Start line:  
End line:  

Snippet Preview

Snippet HTML Code

Stack Overflow Questions
   // Revision 1.120
   
   /*
    * Written by Doug Lea with assistance from members of JCP JSR-166
    * Expert Group and released to the public domain, as explained at
    * http://creativecommons.org/publicdomain/zero/1.0/
    */
   
   package org.infinispan.commons.util.concurrent.jdk8backported;
  
  
  import java.util.*;
A hash table supporting full concurrency of retrievals and high expected concurrency for updates. This class obeys the same functional specification as java.util.Hashtable, and includes versions of methods corresponding to each method of Hashtable. However, even though all operations are thread-safe, retrieval operations do not entail locking, and there is not any support for locking the entire table in a way that prevents all access. This class is fully interoperable with Hashtable in programs that rely on its thread safety but not on its synchronization details.

Retrieval operations (including get) generally do not block, so may overlap with update operations (including put and remove). Retrievals reflect the results of the most recently completed update operations holding upon their onset. (More formally, an update operation for a given key bears a happens-before relation with any (non-null) retrieval for that key reporting the updated value.) For aggregate operations such as putAll and clear, concurrent retrievals may reflect insertion or removal of only some entries. Similarly, Iterators and Enumerations return elements reflecting the state of the hash table at some point at or since the creation of the iterator/enumeration. They do not throw java.util.ConcurrentModificationException. However, iterators are designed to be used by only one thread at a time. Bear in mind that the results of aggregate status methods including size, isEmpty, and containsValue are typically useful only when a map is not undergoing concurrent updates in other threads. Otherwise the results of these methods reflect transient states that may be adequate for monitoring or estimation purposes, but not for program control.

The table is dynamically expanded when there are too many collisions (i.e., keys that have distinct hash codes but fall into the same slot modulo the table size), with the expected average effect of maintaining roughly two bins per mapping (corresponding to a 0.75 load factor threshold for resizing). There may be much variance around this average as mappings are added and removed, but overall, this maintains a commonly accepted time/space tradeoff for hash tables. However, resizing this or any other kind of hash table may be a relatively slow operation. When possible, it is a good idea to provide a size estimate as an optional initialCapacity constructor argument. An additional optional loadFactor constructor argument provides a further means of customizing initial table capacity by specifying the table density to be used in calculating the amount of space to allocate for the given number of elements. Also, for compatibility with previous versions of this class, constructors may optionally specify an expected concurrencyLevel as an additional hint for internal sizing. Note that using many keys with exactly the same hashCode() is a sure way to slow down performance of any hash table. To ameliorate impact, when keys are java.lang.Comparable, this class may use comparison order among keys to help break ties.

A java.util.Set projection of a EquivalentConcurrentHashMapV8 may be created (using newKeySet(org.infinispan.commons.equivalence.Equivalence) or newKeySet(int,org.infinispan.commons.equivalence.Equivalence)), or viewed (using keySet(java.lang.Object) when only keys are of interest, and the mapped values are (perhaps transiently) not used or all take the same mapping value.

This class and its views and iterators implement all of the optional methods of the java.util.Map and java.util.Iterator interfaces.

Like java.util.Hashtable but unlike java.util.HashMap, this class does not allow null to be used as a key or value.

ConcurrentHashMapV8s support a set of sequential and parallel bulk operations that are designed to be safely, and often sensibly, applied even with maps that are being concurrently updated by other threads; for example, when computing a snapshot summary of the values in a shared registry. There are three kinds of operation, each with four forms, accepting functions with Keys, Values, Entries, and (Key, Value) arguments and/or return values. Because the elements of a EquivalentConcurrentHashMapV8 are not ordered in any particular way, and may be processed in different orders in different parallel executions, the correctness of supplied functions should not depend on any ordering, or on any other objects or values that may transiently change while computation is in progress; and except for forEach actions, should ideally be side-effect-free. Bulk operations on java.util.Map.Entry objects do not support method setValue.

  • forEach: Perform a given action on each element. A variant form applies a given transformation on each element before performing the action.
  • search: Return the first available non-null result of applying a given function on each element; skipping further search when a result is found.
  • reduce: Accumulate each element. The supplied reduction function cannot rely on ordering (more formally, it should be both associative and commutative). There are five variants:
    • Plain reductions. (There is not a form of this method for (key, value) function arguments since there is no corresponding return type.)
    • Mapped reductions that accumulate the results of a given function applied to each element.
    • Reductions to scalar doubles, longs, and ints, using a given basis value.

These bulk operations accept a parallelismThreshold argument. Methods proceed sequentially if the current map size is estimated to be less than the given threshold. Using a value of Long.MAX_VALUE suppresses all parallelism. Using a value of 1 results in maximal parallelism by partitioning into enough subtasks to fully utilize the ForkJoinPool.commonPool() that is used for all parallel computations. Normally, you would initially choose one of these extreme values, and then measure performance of using in-between values that trade off overhead versus throughput.

The concurrency properties of bulk operations follow from those of EquivalentConcurrentHashMapV8: Any non-null result returned from get(key) and related access methods bears a happens-before relation with the associated insertion or update. The result of any bulk operation reflects the composition of these per-element relations (but is not necessarily atomic with respect to the map as a whole unless it is somehow known to be quiescent). Conversely, because keys and values in the map are never null, null serves as a reliable atomic indicator of the current lack of any result. To maintain this property, null serves as an implicit basis for all non-scalar reduction operations. For the double, long, and int versions, the basis should be one that, when combined with any other value, returns that other value (more formally, it should be the identity element for the reduction). Most common reductions have these properties; for example, computing a sum with basis 0 or a minimum with basis MAX_VALUE.

Search and transformation functions provided as arguments should similarly return null to indicate the lack of any result (in which case it is not used). In the case of mapped reductions, this also enables transformations to serve as filters, returning null (or, in the case of primitive specializations, the identity basis) if the element should not be combined. You can create compound transformations and filterings by composing them yourself under this "null means there is nothing there now" rule before using them in search or reduce operations.

