = this.reducer) != null) {
5322	>	for (int i = baseIndex, f, h; batch > 0 &&
5323	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
5324	>	addToPendingCount(1);
5325	>	(rights = new MapReduceKeysTask<K,V,U>
5326	>	(this, batch >>>= 1, baseLimit = h, f, tab,
5327	>	rights, transformer, reducer)).fork();
5328	>	}
5329	>	U r = null;
5330	>	for (Node<K,V> p; (p = advance()) != null; ) {
5331	>	U u;
5332	>	if ((u = transformer.apply(p.key)) != null)
5333	>	r = (r == null) ? u : reducer.apply(r, u);
5334	>	}
5335	>	result = r;
5336	>	CountedCompleter<?> c;
5337	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5338	>	@SuppressWarnings("unchecked") MapReduceKeysTask<K,V,U>
5339	>	t = (MapReduceKeysTask<K,V,U>)c,
5340	>	s = t.rights;
5341	>	while (s != null) {
5342	>	U tr, sr;
5343	>	if ((sr = s.result) != null)
5344	>	t.result = (((tr = t.result) == null) ? sr :
5345	>	reducer.apply(tr, sr));
5346	>	s = t.rights = s.nextRight;
5347	>	}
5348	>	}
5349	>	}
5350		}
5351	<	public Iterator<Map.Entry<K,V>> iterator() {
5352	<	return new EntryIterator();
5351	>	}
5352	>
5353	>	@SuppressWarnings("serial")
5354	>	static final class MapReduceValuesTask<K,V,U>
5355	>	extends BulkTask<K,V,U> {
5356	>	final Fun<? super V, ? extends U> transformer;
5357	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5358	>	U result;
5359	>	MapReduceValuesTask<K,V,U> rights, nextRight;
5360	>	MapReduceValuesTask
5361	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
5362	>	MapReduceValuesTask<K,V,U> nextRight,
5363	>	Fun<? super V, ? extends U> transformer,
5364	>	BiFun<? super U, ? super U, ? extends U> reducer) {
5365	>	super(p, b, i, f, t); this.nextRight = nextRight;
5366	>	this.transformer = transformer;
5367	>	this.reducer = reducer;
5368	>	}
5369	>	public final U getRawResult() { return result; }
5370	>	public final void compute() {
5371	>	final Fun<? super V, ? extends U> transformer;
5372	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5373	>	if ((transformer = this.transformer) != null &&
5374	>	(reducer = this.reducer) != null) {
5375	>	for (int i = baseIndex, f, h; batch > 0 &&
5376	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
5377	>	addToPendingCount(1);
5378	>	(rights = new MapReduceValuesTask<K,V,U>
5379	>	(this, batch >>>= 1, baseLimit = h, f, tab,
5380	>	rights, transformer, reducer)).fork();
5381	>	}
5382	>	U r = null;
5383	>	for (Node<K,V> p; (p = advance()) != null; ) {
5384	>	U u;
5385	>	if ((u = transformer.apply(p.val)) != null)
5386	>	r = (r == null) ? u : reducer.apply(r, u);
5387	>	}
5388	>	result = r;
5389	>	CountedCompleter<?> c;
5390	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5391	>	@SuppressWarnings("unchecked") MapReduceValuesTask<K,V,U>
5392	>	t = (MapReduceValuesTask<K,V,U>)c,
5393	>	s = t.rights;
5394	>	while (s != null) {
5395	>	U tr, sr;
5396	>	if ((sr = s.result) != null)
5397	>	t.result = (((tr = t.result) == null) ? sr :
5398	>	reducer.apply(tr, sr));
5399	>	s = t.rights = s.nextRight;
5400	>	}
5401	>	}
5402	>	}
5403		}
5404	<	public boolean contains(Object o) {
5405	<	if (!(o instanceof Map.Entry))
5406	<	return false;
5407	<	Map.Entry<?,?> e = (Map.Entry<?,?>)o;
5408	<	V v = ConcurrentHashMapV8.this.get(e.getKey());
5409	<	return v != null && v.equals(e.getValue());
5404	>	}
5405	>
5406	>	@SuppressWarnings("serial")
5407	>	static final class MapReduceEntriesTask<K,V,U>
5408	>	extends BulkTask<K,V,U> {
5409	>	final Fun<Map.Entry<K,V>, ? extends U> transformer;
5410	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5411	>	U result;
5412	>	MapReduceEntriesTask<K,V,U> rights, nextRight;
5413	>	MapReduceEntriesTask
5414	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
5415	>	MapReduceEntriesTask<K,V,U> nextRight,
5416	>	Fun<Map.Entry<K,V>, ? extends U> transformer,
5417	>	BiFun<? super U, ? super U, ? extends U> reducer) {
5418	>	super(p, b, i, f, t); this.nextRight = nextRight;
5419	>	this.transformer = transformer;
5420	>	this.reducer = reducer;
5421	>	}
5422	>	public final U getRawResult() { return result; }
5423	>	public final void compute() {
5424	>	final Fun<Map.Entry<K,V>, ? extends U> transformer;
5425	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5426	>	if ((transformer = this.transformer) != null &&
5427	>	(reducer = this.reducer) != null) {
5428	>	for (int i = baseIndex, f, h; batch > 0 &&
5429	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
5430	>	addToPendingCount(1);
5431	>	(rights = new MapReduceEntriesTask<K,V,U>
5432	>	(this, batch >>>= 1, baseLimit = h, f, tab,
5433	>	rights, transformer, reducer)).fork();
5434	>	}
5435	>	U r = null;
5436	>	for (Node<K,V> p; (p = advance()) != null; ) {
5437	>	U u;
5438	>	if ((u = transformer.apply(p)) != null)
5439	>	r = (r == null) ? u : reducer.apply(r, u);
5440	>	}
5441	>	result = r;
5442	>	CountedCompleter<?> c;
5443	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5444	>	@SuppressWarnings("unchecked") MapReduceEntriesTask<K,V,U>
5445	>	t = (MapReduceEntriesTask<K,V,U>)c,
5446	>	s = t.rights;
5447	>	while (s != null) {
5448	>	U tr, sr;
5449	>	if ((sr = s.result) != null)
5450	>	t.result = (((tr = t.result) == null) ? sr :
5451	>	reducer.apply(tr, sr));
5452	>	s = t.rights = s.nextRight;
5453	>	}
5454	>	}
5455	>	}
5456		}
5457	<	public boolean remove(Object o) {
5458	<	if (!(o instanceof Map.Entry))
5459	<	return false;
5460	<	Map.Entry<?,?> e = (Map.Entry<?,?>)o;
5461	<	return ConcurrentHashMapV8.this.remove(e.getKey(), e.getValue());
5457	>	}
5458	>
5459	>	@SuppressWarnings("serial")
5460	>	static final class MapReduceMappingsTask<K,V,U>
5461	>	extends BulkTask<K,V,U> {
5462	>	final BiFun<? super K, ? super V, ? extends U> transformer;
5463	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5464	>	U result;
5465	>	MapReduceMappingsTask<K,V,U> rights, nextRight;
5466	>	MapReduceMappingsTask
5467	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
5468	>	MapReduceMappingsTask<K,V,U> nextRight,
5469	>	BiFun<? super K, ? super V, ? extends U> transformer,
5470	>	BiFun<? super U, ? super U, ? extends U> reducer) {
5471	>	super(p, b, i, f, t); this.nextRight = nextRight;
5472	>	this.transformer = transformer;
5473	>	this.reducer = reducer;
5474	>	}
5475	>	public final U getRawResult() { return result; }
5476	>	public final void compute() {
5477	>	final BiFun<? super K, ? super V, ? extends U> transformer;
5478	>	final BiFun<? super U, ? super U, ? extends U> reducer;
5479	>	if ((transformer = this.transformer) != null &&
5480	>	(reducer = this.reducer) != null) {
5481	>	for (int i = baseIndex, f, h; batch > 0 &&
5482	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
5483	>	addToPendingCount(1);
5484	>	(rights = new MapReduceMappingsTask<K,V,U>
5485	>	(this, batch >>>= 1, baseLimit = h, f, tab,
5486	>	rights, transformer, reducer)).fork();
5487	>	}
5488	>	U r = null;
5489	>	for (Node<K,V> p; (p = advance()) != null; ) {
5490	>	U u;
5491	>	if ((u = transformer.apply(p.key, p.val)) != null)
5492	>	r = (r == null) ? u : reducer.apply(r, u);
5493	>	}
5494	>	result = r;
5495	>	CountedCompleter<?> c;
5496	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5497	>	@SuppressWarnings("unchecked") MapReduceMappingsTask<K,V,U>
5498	>	t = (MapReduceMappingsTask<K,V,U>)c,
5499	>	s = t.rights;
5500	>	while (s != null) {
5501	>	U tr, sr;
5502	>	if ((sr = s.result) != null)
5503	>	t.result = (((tr = t.result) == null) ? sr :
5504	>	reducer.apply(tr, sr));
5505	>	s = t.rights = s.nextRight;
5506	>	}
5507	>	}
5508	>	}
5509	>	}
5510	>	}
5511	>
5512	>	@SuppressWarnings("serial")
5513	>	static final class MapReduceKeysToDoubleTask<K,V>
5514	>	extends BulkTask<K,V,Double> {
5515	>	final ObjectToDouble<? super K> transformer;
5516	>	final DoubleByDoubleToDouble reducer;
5517	>	final double basis;
5518	>	double result;
5519	>	MapReduceKeysToDoubleTask<K,V> rights, nextRight;
5520	>	MapReduceKeysToDoubleTask
5521	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
5522	>	MapReduceKeysToDoubleTask<K,V> nextRight,
5523	>	ObjectToDouble<? super K> transformer,
5524	>	double basis,
5525	>	DoubleByDoubleToDouble reducer) {
5526	>	super(p, b, i, f, t); this.nextRight = nextRight;
5527	>	this.transformer = transformer;
5528	>	this.basis = basis; this.reducer = reducer;
5529	>	}
5530	>	public final Double getRawResult() { return result; }
5531	>	public final void compute() {
5532	>	final ObjectToDouble<? super K> transformer;
5533	>	final DoubleByDoubleToDouble reducer;
5534	>	if ((transformer = this.transformer) != null &&
5535	>	(reducer = this.reducer) != null) {
5536	>	double r = this.basis;
5537	>	for (int i = baseIndex, f, h; batch > 0 &&
5538	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
5539	>	addToPendingCount(1);
5540	>	(rights = new MapReduceKeysToDoubleTask<K,V>
5541	>	(this, batch >>>= 1, baseLimit = h, f, tab,
5542	>	rights, transformer, r, reducer)).fork();
5543	>	}
5544	>	for (Node<K,V> p; (p = advance()) != null; )
5545	>	r = reducer.apply(r, transformer.apply(p.key));
5546	>	result = r;
5547	>	CountedCompleter<?> c;
5548	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5549	>	@SuppressWarnings("unchecked") MapReduceKeysToDoubleTask<K,V>
5550	>	t = (MapReduceKeysToDoubleTask<K,V>)c,
5551	>	s = t.rights;
5552	>	while (s != null) {
5553	>	t.result = reducer.apply(t.result, s.result);
5554	>	s = t.rights = s.nextRight;
5555	>	}
5556	>	}
5557	>	}
5558		}
5559		}
5560
5561	<	/* ---------------- Serialization Support -------------- */
5561	>	@SuppressWarnings("serial")
5562	>	static final class MapReduceValuesToDoubleTask<K,V>
5563	>	extends BulkTask<K,V,Double> {
5564	>	final ObjectToDouble<? super V> transformer;
5565	>	final DoubleByDoubleToDouble reducer;
5566	>	final double basis;
5567	>	double result;
5568	>	MapReduceValuesToDoubleTask<K,V> rights, nextRight;
5569	>	MapReduceValuesToDoubleTask
5570	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
5571	>	MapReduceValuesToDoubleTask<K,V> nextRight,
5572	>	ObjectToDouble<? super V> transformer,
5573	>	double basis,
5574	>	DoubleByDoubleToDouble reducer) {
5575	>	super(p, b, i, f, t); this.nextRight = nextRight;