Methods accepting and/or returning Entry arguments maintain key-value associations. They may be useful for example when finding the key for the greatest value. Note that "plain" Entry arguments can be supplied using new AbstractMap.SimpleEntry(k,v).

Bulk operations may complete abruptly, throwing an exception encountered in the application of a supplied function. Bear in mind when handling such exceptions that other concurrently executing functions could also have thrown exceptions, or would have done so if the first exception had not occurred.

Speedups for parallel compared to sequential forms are common but not guaranteed. Parallel operations involving brief functions on small maps may execute more slowly than sequential forms if the underlying work to parallelize the computation is more expensive than the computation itself. Similarly, parallelization may not lead to much actual parallelism if all processors are busy performing unrelated tasks.

All arguments to all task methods must be non-null.

jsr166e note: During transition, this class uses nested functional interfaces with different names but the same forms as those expected for JDK8.

This class is a member of the Java Collections Framework. NOTE: This map has been tweaked so that equality and hash code calculations are done based on a passed org.infinispan.commons.equivalence.Equivalence function implementation for keys and values, as opposed to relying on their own equals/hashCode/toString implementations. This is handy when using key/values whose mentioned methods cannot be overriden, i.e. arrays, and in situations where users want to avoid using wrapper objects. To help with future revisions of this class, changes other than constructor changes have been marked with 'EQUIVALENCE_MOD' comment.

Parameters:
<K> the type of keys maintained by this map
<V> the type of mapped values
Author(s):
Doug Lea
Galder ZamarreƱo
Since:
1.5
 
 @SuppressWarnings("restriction")
 public class EquivalentConcurrentHashMapV8<K,V> extends AbstractMap<K,V>
       implements ConcurrentMap<K,V>, Serializable {
    private static final long serialVersionUID = 7249069246763182397L;

   
An object for traversing and partitioning elements of a source. This interface provides a subset of the functionality of JDK8 java.util.Spliterator.
 
    public static interface ConcurrentHashMapSpliterator<T> {
      
If possible, returns a new spliterator covering approximately one half of the elements, which will not be covered by this spliterator. Returns null if cannot be split.
 
       ConcurrentHashMapSpliterator<T> trySplit();
      
Returns an estimate of the number of elements covered by this Spliterator.
 
       long estimateSize();

      
Applies the action to each untraversed element
 
       void forEachRemaining(Action<? super T> action);
      
If an element remains, applies the action and returns true.
 