5576	>	this.transformer = transformer;
5577	>	this.basis = basis; this.reducer = reducer;
5578	>	}
5579	>	public final Double getRawResult() { return result; }
5580	>	public final void compute() {
5581	>	final ObjectToDouble<? super V> transformer;
5582	>	final DoubleByDoubleToDouble reducer;
5583	>	if ((transformer = this.transformer) != null &&
5584	>	(reducer = this.reducer) != null) {
5585	>	double r = this.basis;
5586	>	for (int i = baseIndex, f, h; batch > 0 &&
5587	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
5588	>	addToPendingCount(1);
5589	>	(rights = new MapReduceValuesToDoubleTask<K,V>
5590	>	(this, batch >>>= 1, baseLimit = h, f, tab,
5591	>	rights, transformer, r, reducer)).fork();
5592	>	}
5593	>	for (Node<K,V> p; (p = advance()) != null; )
5594	>	r = reducer.apply(r, transformer.apply(p.val));
5595	>	result = r;
5596	>	CountedCompleter<?> c;
5597	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5598	>	@SuppressWarnings("unchecked") MapReduceValuesToDoubleTask<K,V>
5599	>	t = (MapReduceValuesToDoubleTask<K,V>)c,
5600	>	s = t.rights;
5601	>	while (s != null) {
5602	>	t.result = reducer.apply(t.result, s.result);
5603	>	s = t.rights = s.nextRight;
5604	>	}
5605	>	}
5606	>	}
5607	>	}
5608	>	}
5609	>
5610	>	@SuppressWarnings("serial")
5611	>	static final class MapReduceEntriesToDoubleTask<K,V>
5612	>	extends BulkTask<K,V,Double> {
5613	>	final ObjectToDouble<Map.Entry<K,V>> transformer;
5614	>	final DoubleByDoubleToDouble reducer;
5615	>	final double basis;
5616	>	double result;
5617	>	MapReduceEntriesToDoubleTask<K,V> rights, nextRight;
5618	>	MapReduceEntriesToDoubleTask
5619	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
5620	>	MapReduceEntriesToDoubleTask<K,V> nextRight,
5621	>	ObjectToDouble<Map.Entry<K,V>> transformer,
5622	>	double basis,
5623	>	DoubleByDoubleToDouble reducer) {
5624	>	super(p, b, i, f, t); this.nextRight = nextRight;
5625	>	this.transformer = transformer;
5626	>	this.basis = basis; this.reducer = reducer;
5627	>	}
5628	>	public final Double getRawResult() { return result; }
5629	>	public final void compute() {
5630	>	final ObjectToDouble<Map.Entry<K,V>> transformer;
5631	>	final DoubleByDoubleToDouble reducer;
5632	>	if ((transformer = this.transformer) != null &&
5633	>	(reducer = this.reducer) != null) {
5634	>	double r = this.basis;
5635	>	for (int i = baseIndex, f, h; batch > 0 &&
5636	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
5637	>	addToPendingCount(1);
5638	>	(rights = new MapReduceEntriesToDoubleTask<K,V>
5639	>	(this, batch >>>= 1, baseLimit = h, f, tab,
5640	>	rights, transformer, r, reducer)).fork();
5641	>	}
5642	>	for (Node<K,V> p; (p = advance()) != null; )
5643	>	r = reducer.apply(r, transformer.apply(p));
5644	>	result = r;
5645	>	CountedCompleter<?> c;
5646	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5647	>	@SuppressWarnings("unchecked") MapReduceEntriesToDoubleTask<K,V>
5648	>	t = (MapReduceEntriesToDoubleTask<K,V>)c,
5649	>	s = t.rights;
5650	>	while (s != null) {
5651	>	t.result = reducer.apply(t.result, s.result);
5652	>	s = t.rights = s.nextRight;
5653	>	}
5654	>	}
5655	>	}
5656	>	}
5657	>	}
5658	>
5659	>	@SuppressWarnings("serial")
5660	>	static final class MapReduceMappingsToDoubleTask<K,V>
5661	>	extends BulkTask<K,V,Double> {
5662	>	final ObjectByObjectToDouble<? super K, ? super V> transformer;
5663	>	final DoubleByDoubleToDouble reducer;
5664	>	final double basis;
5665	>	double result;
5666	>	MapReduceMappingsToDoubleTask<K,V> rights, nextRight;
5667	>	MapReduceMappingsToDoubleTask
5668	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
5669	>	MapReduceMappingsToDoubleTask<K,V> nextRight,
5670	>	ObjectByObjectToDouble<? super K, ? super V> transformer,
5671	>	double basis,
5672	>	DoubleByDoubleToDouble reducer) {
5673	>	super(p, b, i, f, t); this.nextRight = nextRight;
5674	>	this.transformer = transformer;
5675	>	this.basis = basis; this.reducer = reducer;
5676	>	}
5677	>	public final Double getRawResult() { return result; }
5678	>	public final void compute() {
5679	>	final ObjectByObjectToDouble<? super K, ? super V> transformer;
5680	>	final DoubleByDoubleToDouble reducer;
5681	>	if ((transformer = this.transformer) != null &&
5682	>	(reducer = this.reducer) != null) {
5683	>	double r = this.basis;
5684	>	for (int i = baseIndex, f, h; batch > 0 &&
5685	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
5686	>	addToPendingCount(1);
5687	>	(rights = new MapReduceMappingsToDoubleTask<K,V>
5688	>	(this, batch >>>= 1, baseLimit = h, f, tab,
5689	>	rights, transformer, r, reducer)).fork();
5690	>	}
5691	>	for (Node<K,V> p; (p = advance()) != null; )
5692	>	r = reducer.apply(r, transformer.apply(p.key, p.val));
5693	>	result = r;
5694	>	CountedCompleter<?> c;
5695	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5696	>	@SuppressWarnings("unchecked") MapReduceMappingsToDoubleTask<K,V>
5697	>	t = (MapReduceMappingsToDoubleTask<K,V>)c,
5698	>	s = t.rights;
5699	>	while (s != null) {
5700	>	t.result = reducer.apply(t.result, s.result);
5701	>	s = t.rights = s.nextRight;
5702	>	}
5703	>	}
5704	>	}
5705	>	}
5706	>	}
5707	>
5708	>	@SuppressWarnings("serial")
5709	>	static final class MapReduceKeysToLongTask<K,V>
5710	>	extends BulkTask<K,V,Long> {
5711	>	final ObjectToLong<? super K> transformer;
5712	>	final LongByLongToLong reducer;
5713	>	final long basis;
5714	>	long result;
5715	>	MapReduceKeysToLongTask<K,V> rights, nextRight;
5716	>	MapReduceKeysToLongTask
5717	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
5718	>	MapReduceKeysToLongTask<K,V> nextRight,
5719	>	ObjectToLong<? super K> transformer,
5720	>	long basis,
5721	>	LongByLongToLong reducer) {
5722	>	super(p, b, i, f, t); this.nextRight = nextRight;
5723	>	this.transformer = transformer;
5724	>	this.basis = basis; this.reducer = reducer;
5725	>	}
5726	>	public final Long getRawResult() { return result; }
5727	>	public final void compute() {
5728	>	final ObjectToLong<? super K> transformer;
5729	>	final LongByLongToLong reducer;
5730	>	if ((transformer = this.transformer) != null &&
5731	>	(reducer = this.reducer) != null) {
5732	>	long r = this.basis;
5733	>	for (int i = baseIndex, f, h; batch > 0 &&
5734	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
5735	>	addToPendingCount(1);
5736	>	(rights = new MapReduceKeysToLongTask<K,V>
5737	>	(this, batch >>>= 1, baseLimit = h, f, tab,
5738	>	rights, transformer, r, reducer)).fork();
5739	>	}
5740	>	for (Node<K,V> p; (p = advance()) != null; )
5741	>	r = reducer.apply(r, transformer.apply(p.key));
5742	>	result = r;
5743	>	CountedCompleter<?> c;
5744	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5745	>	@SuppressWarnings("unchecked") MapReduceKeysToLongTask<K,V>
5746	>	t = (MapReduceKeysToLongTask<K,V>)c,
5747	>	s = t.rights;
5748	>	while (s != null) {
5749	>	t.result = reducer.apply(t.result, s.result);
5750	>	s = t.rights = s.nextRight;
5751	>	}
5752	>	}
5753	>	}
5754	>	}
5755	>	}
5756	>
5757	>	@SuppressWarnings("serial")
5758	>	static final class MapReduceValuesToLongTask<K,V>
5759	>	extends BulkTask<K,V,Long> {
5760	>	final ObjectToLong<? super V> transformer;
5761	>	final LongByLongToLong reducer;
5762	>	final long basis;
5763	>	long result;
5764	>	MapReduceValuesToLongTask<K,V> rights, nextRight;
5765	>	MapReduceValuesToLongTask
5766	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
5767	>	MapReduceValuesToLongTask<K,V> nextRight,
5768	>	ObjectToLong<? super V> transformer,
5769	>	long basis,
5770	>	LongByLongToLong reducer) {
5771	>	super(p, b, i, f, t); this.nextRight = nextRight;
5772	>	this.transformer = transformer;
5773	>	this.basis = basis; this.reducer = reducer;
5774	>	}
5775	>	public final Long getRawResult() { return result; }
5776	>	public final void compute() {
5777	>	final ObjectToLong<? super V> transformer;
5778	>	final LongByLongToLong reducer;
5779	>	if ((transformer = this.transformer) != null &&
5780	>	(reducer = this.reducer) != null) {
5781	>	long r = this.basis;
5782	>	for (int i = baseIndex, f, h; batch > 0 &&
5783	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
5784	>	addToPendingCount(1);
5785	>	(rights = new MapReduceValuesToLongTask<K,V>
5786	>	(this, batch >>>= 1, baseLimit = h, f, tab,
5787	>	rights, transformer, r, reducer)).fork();
5788	>	}
5789	>	for (Node<K,V> p; (p = advance()) != null; )
5790	>	r = reducer.apply(r, transformer.apply(p.val));
5791	>	result = r;
5792	>	CountedCompleter<?> c;
5793	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5794	>	@SuppressWarnings("unchecked") MapReduceValuesToLongTask<K,V>
5795	>	t = (MapReduceValuesToLongTask<K,V>)c,
5796	>	s = t.rights;
5797	>	while (s != null) {
5798	>	t.result = reducer.apply(t.result, s.result);
5799	>	s = t.rights = s.nextRight;
5800	>	}
5801	>	}
5802	>	}
5803	>	}
5804	>	}
5805	>
5806	>	@SuppressWarnings("serial")
5807	>	static final class MapReduceEntriesToLongTask<K,V>
5808	>	extends BulkTask<K,V,Long> {
5809	>	final ObjectToLong<Map.Entry<K,V>> transformer;
5810	>	final LongByLongToLong reducer;
5811	>	final long basis;
5812	>	long result;
5813	>	MapReduceEntriesToLongTask<K,V> rights, nextRight;
5814	>	MapReduceEntriesToLongTask
5815	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
5816	>	MapReduceEntriesToLongTask<K,V> nextRight,
5817	>	ObjectToLong<Map.Entry<K,V>> transformer,
5818	>	long basis,
5819	>	LongByLongToLong reducer) {
5820	>	super(p, b, i, f, t); this.nextRight = nextRight;
5821	>	this.transformer = transformer;
5822	>	this.basis = basis; this.reducer = reducer;
5823	>	}
5824	>	public final Long getRawResult() { return result; }
5825	>	public final void compute() {
5826	>	final ObjectToLong<Map.Entry<K,V>> transformer;
5827	>	final LongByLongToLong reducer;
5828	>	if ((transformer = this.transformer) != null &&
5829	>	(reducer = this.reducer) != null) {
5830	>	long r = this.basis;
5831	>	for (int i = baseIndex, f, h; batch > 0 &&
5832	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
5833	>	addToPendingCount(1);
5834	>	(rights = new MapReduceEntriesToLongTask<K,V>