       boolean tryAdvance(Action<? super T> action);
    }
 
    // Sams
    
Interface describing a void action of one argument
 
    public interface Action<A> { void apply(A a); }
   
Interface describing a void action of two arguments
 
    public interface BiAction<A,B> { void apply(A a, B b); }
   
Interface describing a function of one argument
 
    public interface Fun<A,T> { T apply(A a); }
   
Interface describing a function of two arguments
 
    public interface BiFun<A,B,T> { T apply(A a, B b); }
   
Interface describing a function mapping its argument to a double
 
    public interface ObjectToDouble<A> { double apply(A a); }
   
Interface describing a function mapping its argument to a long
 
    public interface ObjectToLong<A> { long apply(A a); }
   
Interface describing a function mapping its argument to an int
 
    public interface ObjectToInt<A> {int apply(A a); }
   
Interface describing a function mapping two arguments to a double
 
    public interface ObjectByObjectToDouble<A,B> { double apply(A a, B b); }
   
Interface describing a function mapping two arguments to a long
 
    public interface ObjectByObjectToLong<A,B> { long apply(A a, B b); }
   
Interface describing a function mapping two arguments to an int
 
    public interface ObjectByObjectToInt<A,B> {int apply(A a, B b); }
   
Interface describing a function mapping two doubles to a double
 
    public interface DoubleByDoubleToDouble { double apply(double adouble b); }
   
Interface describing a function mapping two longs to a long
 
    public interface LongByLongToLong { long apply(long along b); }
   
Interface describing a function mapping two ints to an int
 
    public interface IntByIntToInt { int apply(int aint b); }
 
 
     /*
      * Overview:
      *
      * The primary design goal of this hash table is to maintain
      * concurrent readability (typically method get(), but also
      * iterators and related methods) while minimizing update
      * contention. Secondary goals are to keep space consumption about
      * the same or better than java.util.HashMap, and to support high
      * initial insertion rates on an empty table by many threads.
      *
      * This map usually acts as a binned (bucketed) hash table.  Each
      * key-value mapping is held in a Node.  Most nodes are instances
      * of the basic Node class with hash, key, value, and next
      * fields. However, various subclasses exist: TreeNodes are
      * arranged in balanced trees, not lists.  TreeBins hold the roots
      * of sets of TreeNodes. ForwardingNodes are placed at the heads
      * of bins during resizing. ReservationNodes are used as
      * placeholders while establishing values in computeIfAbsent and
      * related methods.  The types TreeBin, ForwardingNode, and
      * ReservationNode do not hold normal user keys, values, or
      * hashes, and are readily distinguishable during search etc
      * because they have negative hash fields and null key and value
      * fields. (These special nodes are either uncommon or transient,
      * so the impact of carrying around some unused fields is
      * insignificant.)
      *
      * The table is lazily initialized to a power-of-two size upon the
      * first insertion.  Each bin in the table normally contains a
      * list of Nodes (most often, the list has only zero or one Node).
      * Table accesses require volatile/atomic reads, writes, and
      * CASes.  Because there is no other way to arrange this without
      * adding further indirections, we use intrinsics
      * (sun.misc.Unsafe) operations.
      *
      * We use the top (sign) bit of Node hash fields for control
      * purposes -- it is available anyway because of addressing
      * constraints.  Nodes with negative hash fields are specially
      * handled or ignored in map methods.
      *
      * Insertion (via put or its variants) of the first node in an
      * empty bin is performed by just CASing it to the bin.  This is
      * by far the most common case for put operations under most
      * key/hash distributions.  Other update operations (insert,
      * delete, and replace) require locks.  We do not want to waste
      * the space required to associate a distinct lock object with
      * each bin, so instead use the first node of a bin list itself as
      * a lock. Locking support for these locks relies on builtin
      * "synchronized" monitors.
      *
      * Using the first node of a list as a lock does not by itself
      * suffice though: When a node is locked, any update must first
      * validate that it is still the first node after locking it, and
      * retry if not. Because new nodes are always appended to lists,
      * once a node is first in a bin, it remains first until deleted
      * or the bin becomes invalidated (upon resizing).
      *
      * The main disadvantage of per-bin locks is that other update
      * operations on other nodes in a bin list protected by the same
      * lock can stall, for example when user equals() or mapping
      * functions take a long time.  However, statistically, under
      * random hash codes, this is not a common problem.  Ideally, the
      * frequency of nodes in bins follows a Poisson distribution
      * (http://en.wikipedia.org/wiki/Poisson_distribution) with a
      * parameter of about 0.5 on average, given the resizing threshold
      * of 0.75, although with a large variance because of resizing
      * granularity. Ignoring variance, the expected occurrences of
      * list size k are (exp(-0.5) * pow(0.5, k) / factorial(k)). The
      * first values are:
      *
      * 0:    0.60653066
      * 1:    0.30326533
      * 2:    0.07581633
      * 3:    0.01263606
      * 4:    0.00157952
      * 5:    0.00015795
      * 6:    0.00001316
      * 7:    0.00000094
      * 8:    0.00000006
      * more: less than 1 in ten million
      *
      * Lock contention probability for two threads accessing distinct
      * elements is roughly 1 / (8 * #elements) under random hashes.
      *
      * Actual hash code distributions encountered in practice
      * sometimes deviate significantly from uniform randomness.  This
      * includes the case when N > (1<<30), so some keys MUST collide.
      * Similarly for dumb or hostile usages in which multiple keys are
      * designed to have identical hash codes or ones that differs only
      * in masked-out high bits. So we use a secondary strategy that
      * applies when the number of nodes in a bin exceeds a
      * threshold. These TreeBins use a balanced tree to hold nodes (a
      * specialized form of red-black trees), bounding search time to
      * O(log N).  Each search step in a TreeBin is at least twice as
      * slow as in a regular list, but given that N cannot exceed
      * (1<<64) (before running out of addresses) this bounds search
      * steps, lock hold times, etc, to reasonable constants (roughly
      * 100 nodes inspected per operation worst case) so long as keys
      * are Comparable (which is very common -- String, Long, etc).
      * TreeBin nodes (TreeNodes) also maintain the same "next"
      * traversal pointers as regular nodes, so can be traversed in
      * iterators in the same way.
      *
      * The table is resized when occupancy exceeds a percentage
      * threshold (nominally, 0.75, but see below).  Any thread
      * noticing an overfull bin may assist in resizing after the
      * initiating thread allocates and sets up the replacement array.
      * However, rather than stalling, these other threads may proceed
      * with insertions etc.  The use of TreeBins shields us from the
      * worst case effects of overfilling while resizes are in
      * progress.  Resizing proceeds by transferring bins, one by one,
      * from the table to the next table. However, threads claim small
      * blocks of indices to transfer (via field transferIndex) before
      * doing so, reducing contention.  A generation stamp in field
      * sizeCtl ensures that resizings do not overlap. Because we are
      * using power-of-two expansion, the elements from each bin must
      * either stay at same index, or move with a power of two
      * offset. We eliminate unnecessary node creation by catching
      * cases where old nodes can be reused because their next fields
      * won't change.  On average, only about one-sixth of them need
      * cloning when a table doubles. The nodes they replace will be
      * garbage collectable as soon as they are no longer referenced by
      * any reader thread that may be in the midst of concurrently
      * traversing table.  Upon transfer, the old table bin contains
      * only a special forwarding node (with hash field "MOVED") that
      * contains the next table as its key. On encountering a
      * forwarding node, access and update operations restart, using
      * the new table.
      *
      * Each bin transfer requires its bin lock, which can stall
      * waiting for locks while resizing. However, because other
      * threads can join in and help resize rather than contend for
      * locks, average aggregate waits become shorter as resizing
      * progresses.  The transfer operation must also ensure that all
      * accessible bins in both the old and new table are usable by any
      * traversal.  This is arranged in part by proceeding from the
      * last bin (table.length - 1) up towards the first.  Upon seeing
      * a forwarding node, traversals (see class Traverser) arrange to
      * move to the new table without revisiting nodes.  To ensure that
      * no intervening nodes are skipped even when moved out of order,
      * a stack (see class TableStack) is created on first encounter of
      * a forwarding node during a traversal, to maintain its place if
      * later processing the current table. The need for these
      * save/restore mechanics is relatively rare, but when one
      * forwarding node is encountered, typically many more will be.
      * So Traversers use a simple caching scheme to avoid creating so
      * many new TableStack nodes. (Thanks to Peter Levart for
      * suggesting use of a stack here.)
      *
      * The traversal scheme also applies to partial traversals of
      * ranges of bins (via an alternate Traverser constructor)
      * to support partitioned aggregate operations.  Also, read-only
      * operations give up if ever forwarded to a null table, which
      * provides support for shutdown-style clearing, which is also not
      * currently implemented.
      *
      * Lazy table initialization minimizes footprint until first use,
      * and also avoids resizings when the first operation is from a
      * putAll, constructor with map argument, or deserialization.
      * These cases attempt to override the initial capacity settings,
      * but harmlessly fail to take effect in cases of races.
      *
      * The element count is maintained using a specialization of
      * LongAdder. We need to incorporate a specialization rather than
      * just use a LongAdder in order to access implicit
      * contention-sensing that leads to creation of multiple
      * CounterCells.  The counter mechanics avoid contention on
      * updates but can encounter cache thrashing if read too
      * frequently during concurrent access. To avoid reading so often,
      * resizing under contention is attempted only upon adding to a
      * bin already holding two or more nodes. Under uniform hash
      * distributions, the probability of this occurring at threshold
      * is around 13%, meaning that only about 1 in 8 puts check
      * threshold (and after resizing, many fewer do so).
      *
      * TreeBins use a special form of comparison for search and
      * related operations (which is the main reason we cannot use
      * existing collections such as TreeMaps). TreeBins contain
      * Comparable elements, but may contain others, as well as
      * elements that are Comparable but not necessarily Comparable for
      * the same T, so we cannot invoke compareTo among them. To handle
      * this, the tree is ordered primarily by hash value, then by
      * Comparable.compareTo order if applicable.  On lookup at a node,
      * if elements are not comparable or compare as 0 then both left
      * and right children may need to be searched in the case of tied
      * hash values. (This corresponds to the full list search that
      * would be necessary if all elements were non-Comparable and had
      * tied hashes.) On insertion, to keep a total ordering (or as
      * close as is required here) across rebalancings, we compare
      * classes and identityHashCodes as tie-breakers. The red-black
      * balancing code is updated from pre-jdk-collections
      * (http://gee.cs.oswego.edu/dl/classes/collections/RBCell.java)
      * based in turn on Cormen, Leiserson, and Rivest "Introduction to
      * Algorithms" (CLR).
      *
      * TreeBins also require an additional locking mechanism.  While
      * list traversal is always possible by readers even during
      * updates, tree traversal is not, mainly because of tree-rotations
      * that may change the root node and/or its linkages.  TreeBins
      * include a simple read-write lock mechanism parasitic on the
      * main bin-synchronization strategy: Structural adjustments
      * associated with an insertion or removal are already bin-locked
      * (and so cannot conflict with other writers) but must wait for
      * ongoing readers to finish. Since there can be only one such
      * waiter, we use a simple scheme using a single "waiter" field to
      * block writers.  However, readers need never block.  If the root
      * lock is held, they proceed along the slow traversal path (via
      * next-pointers) until the lock becomes available or the list is
      * exhausted, whichever comes first. These cases are not fast, but
      * maximize aggregate expected throughput.
      *
      * Maintaining API and serialization compatibility with previous
      * versions of this class introduces several oddities. Mainly: We
      * leave untouched but unused constructor arguments refering to
      * concurrencyLevel. We accept a loadFactor constructor argument,
      * but apply it only to initial table capacity (which is the only
      * time that we can guarantee to honor it.) We also declare an
      * unused "Segment" class that is instantiated in minimal form
      * only when serializing.
      *
      * Also, solely for compatibility with previous versions of this
      * class, it extends AbstractMap, even though all of its methods
      * are overridden, so it is just useless baggage.
      *
      * This file is organized to make things a little easier to follow
      * while reading than they might otherwise: First the main static
      * declarations and utilities, then fields, then main public
      * methods (with a few factorings of multiple public methods into
      * internal ones), then sizing methods, trees, traversers, and
      * bulk operations.
      */
 