5835	>	(this, batch >>>= 1, baseLimit = h, f, tab,
5836	>	rights, transformer, r, reducer)).fork();
5837	>	}
5838	>	for (Node<K,V> p; (p = advance()) != null; )
5839	>	r = reducer.apply(r, transformer.apply(p));
5840	>	result = r;
5841	>	CountedCompleter<?> c;
5842	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5843	>	@SuppressWarnings("unchecked") MapReduceEntriesToLongTask<K,V>
5844	>	t = (MapReduceEntriesToLongTask<K,V>)c,
5845	>	s = t.rights;
5846	>	while (s != null) {
5847	>	t.result = reducer.apply(t.result, s.result);
5848	>	s = t.rights = s.nextRight;
5849	>	}
5850	>	}
5851	>	}
5852	>	}
5853	>	}
5854	>
5855	>	@SuppressWarnings("serial")
5856	>	static final class MapReduceMappingsToLongTask<K,V>
5857	>	extends BulkTask<K,V,Long> {
5858	>	final ObjectByObjectToLong<? super K, ? super V> transformer;
5859	>	final LongByLongToLong reducer;
5860	>	final long basis;
5861	>	long result;
5862	>	MapReduceMappingsToLongTask<K,V> rights, nextRight;
5863	>	MapReduceMappingsToLongTask
5864	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
5865	>	MapReduceMappingsToLongTask<K,V> nextRight,
5866	>	ObjectByObjectToLong<? super K, ? super V> transformer,
5867	>	long basis,
5868	>	LongByLongToLong reducer) {
5869	>	super(p, b, i, f, t); this.nextRight = nextRight;
5870	>	this.transformer = transformer;
5871	>	this.basis = basis; this.reducer = reducer;
5872	>	}
5873	>	public final Long getRawResult() { return result; }
5874	>	public final void compute() {
5875	>	final ObjectByObjectToLong<? super K, ? super V> transformer;
5876	>	final LongByLongToLong reducer;
5877	>	if ((transformer = this.transformer) != null &&
5878	>	(reducer = this.reducer) != null) {
5879	>	long r = this.basis;
5880	>	for (int i = baseIndex, f, h; batch > 0 &&
5881	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
5882	>	addToPendingCount(1);
5883	>	(rights = new MapReduceMappingsToLongTask<K,V>
5884	>	(this, batch >>>= 1, baseLimit = h, f, tab,
5885	>	rights, transformer, r, reducer)).fork();
5886	>	}
5887	>	for (Node<K,V> p; (p = advance()) != null; )
5888	>	r = reducer.apply(r, transformer.apply(p.key, p.val));
5889	>	result = r;
5890	>	CountedCompleter<?> c;
5891	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5892	>	@SuppressWarnings("unchecked") MapReduceMappingsToLongTask<K,V>
5893	>	t = (MapReduceMappingsToLongTask<K,V>)c,
5894	>	s = t.rights;
5895	>	while (s != null) {
5896	>	t.result = reducer.apply(t.result, s.result);
5897	>	s = t.rights = s.nextRight;
5898	>	}
5899	>	}
5900	>	}
5901	>	}
5902	>	}
5903	>
5904	>	@SuppressWarnings("serial")
5905	>	static final class MapReduceKeysToIntTask<K,V>
5906	>	extends BulkTask<K,V,Integer> {
5907	>	final ObjectToInt<? super K> transformer;
5908	>	final IntByIntToInt reducer;
5909	>	final int basis;
5910	>	int result;
5911	>	MapReduceKeysToIntTask<K,V> rights, nextRight;
5912	>	MapReduceKeysToIntTask
5913	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
5914	>	MapReduceKeysToIntTask<K,V> nextRight,
5915	>	ObjectToInt<? super K> transformer,
5916	>	int basis,
5917	>	IntByIntToInt reducer) {
5918	>	super(p, b, i, f, t); this.nextRight = nextRight;
5919	>	this.transformer = transformer;
5920	>	this.basis = basis; this.reducer = reducer;
5921	>	}
5922	>	public final Integer getRawResult() { return result; }
5923	>	public final void compute() {
5924	>	final ObjectToInt<? super K> transformer;
5925	>	final IntByIntToInt reducer;
5926	>	if ((transformer = this.transformer) != null &&
5927	>	(reducer = this.reducer) != null) {
5928	>	int r = this.basis;
5929	>	for (int i = baseIndex, f, h; batch > 0 &&
5930	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
5931	>	addToPendingCount(1);
5932	>	(rights = new MapReduceKeysToIntTask<K,V>
5933	>	(this, batch >>>= 1, baseLimit = h, f, tab,
5934	>	rights, transformer, r, reducer)).fork();
5935	>	}
5936	>	for (Node<K,V> p; (p = advance()) != null; )
5937	>	r = reducer.apply(r, transformer.apply(p.key));
5938	>	result = r;
5939	>	CountedCompleter<?> c;
5940	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5941	>	@SuppressWarnings("unchecked") MapReduceKeysToIntTask<K,V>
5942	>	t = (MapReduceKeysToIntTask<K,V>)c,
5943	>	s = t.rights;
5944	>	while (s != null) {
5945	>	t.result = reducer.apply(t.result, s.result);
5946	>	s = t.rights = s.nextRight;
5947	>	}
5948	>	}
5949	>	}
5950	>	}
5951	>	}
5952	>
5953	>	@SuppressWarnings("serial")
5954	>	static final class MapReduceValuesToIntTask<K,V>
5955	>	extends BulkTask<K,V,Integer> {
5956	>	final ObjectToInt<? super V> transformer;
5957	>	final IntByIntToInt reducer;
5958	>	final int basis;
5959	>	int result;
5960	>	MapReduceValuesToIntTask<K,V> rights, nextRight;
5961	>	MapReduceValuesToIntTask
5962	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
5963	>	MapReduceValuesToIntTask<K,V> nextRight,
5964	>	ObjectToInt<? super V> transformer,
5965	>	int basis,
5966	>	IntByIntToInt reducer) {
5967	>	super(p, b, i, f, t); this.nextRight = nextRight;
5968	>	this.transformer = transformer;
5969	>	this.basis = basis; this.reducer = reducer;
5970	>	}