     /* ---------------- Constants -------------- */

   
The largest possible table capacity. This value must be exactly 1<<30 to stay within Java array allocation and indexing bounds for power of two table sizes, and is further required because the top two bits of 32bit hash fields are used for control purposes.
 
    private static final int MAXIMUM_CAPACITY = 1 << 30;

   
The default initial table capacity. Must be a power of 2 (i.e., at least 1) and at most MAXIMUM_CAPACITY.
 
    private static final int DEFAULT_CAPACITY = 16;

   
The largest possible (non-power of two) array size. Needed by toArray and related methods.
 
    static final int MAX_ARRAY_SIZE = . - 8;

   
The default concurrency level for this table. Unused but defined for compatibility with previous versions of this class.
 
    private static final int DEFAULT_CONCURRENCY_LEVEL = 16;

   
The load factor for this table. Overrides of this value in constructors affect only the initial table capacity. The actual floating point value isn't normally used -- it is simpler to use expressions such as n - (n >>> 2) for the associated resizing threshold.
 
    private static final float LOAD_FACTOR = 0.75f;

   
The bin count threshold for using a tree rather than list for a bin. Bins are converted to trees when adding an element to a bin with at least this many nodes. The value must be greater than 2, and should be at least 8 to mesh with assumptions in tree removal about conversion back to plain bins upon shrinkage.
 
    static final int TREEIFY_THRESHOLD = 8;

   
The bin count threshold for untreeifying a (split) bin during a resize operation. Should be less than TREEIFY_THRESHOLD, and at most 6 to mesh with shrinkage detection under removal.
 
    static final int UNTREEIFY_THRESHOLD = 6;

   
The smallest table capacity for which bins may be treeified. (Otherwise the table is resized if too many nodes in a bin.) The value should be at least 4 * TREEIFY_THRESHOLD to avoid conflicts between resizing and treeification thresholds.
 
    static final int MIN_TREEIFY_CAPACITY = 64;

   
Minimum number of rebinnings per transfer step. Ranges are subdivided to allow multiple resizer threads. This value serves as a lower bound to avoid resizers encountering excessive memory contention. The value should be at least DEFAULT_CAPACITY.
 
    private static final int MIN_TRANSFER_STRIDE = 16;

   
The number of bits used for generation stamp in sizeCtl. Must be at least 6 for 32bit arrays.
 
    private static int RESIZE_STAMP_BITS = 16;

   
The maximum number of threads that can help resize. Must fit in 32 - RESIZE_STAMP_BITS bits.
 
    private static final int MAX_RESIZERS = (1 << (32 - )) - 1;

   
The bit shift for recording size stamp in sizeCtl.
 
    private static final int RESIZE_STAMP_SHIFT = 32 - ;
 