5971	>	public final Integer getRawResult() { return result; }
5972	>	public final void compute() {
5973	>	final ObjectToInt<? super V> transformer;
5974	>	final IntByIntToInt reducer;
5975	>	if ((transformer = this.transformer) != null &&
5976	>	(reducer = this.reducer) != null) {
5977	>	int r = this.basis;
5978	>	for (int i = baseIndex, f, h; batch > 0 &&
5979	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
5980	>	addToPendingCount(1);
5981	>	(rights = new MapReduceValuesToIntTask<K,V>
5982	>	(this, batch >>>= 1, baseLimit = h, f, tab,
5983	>	rights, transformer, r, reducer)).fork();
5984	>	}
5985	>	for (Node<K,V> p; (p = advance()) != null; )
5986	>	r = reducer.apply(r, transformer.apply(p.val));
5987	>	result = r;
5988	>	CountedCompleter<?> c;
5989	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
5990	>	@SuppressWarnings("unchecked") MapReduceValuesToIntTask<K,V>
5991	>	t = (MapReduceValuesToIntTask<K,V>)c,
5992	>	s = t.rights;
5993	>	while (s != null) {
5994	>	t.result = reducer.apply(t.result, s.result);
5995	>	s = t.rights = s.nextRight;
5996	>	}
5997	>	}
5998	>	}
5999	>	}
6000	>	}
6001	>
6002	>	@SuppressWarnings("serial")
6003	>	static final class MapReduceEntriesToIntTask<K,V>
6004	>	extends BulkTask<K,V,Integer> {
6005	>	final ObjectToInt<Map.Entry<K,V>> transformer;
6006	>	final IntByIntToInt reducer;
6007	>	final int basis;
6008	>	int result;
6009	>	MapReduceEntriesToIntTask<K,V> rights, nextRight;
6010	>	MapReduceEntriesToIntTask
6011	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
6012	>	MapReduceEntriesToIntTask<K,V> nextRight,
6013	>	ObjectToInt<Map.Entry<K,V>> transformer,
6014	>	int basis,
6015	>	IntByIntToInt reducer) {
6016	>	super(p, b, i, f, t); this.nextRight = nextRight;
6017	>	this.transformer = transformer;
6018	>	this.basis = basis; this.reducer = reducer;
6019	>	}
6020	>	public final Integer getRawResult() { return result; }
6021	>	public final void compute() {
6022	>	final ObjectToInt<Map.Entry<K,V>> transformer;
6023	>	final IntByIntToInt reducer;
6024	>	if ((transformer = this.transformer) != null &&
6025	>	(reducer = this.reducer) != null) {
6026	>	int r = this.basis;
6027	>	for (int i = baseIndex, f, h; batch > 0 &&
6028	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
6029	>	addToPendingCount(1);
6030	>	(rights = new MapReduceEntriesToIntTask<K,V>
6031	>	(this, batch >>>= 1, baseLimit = h, f, tab,
6032	>	rights, transformer, r, reducer)).fork();
6033	>	}
6034	>	for (Node<K,V> p; (p = advance()) != null; )
6035	>	r = reducer.apply(r, transformer.apply(p));
6036	>	result = r;
6037	>	CountedCompleter<?> c;
6038	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6039	>	@SuppressWarnings("unchecked") MapReduceEntriesToIntTask<K,V>
6040	>	t = (MapReduceEntriesToIntTask<K,V>)c,
6041	>	s = t.rights;
6042	>	while (s != null) {
6043	>	t.result = reducer.apply(t.result, s.result);
6044	>	s = t.rights = s.nextRight;
6045	>	}
6046	>	}
6047	>	}
6048	>	}
6049	>	}
6050	>
6051	>	@SuppressWarnings("serial")
6052	>	static final class MapReduceMappingsToIntTask<K,V>
6053	>	extends BulkTask<K,V,Integer> {
6054	>	final ObjectByObjectToInt<? super K, ? super V> transformer;
6055	>	final IntByIntToInt reducer;
6056	>	final int basis;
6057	>	int result;
6058	>	MapReduceMappingsToIntTask<K,V> rights, nextRight;
6059	>	MapReduceMappingsToIntTask
6060	>	(BulkTask<K,V,?> p, int b, int i, int f, Node<K,V>[] t,
6061	>	MapReduceMappingsToIntTask<K,V> nextRight,
6062	>	ObjectByObjectToInt<? super K, ? super V> transformer,
6063	>	int basis,
6064	>	IntByIntToInt reducer) {
6065	>	super(p, b, i, f, t); this.nextRight = nextRight;
6066	>	this.transformer = transformer;
6067	>	this.basis = basis; this.reducer = reducer;
6068	>	}
6069	>	public final Integer getRawResult() { return result; }
6070	>	public final void compute() {
6071	>	final ObjectByObjectToInt<? super K, ? super V> transformer;
6072	>	final IntByIntToInt reducer;
6073	>	if ((transformer = this.transformer) != null &&
6074	>	(reducer = this.reducer) != null) {
6075	>	int r = this.basis;
6076	>	for (int i = baseIndex, f, h; batch > 0 &&
6077	>	(h = ((f = baseLimit) + i) >>> 1) > i;) {
6078	>	addToPendingCount(1);
6079	>	(rights = new MapReduceMappingsToIntTask<K,V>
6080	>	(this, batch >>>= 1, baseLimit = h, f, tab,
6081	>	rights, transformer, r, reducer)).fork();
6082	>	}
6083	>	for (Node<K,V> p; (p = advance()) != null; )
6084	>	r = reducer.apply(r, transformer.apply(p.key, p.val));
6085	>	result = r;
6086	>	CountedCompleter<?> c;
6087	>	for (c = firstComplete(); c != null; c = c.nextComplete()) {
6088	>	@SuppressWarnings("unchecked") MapReduceMappingsToIntTask<K,V>
6089	>	t = (MapReduceMappingsToIntTask<K,V>)c,
6090	>	s = t.rights;
6091	>	while (s != null) {
6092	>	t.result = reducer.apply(t.result, s.result);
6093	>	s = t.rights = s.nextRight;
6094	>	}
6095	>	}
6096	>	}
6097	>	}
6098	>	}
6099	>
6100	>	/* ---------------- Counters -------------- */
6101	>
6102	>	// Adapted from LongAdder and Striped64.
6103	>	// See their internal docs for explanation.