    /*
      * Encodings for Node hash fields. See above for explanation.
      */
    static final int MOVED     = -1; // hash for forwarding nodes
    static final int TREEBIN   = -2; // hash for roots of trees
    static final int RESERVED  = -3; // hash for transient reservations
    static final int HASH_BITS = 0x7fffffff; // usable bits of normal node hash
 
   
Number of CPUS, to place bounds on some sizings
 
    static final int NCPU = Runtime.getRuntime().availableProcessors();

   
For serialization compatibility.
 
    private static final ObjectStreamField[] serialPersistentFields = {
          new ObjectStreamField("segments", Segment[].class),
          new ObjectStreamField("segmentMask".),
          new ObjectStreamField("segmentShift".)
    };
 
     /* ---------------- Nodes -------------- */
 
    static class NodeEquivalence<K,V> {
       final Equivalence<K> keyEq;
       final Equivalence<V> valueEq;
 
       NodeEquivalence(Equivalence<K> keyEqEquivalence<V> valueEq) {
          this. = keyEq;
          this. = valueEq;
       }
    }

   
Key-value entry. This class is never exported out as a user-mutable Map.Entry (i.e., one supporting setValue; see MapEntry below), but can be used for read-only traversals used in bulk tasks. Subclasses of Node with a negative hash field are special, and contain null keys and values (but are never exported). Otherwise, keys and vals are never null.
 
    static class Node<K,V> implements Map.Entry<K,V> {
       final int hash;
       final K key;
       final NodeEquivalence<K, V> nodeEq// EQUIVALENCE_MOD
       volatile V val;
       volatile Node<K,V> next;
 
       Node(int hashNodeEquivalence<K, V> nodeEq, K key, V valNode<K,V> next) { // EQUIVALENCE_MOD
          this. = hash;
          this. = key;
          this. = val;
          this. = next;
          this. = nodeEq// EQUIVALENCE_MOD
       }
 
       public final K getKey()       { return ; }
       public final V getValue()     { return ; }
       public final int hashCode()   { return ..hashCode() ^ ..hashCode(); } // EQUIVALENCE_MOD
       public final String toString(){ return  + "=" + ; }
       public final V setValue(V value) {
          throw new UnsupportedOperationException();
       }
 
       @SuppressWarnings("unchecked")
       public final boolean equals(Object o) {
          Object kvuMap.Entry<?,?> e;
          return ((o instanceof Map.Entry) &&
                        (k = (e = (Map.Entry<?,?>)o).getKey()) != null &&
                        (v = e.getValue()) != null &&
                        (k ==  || ..equals(k)) && // EQUIVALENCE_MOD
                        (v == (u = ) || ..equals((V) uv))); // EQUIVALENCE_MOD
       }

      
Virtualized support for map.get(); overridden in subclasses.
 
       Node<K,V> find(int hObject k) {
          Node<K,V> e = this;
          if (k != null) {
             do {
                K ek;
                if (e.hash == h &&
                      ((ek = e.key) == k || (ek != null && ..equals(ekk)))) // EQUIVALENCE_MOD
                   return e;
             } while ((e = e.next) != null);
          }
          return null;
       }
    }
 
     /* ---------------- Static utilities -------------- */

   
Spreads (XORs) higher bits of hash to lower and also forces top bit to 0. Because the table uses power-of-two masking, sets of hashes that vary only in bits above the current mask will always collide. (Among known examples are sets of Float keys holding consecutive whole numbers in small tables.) So we apply a transform that spreads the impact of higher bits downward. There is a tradeoff between speed, utility, and quality of bit-spreading. Because many common sets of hashes are already reasonably distributed (so don't benefit from spreading), and because we use trees to handle large sets of collisions in bins, we just XOR some shifted bits in the cheapest possible way to reduce systematic lossage, as well as to incorporate impact of the highest bits that would otherwise never be used in index calculations because of table bounds.
 
    static final int spread(int h) {
       return (h ^ (h >>> 16)) & ;
    }

   
Returns a power of two table size for the given desired capacity. See Hackers Delight, sec 3.2
 
    private static final int tableSizeFor(int c) {
       int n = c - 1;
       n |= n >>> 1;
       n |= n >>> 2;
       n |= n >>> 4;
       n |= n >>> 8;
       n |= n >>> 16;
       return (n < 0) ? 1 : (n >= ) ?  : n + 1;
    }

   
Returns x's Class if it is of the form "class C implements Comparable<C>", else null.
 
    static Class<?> comparableClassFor(Object xEquivalence<?> eq) { // EQUIVALENCE_MOD
       if (eq.isComparable(x)) { // EQUIVALENCE_MOD
          Class<?> cType[] tsasType tParameterizedType p;
          if ((c = x.getClass()) == String.class// bypass checks
             return c;
          if ((ts = c.getGenericInterfaces()) != null) {
             for (int i = 0; i < ts.length; ++i) {
                if (((t = ts[i]) instanceof ParameterizedType) &&
                      ((p = (ParameterizedType)t).getRawType() ==
                             Comparable.class) &&
                      (as = p.getActualTypeArguments()) != null &&
                      as.length == 1 && as[0] == c// type arg is c
                   return c;
             }
          }
       }
       return null;
    }

   
Returns k.compareTo(x) if x matches kc (k's screened comparable class), else 0.
 