6104	>
6105	>	// A padded cell for distributing counts
6106	>	static final class CounterCell {
6107	>	volatile long p0, p1, p2, p3, p4, p5, p6;
6108	>	volatile long value;
6109	>	volatile long q0, q1, q2, q3, q4, q5, q6;
6110	>	CounterCell(long x) { value = x; }
6111	>	}
6112
6113		/**
6114	<	* Helper class used in previous version, declared for the sake of
6115	<	* serialization compatibility
6114	>	* Holder for the thread-local hash code determining which
6115	>	* CounterCell to use. The code is initialized via the
6116	>	* counterHashCodeGenerator, but may be moved upon collisions.
6117		*/
6118	<	static class Segment<K,V> extends java.util.concurrent.locks.ReentrantLock
6119	<	implements Serializable {
1515	<	private static final long serialVersionUID = 2249069246763182397L;
1516	<	final float loadFactor;
1517	<	Segment(float lf) { this.loadFactor = lf; }
6118	>	static final class CounterHashCode {
6119	>	int code;
6120		}
6121
6122		/**
6123	<	* Saves the state of the {@code ConcurrentHashMapV8} instance to a
1522	<	* stream (i.e., serializes it).
1523	<	* @param s the stream
1524	<	* @serialData
1525	<	* the key (Object) and value (Object)
1526	<	* for each key-value mapping, followed by a null pair.
1527	<	* The key-value mappings are emitted in no particular order.
6123	>	* Generates initial value for per-thread CounterHashCodes.
6124		*/
6125	<	@SuppressWarnings("unchecked")
1530	<	private void writeObject(java.io.ObjectOutputStream s)
1531	<	throws java.io.IOException {
1532	<	if (segments == null) { // for serialization compatibility
1533	<	segments = (Segment<K,V>[])
1534	<	new Segment<?,?>[DEFAULT_CONCURRENCY_LEVEL];
1535	<	for (int i = 0; i < segments.length; ++i)
1536	<	segments[i] = new Segment<K,V>(loadFactor);
1537	<	}
1538	<	s.defaultWriteObject();
1539	<	new HashIterator().writeEntries(s);
1540	<	s.writeObject(null);
1541	<	s.writeObject(null);
1542	<	segments = null; // throw away
1543	<	}
6125	>	static final AtomicInteger counterHashCodeGenerator = new AtomicInteger();
6126
6127		/**
6128	<	* Reconstitutes the instance from a stream (that is, deserializes it).
6129	<	* @param s the stream
6128	>	* Increment for counterHashCodeGenerator. See class ThreadLocal
6129	>	* for explanation.
6130		*/
6131	<	@SuppressWarnings("unchecked")
6132	<	private void readObject(java.io.ObjectInputStream s)
6133	<	throws java.io.IOException, ClassNotFoundException {
6134	<	s.defaultReadObject();
6135	<	// find load factor in a segment, if one exists
6136	<	if (segments != null && segments.length != 0)
6137	<	this.loadFactor = segments[0].loadFactor;
6138	<	else
1557	<	this.loadFactor = DEFAULT_LOAD_FACTOR;
1558	<	this.initCap = DEFAULT_CAPACITY;
1559	<	LongAdder ct = new LongAdder(); // force final field write
1560	<	UNSAFE.putObjectVolatile(this, counterOffset, ct);
1561	<	this.segments = null; // unneeded
6131	>	static final int SEED_INCREMENT = 0x61c88647;
6132	>
6133	>	/**
6134	>	* Per-thread counter hash codes. Shared across all instances.
6135	>	*/
6136	>	static final ThreadLocal<CounterHashCode> threadCounterHashCode =
6137	>	new ThreadLocal<CounterHashCode>();
6138	>
6139
6140	<	// Read the keys and values, and put the mappings in the table
6140	>	final long sumCount() {
6141	>	CounterCell[] as = counterCells; CounterCell a;
6142	>	long sum = baseCount;
6143	>	if (as != null) {
6144	>	for (int i = 0; i < as.length; ++i) {
6145	>	if ((a = as[i]) != null)
6146	>	sum += a.value;
6147	>	}
6148	>	}
6149	>	return sum;
6150	>	}
6151	>
6152	>	// See LongAdder version for explanation
6153	>	private final void fullAddCount(long x, CounterHashCode hc,
6154	>	boolean wasUncontended) {
6155	>	int h;
6156	>	if (hc == null) {
6157	>	hc = new CounterHashCode();
6158	>	int s = counterHashCodeGenerator.addAndGet(SEED_INCREMENT);
6159	>	h = hc.code = (s == 0) ? 1 : s; // Avoid zero
6160	>	threadCounterHashCode.set(hc);
6161	>	}
6162	>	else
6163	>	h = hc.code;
6164	>	boolean collide = false;                // True if last slot nonempty
6165		for (;;) {
6166	<	K key = (K) s.readObject();
6167	<	V value = (V) s.readObject();
6168	<	if (key == null)
6169	<	break;
6170	<	put(key, value);
6166	>	CounterCell[] as; CounterCell a; int n; long v;
6167	>	if ((as = counterCells) != null && (n = as.length) > 0) {
6168	>	if ((a = as[(n - 1) & h]) == null) {
6169	>	if (cellsBusy == 0) {            // Try to attach new Cell
6170	>	CounterCell r = new CounterCell(x); // Optimistic create
6171	>	if (cellsBusy == 0 &&
6172	>	U.compareAndSwapInt(this, CELLSBUSY, 0, 1)) {
6173	>	boolean created = false;
6174	>	try {               // Recheck under lock
6175	>	CounterCell[] rs; int m, j;
6176	>	if ((rs = counterCells) != null &&
6177	>	(m = rs.length) > 0 &&
6178	>	rs[j = (m - 1) & h] == null) {
6179	>	rs[j] = r;
6180	>	created = true;
6181	>	}
6182	>	} finally {
6183	>	cellsBusy = 0;
6184	>	}
6185	>	if (created)
6186	>	break;
6187	>	continue;           // Slot is now non-empty
6188	>	}
6189	>	}
6190	>	collide = false;
6191	>	}
6192	>	else if (!wasUncontended)       // CAS already known to fail