    @SuppressWarnings({"rawtypes","unchecked"}) // for cast to Comparable
    static int compareComparables(Class<?> kcObject kObject xEquivalence<Objecteq) { // EQUIVALENCE_MOD
       return (x == null || x.getClass() != kc ? 0 : eq.compare(kx)); // EQUIVALENCE_MOD
    }
 
     /* ---------------- Table element access -------------- */
 
     /*
      * Volatile access methods are used for table elements as well as
      * elements of in-progress next table while resizing.  All uses of
      * the tab arguments must be null checked by callers.  All callers
      * also paranoically precheck that tab's length is not zero (or an
      * equivalent check), thus ensuring that any index argument taking
      * the form of a hash value anded with (length - 1) is a valid
      * index.  Note that, to be correct wrt arbitrary concurrency
      * errors by users, these checks must operate on local variables,
      * which accounts for some odd-looking inline assignments below.
      * Note that calls to setTabAt always occur within locked regions,
      * and so in principle require only release ordering, not
      * full volatile semantics, but are currently coded as volatile
      * writes to be conservative.
      */
 
    @SuppressWarnings("unchecked")
    static final <K,V> Node<K,V> tabAt(Node<K,V>[] tabint i) {
       return (Node<K,V>).getObjectVolatile(tab, ((long)i << ) + );
    }
 
    static final <K,V> boolean casTabAt(Node<K,V>[] tabint i,
          Node<K,V> cNode<K,V> v) {
       return .compareAndSwapObject(tab, ((long)i << ) + cv);
    }
 
    static final <K,V> void setTabAt(Node<K,V>[] tabint iNode<K,V> v) {
       .putObjectVolatile(tab, ((long)i << ) + v);
    }
 
     /* ---------------- Fields -------------- */

   
The array of bins. Lazily initialized upon first insertion. Size is always a power of two. Accessed directly by iterators.
 
    transient volatile Node<K,V>[] table;

   
The next table to use; non-null only while resizing.
 
    private transient volatile Node<K,V>[] nextTable;

   
Base counter value, used mainly when there is no contention, but also as a fallback during table initialization races. Updated via CAS.
 
    private transient volatile long baseCount;

   
Table initialization and resizing control. When negative, the table is being initialized or resized: -1 for initialization, else -(1 + the number of active resizing threads). Otherwise, when table is null, holds the initial table size to use upon creation, or 0 for default. After initialization, holds the next element count value upon which to resize the table.
 
    private transient volatile int sizeCtl;

   
The next table index (plus one) to split while resizing.
 
    private transient volatile int transferIndex;

   
Spinlock (locked via CAS) used when resizing and/or creating CounterCells.
 
    private transient volatile int cellsBusy;

   
Table of counter cells. When non-null, size is a power of 2.
 
    private transient volatile CounterCell[] counterCells;
 
    // views
    private transient KeySetView<K,V> keySet;
    private transient ValuesView<K,V> values;
    private transient EntrySetView<K,V> entrySet;
 
    Equivalence<K> keyEq;
    Equivalence<V> valueEq;
    transient NodeEquivalence<K, V> nodeEq;
 
     /* ---------------- Public operations -------------- */

   
Creates a new, empty map with the default initial table size (16).
 
          Equivalence<K> keyEquivalenceEquivalence<V> valueEquivalence) {
       this. = keyEquivalence// EQUIVALENCE_MOD
       this. = valueEquivalence// EQUIVALENCE_MOD
       this. = new NodeEquivalence<K, V>(this.this.); // EQUIVALENCE_MOD
    }

   
Creates a new, empty map with an initial table size accommodating the specified number of elements without the need to dynamically resize.

Parameters:
initialCapacity The implementation performs internal sizing to accommodate this many elements.
Throws:
java.lang.IllegalArgumentException if the initial capacity of elements is negative
 
    public EquivalentConcurrentHashMapV8(int initialCapacity,
          Equivalence<K> keyEquivalenceEquivalence<V> valueEquivalence) {
       this(keyEquivalencevalueEquivalence); // EQUIVALENCE_MOD
       if (initialCapacity < 0)
          throw new IllegalArgumentException();
       int cap = ((initialCapacity >= ( >>> 1)) ?
                         :
                        tableSizeFor(initialCapacity + (initialCapacity >>> 1) + 1));
       this. = cap;
    }

   
Creates a new map with the same mappings as the given map.

Parameters:
m the map
 
    public EquivalentConcurrentHashMapV8(Map<? extends K, ? extends V> m,
          Equivalence<K> keyEquivalenceEquivalence<V> valueEquivalence) {
       this(keyEquivalencevalueEquivalence); // EQUIVALENCE_MOD
       this. = ;
       putAll(m);
    }

   
Creates a new, empty map with an initial table size based on the given number of elements (initialCapacity) and initial table density (loadFactor).

Parameters:
initialCapacity the initial capacity. The implementation performs internal sizing to accommodate this many elements, given the specified load factor.
loadFactor the load factor (table density) for establishing the initial table size
Throws:
java.lang.IllegalArgumentException if the initial capacity of elements is negative or the load factor is nonpositive
Since:
1.6
 
    public EquivalentConcurrentHashMapV8(int initialCapacityfloat loadFactor,
          Equivalence<K> keyEquivalenceEquivalence<V> valueEquivalence) {
       this(initialCapacityloadFactor, 1, keyEquivalencevalueEquivalence); // EQUIVALENCE_MOD
    }

   
Creates a new, empty map with an initial table size based on the given number of elements (initialCapacity), table density (loadFactor), and number of concurrently updating threads (concurrencyLevel).

Parameters:
initialCapacity the initial capacity. The implementation performs internal sizing to accommodate this many elements, given the specified load factor.
loadFactor the load factor (table density) for establishing the initial table size
concurrencyLevel the estimated number of concurrently updating threads. The implementation may use this value as a sizing hint.
Throws:
java.lang.IllegalArgumentException if the initial capacity is negative or the load factor or concurrencyLevel are nonpositive
 
    public EquivalentConcurrentHashMapV8(int initialCapacity,
          float loadFactorint concurrencyLevel,
          Equivalence<K> keyEquivalenceEquivalence<V> valueEquivalence) {
       this(keyEquivalencevalueEquivalence); // EQUIVALENCE_MOD
       if (!(loadFactor > 0.0f) || initialCapacity < 0 || concurrencyLevel <= 0)
          throw new IllegalArgumentException();
       if (initialCapacity < concurrencyLevel)   // Use at least as many bins
          initialCapacity = concurrencyLevel;   // as estimated threads
       long size = (long)(1.0 + (long)initialCapacity / loadFactor);
       int cap = (size >= (long)) ?
              : tableSizeFor((int)size);
       this. = cap;
    }
 
    // Original (since JDK1.2) Map methods
 
   
 
    public int size() {
       long n = sumCount();
       return ((n < 0L) ? 0 :
                     (n > (long).) ? . :
                           (int)n);
    }

   
 
    public boolean isEmpty() {
       return sumCount() <= 0L; // ignore transient negative values
    }

   
Returns the value to which the specified key is mapped, or null if this map contains no mapping for the key.

More formally, if this map contains a mapping from a key k to a value v such that key.equals(k), then this method returns v; otherwise it returns null. (There can be at most one such mapping.)

Throws:
java.lang.NullPointerException if the specified key is null
 
    public V get(Object key) {
       Node<K,V>[] tabNode<K,V> epint neh; K ek;
       int h = spread(.hashCode(key)); // EQUIVALENCE_MOD
       if ((tab = ) != null && (n = tab.length) > 0 &&
             (e = tabAt(tab, (n - 1) & h)) != null) {
          if ((eh = e.hash) == h) {
             if ((ek = e.key) == key || (ek != null && .equals(ekkey))) // EQUIVALENCE_MOD
                return e.val;
          }
          else if (eh < 0)
             return (p = e.find(hkey)) != null ? p.val : null;
          while ((e = e.next) != null) {
             if (e.hash == h &&
                   ((ek = e.key) == key || (ek != null && .equals(ekkey)))) // EQUIVALENCE_MOD
                return e.val;
          }
       }
       return null;
    }

   
Tests if the specified object is a key in this table.

Parameters:
key possible key
Returns:
true if and only if the specified object is a key in this table, as determined by the equals method; false otherwise
Throws:
java.lang.NullPointerException if the specified key is null
   public boolean containsKey(Object key) {
      return get(key) != null;
   }

   
Returns true if this map maps one or more keys to the specified value. Note: This method may require a full traversal of the map, and is much slower than method containsKey.

Parameters:
value value whose presence in this map is to be tested
Returns:
true if this map maps one or more keys to the specified value
Throws:
java.lang.NullPointerException if the specified value is null
   public boolean containsValue(Object value) {
      if (value == null)
         throw new NullPointerException();
      Node<K,V>[] t;
      if ((t = ) != null) {
         Traverser<K,V> it = new Traverser<K,V>(tt.length, 0, t.length);
         for (Node<K,V> p; (p = it.advance()) != null; ) {
            V v;
            if ((v = p.val) == value || (v != null && .equals(vvalue))) // EQUIVALENCE_MOD
               return true;
         }
      }
      return false;
   }

   
Maps the specified key to the specified value in this table. Neither the key nor the value can be null.

The value can be retrieved by calling the get method with a key that is equal to the original key.

Parameters:
key key with which the specified value is to be associated
value value to be associated with the specified key
Returns:
the previous value associated with key, or null if there was no mapping for key
Throws:
java.lang.NullPointerException if the specified key or value is null
   public V put(K key, V value) {
      return putVal(keyvaluefalse);
   }

   
Implementation for put and putIfAbsent
   final V putVal(K key, V valueboolean onlyIfAbsent) {
      if (key == null || value == nullthrow new NullPointerException();
      int hash = spread(.hashCode(key)); // EQUIVALENCE_MOD
      int binCount = 0;
      for (Node<K,V>[] tab = ;;) {
         Node<K,V> fint nifh;
         if (tab == null || (n = tab.length) == 0)
            tab = initTable();
         else if ((f = tabAt(tabi = (n - 1) & hash)) == null) {
            if (casTabAt(tabinull,
                  new Node<K,V>(hashkeyvaluenull))) // EQUIVALENCE_MOD
               break;                   // no lock when adding to empty bin
         }
         else if ((fh = f.hash) == )
            tab = helpTransfer(tabf);
         else {
            V oldVal = null;
            synchronized (f) {
               if (tabAt(tabi) == f) {
                  if (fh >= 0) {
                     binCount = 1;
                     for (Node<K,V> e = f;; ++binCount) {
                        K ek;
                        if (e.hash == hash &&
                              ((ek = e.key) == key ||
                                     (ek != null && .equals(ekkey)))) { // EQUIVALENCE_MOD
                           oldVal = e.val;
                           if (!onlyIfAbsent)
                              e.val = value;
                           break;
                        }
                        Node<K,V> pred = e;
                        if ((e = e.next) == null) {
                           pred.next = new Node<K,V>(hashkey// EQUIVALENCE_MOD
                                 valuenull);
                           break;
                        }
                     }
                  }
                  else if (f instanceof TreeBin) {
                     Node<K,V> p;
                     binCount = 2;
                     if ((p = ((TreeBin<K,V>)f).putTreeVal(hashkey,
                           value)) != null) {
                        oldVal = p.val;
                        if (!onlyIfAbsent)
                           p.val = value;
                     }
                  }
               }
            }
            if (binCount != 0) {
               if (binCount >= )
                  treeifyBin(tabi);
               if (oldVal != null)
                  return oldVal;
               break;
            }
         }
      }
      addCount(1L, binCount);
      return null;
   }

   
Copies all of the mappings from the specified map to this one. These mappings replace any mappings that this map had for any of the keys currently in the specified map.

Parameters:
m mappings to be stored in this map
   public void putAll(Map<? extends K, ? extends V> m) {
      tryPresize(m.size());
      for (Map.Entry<? extends K, ? extends V> e : m.entrySet())
         putVal(e.getKey(), e.getValue(), false);
   }

   
Removes the key (and its corresponding value) from this map. This method does nothing if the key is not in the map.