6193	>	wasUncontended = true;      // Continue after rehash
6194	>	else if (U.compareAndSwapLong(a, CELLVALUE, v = a.value, v + x))
6195	>	break;
6196	>	else if (counterCells != as || n >= NCPU)
6197	>	collide = false;            // At max size or stale
6198	>	else if (!collide)
6199	>	collide = true;
6200	>	else if (cellsBusy == 0 &&
6201	>	U.compareAndSwapInt(this, CELLSBUSY, 0, 1)) {
6202	>	try {
6203	>	if (counterCells == as) {// Expand table unless stale
6204	>	CounterCell[] rs = new CounterCell[n << 1];
6205	>	for (int i = 0; i < n; ++i)
6206	>	rs[i] = as[i];
6207	>	counterCells = rs;
6208	>	}
6209	>	} finally {
6210	>	cellsBusy = 0;
6211	>	}
6212	>	collide = false;
6213	>	continue;                   // Retry with expanded table
6214	>	}
6215	>	h ^= h << 13;                   // Rehash
6216	>	h ^= h >>> 17;
6217	>	h ^= h << 5;
6218	>	}
6219	>	else if (cellsBusy == 0 && counterCells == as &&
6220	>	U.compareAndSwapInt(this, CELLSBUSY, 0, 1)) {
6221	>	boolean init = false;
6222	>	try {                           // Initialize table
6223	>	if (counterCells == as) {
6224	>	CounterCell[] rs = new CounterCell[2];
6225	>	rs[h & 1] = new CounterCell(x);
6226	>	counterCells = rs;
6227	>	init = true;
6228	>	}
6229	>	} finally {
6230	>	cellsBusy = 0;
6231	>	}
6232	>	if (init)
6233	>	break;
6234	>	}
6235	>	else if (U.compareAndSwapLong(this, BASECOUNT, v = baseCount, v + x))
6236	>	break;                          // Fall back on using base
6237		}
6238	+	hc.code = h;                            // Record index for next time
6239		}
6240
6241		// Unsafe mechanics
6242	<	private static final sun.misc.Unsafe UNSAFE;
6243	<	private static final long counterOffset;
6244	<	private static final long resizingOffset;
6242	>	private static final sun.misc.Unsafe U;
6243	>	private static final long SIZECTL;
6244	>	private static final long TRANSFERINDEX;
6245	>	private static final long BASECOUNT;
6246	>	private static final long CELLSBUSY;
6247	>	private static final long CELLVALUE;
6248		private static final long ABASE;
6249		private static final int ASHIFT;
6250
6251		static {
1581	–	int ss;
6252		try {
6253	<	UNSAFE = getUnsafe();
6253	>	U = getUnsafe();
6254		Class<?> k = ConcurrentHashMapV8.class;
6255	<	counterOffset = UNSAFE.objectFieldOffset
6256	<	(k.getDeclaredField("counter"));
6257	<	resizingOffset = UNSAFE.objectFieldOffset
6258	<	(k.getDeclaredField("resizing"));
6259	<	Class<?> sc = Node[].class;
6260	<	ABASE = UNSAFE.arrayBaseOffset(sc);
6261	<	ss = UNSAFE.arrayIndexScale(sc);
6255	>	SIZECTL = U.objectFieldOffset
6256	>	(k.getDeclaredField("sizeCtl"));
6257	>	TRANSFERINDEX = U.objectFieldOffset
6258	>	(k.getDeclaredField("transferIndex"));
6259	>	BASECOUNT = U.objectFieldOffset
6260	>	(k.getDeclaredField("baseCount"));
6261	>	CELLSBUSY = U.objectFieldOffset
6262	>	(k.getDeclaredField("cellsBusy"));
6263	>	Class<?> ck = CounterCell.class;
6264	>	CELLVALUE = U.objectFieldOffset
6265	>	(ck.getDeclaredField("value"));
6266	>	Class<?> ak = Node[].class;
6267	>	ABASE = U.arrayBaseOffset(ak);
6268	>	int scale = U.arrayIndexScale(ak);
6269	>	if ((scale & (scale - 1)) != 0)
6270	>	throw new Error("data type scale not a power of two");
6271	>	ASHIFT = 31 - Integer.numberOfLeadingZeros(scale);
6272		} catch (Exception e) {
6273		throw new Error(e);
6274		}
1595	–	if ((ss & (ss-1)) != 0)
1596	–	throw new Error("data type scale not a power of two");
1597	–	ASHIFT = 31 - Integer.numberOfLeadingZeros(ss);
6275		}
6276
6277		/**
#	Line 1607 | Line 6284 | public class ConcurrentHashMapV8<K, V>
6284		private static sun.misc.Unsafe getUnsafe() {
6285		try {
6286		return sun.misc.Unsafe.getUnsafe();
6287	<	} catch (SecurityException se) {
6288	<	try {
6289	<	return java.security.AccessController.doPrivileged
6290	<	(new java.security
6291	<	.PrivilegedExceptionAction<sun.misc.Unsafe>() {
6292	<	public sun.misc.Unsafe run() throws Exception {
6293	<	java.lang.reflect.Field f = sun.misc
6294	<	.Unsafe.class.getDeclaredField("theUnsafe");
6295	<	f.setAccessible(true);
6296	<	return (sun.misc.Unsafe) f.get(null);
6297	<	}});
6298	<	} catch (java.security.PrivilegedActionException e) {
6299	<	throw new RuntimeException("Could not initialize intrinsics",
6300	<	e.getCause());
6301	<	}
6287	>	} catch (SecurityException tryReflectionInstead) {}
6288	>	try {
6289	>	return java.security.AccessController.doPrivileged
6290	>	(new java.security.PrivilegedExceptionAction<sun.misc.Unsafe>() {
6291	>	public sun.misc.Unsafe run() throws Exception {
6292	>	Class<sun.misc.Unsafe> k = sun.misc.Unsafe.class;
6293	>	for (java.lang.reflect.Field f : k.getDeclaredFields()) {
6294	>	f.setAccessible(true);
6295	>	Object x = f.get(null);
6296	>	if (k.isInstance(x))
6297	>	return k.cast(x);
6298	>	}
6299	>	throw new NoSuchFieldError("the Unsafe");
6300	>	}});
6301	>	} catch (java.security.PrivilegedActionException e) {
6302	>	throw new RuntimeException("Could not initialize intrinsics",
6303	>	e.getCause());
6304		}
6305		}
1627	–
6306		}

Diff Legend

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