Parameters:
key the key that needs to be removed
Returns:
the previous value associated with key, or null if there was no mapping for key
Throws:
java.lang.NullPointerException if the specified key is null
   public V remove(Object key) {
      return replaceNode(keynullnull);
   }

   
Implementation for the four public remove/replace methods: Replaces node value with v, conditional upon match of cv if non-null. If resulting value is null, delete.
   final V replaceNode(Object key, V valueObject cv) {
      int hash = spread(.hashCode(key)); // EQUIVALENCE_MOD
      for (Node<K,V>[] tab = ;;) {
         Node<K,V> fint nifh;
         if (tab == null || (n = tab.length) == 0 ||
               (f = tabAt(tabi = (n - 1) & hash)) == null)
            break;
         else if ((fh = f.hash) == )
            tab = helpTransfer(tabf);
         else {
            V oldVal = null;
            boolean validated = false;
            synchronized (f) {
               if (tabAt(tabi) == f) {
                  if (fh >= 0) {
                     validated = true;
                     for (Node<K,V> e = fpred = null;;) {
                        K ek;
                        if (e.hash == hash &&
                              ((ek = e.key) == key ||
                                     (ek != null && .equals(ekkey)))) { // EQUIVALENCE_MOD
                           V ev = e.val;
                           if (cv == null || cv == ev ||
                                 (ev != null && .equals(evcv))) { // EQUIVALENCE_MOD
                              oldVal = ev;
                              if (value != null)
                                 e.val = value;
                              else if (pred != null)
                                 pred.next = e.next;
                              else
                                 setTabAt(tabie.next);
                           }
                           break;
                        }
                        pred = e;
                        if ((e = e.next) == null)
                           break;
                     }
                  }
                  else if (f instanceof TreeBin) {
                     validated = true;
                     TreeBin<K,V> t = (TreeBin<K,V>)f;
                     TreeNode<K,V> rp;
                     if ((r = t.root) != null &&
                           (p = r.findTreeNode(hashkeynull)) != null) {
                        V pv = p.val;
                        if (cv == null || cv == pv ||
                              (pv != null && .equals(pvcv))) { // EQUIVALENCE_MOD
                           oldVal = pv;
                           if (value != null)
                              p.val = value;
                           else if (t.removeTreeNode(p))
                              setTabAt(tabiuntreeify(t.first)); // EQUIVALENCE_MOD
                        }
                     }
                  }
               }
            }
            if (validated) {
               if (oldVal != null) {
                  if (value == null)
                     addCount(-1L, -1);
                  return oldVal;
               }
               break;
            }
         }
      }
      return null;
   }

   
Removes all of the mappings from this map.
   public void clear() {
      long delta = 0L; // negative number of deletions
      int i = 0;
      Node<K,V>[] tab = ;
      while (tab != null && i < tab.length) {
         int fh;
         Node<K,V> f = tabAt(tabi);
         if (f == null)
            ++i;
         else if ((fh = f.hash) == ) {
            tab = helpTransfer(tabf);
            i = 0; // restart
         }
         else {
            synchronized (f) {
               if (tabAt(tabi) == f) {
                  Node<K,V> p = (fh >= 0 ? f :
                                       (f instanceof TreeBin) ?
                                             ((TreeBin<K,V>)f). : null);
                  while (p != null) {
                     --delta;
                     p = p.next;
                  }
                  setTabAt(tabi++, null);
               }
            }
         }
      }
      if (delta != 0L)
         addCount(delta, -1);
   }

   
Returns a java.util.Set view of the keys contained in this map. The set is backed by the map, so changes to the map are reflected in the set, and vice-versa. The set supports element removal, which removes the corresponding mapping from this map, via the Iterator.remove, Set.remove, removeAll, retainAll, and clear operations. It does not support the add or addAll operations.

The view's iterator is a "weakly consistent" iterator that will never throw java.util.ConcurrentModificationException, and guarantees to traverse elements as they existed upon construction of the iterator, and may (but is not guaranteed to) reflect any modifications subsequent to construction.

Returns:
the set view
   public KeySetView<K,V> keySet() {
      KeySetView<K,V> ks;
      return (ks = ) != null ? ks : ( = new KeySetView<K,V>(thisnull));
   }

   
Returns a java.util.Collection view of the values contained in this map. The collection is backed by the map, so changes to the map are reflected in the collection, and vice-versa. The collection supports element removal, which removes the corresponding mapping from this map, via the Iterator.remove, Collection.remove, removeAll, retainAll, and clear operations. It does not support the add or addAll operations.

The view's iterator is a "weakly consistent" iterator that will never throw java.util.ConcurrentModificationException, and guarantees to traverse elements as they existed upon construction of the iterator, and may (but is not guaranteed to) reflect any modifications subsequent to construction.

Returns:
the collection view
   public Collection<V> values() {
      ValuesView<K,V> vs;
      return (vs = ) != null ? vs : ( = new ValuesView<K,V>(this));
   }

   
Returns a java.util.Set view of the mappings contained in this map. The set is backed by the map, so changes to the map are reflected in the set, and vice-versa. The set supports element removal, which removes the corresponding mapping from the map, via the Iterator.remove, Set.remove, removeAll, retainAll, and clear operations.

The view's iterator is a "weakly consistent" iterator that will never throw java.util.ConcurrentModificationException, and guarantees to traverse elements as they existed upon construction of the iterator, and may (but is not guaranteed to) reflect any modifications subsequent to construction.

Returns:
the set view
   public Set<Map.Entry<K,V>> entrySet() {
      EntrySetView<K,V> es;
      return (es = ) != null ? es : ( = new EntrySetView<K,V>(this));
   }

   
Returns the hash code value for this java.util.Map, i.e., the sum of, for each key-value pair in the map, key.hashCode() ^ value.hashCode().

Returns:
the hash code value for this map
   public int hashCode() {
      int h = 0;
      Node<K,V>[] t;
      if ((t = ) != null) {
         Traverser<K,V> it = new Traverser<K,V>(tt.length, 0, t.length);
         for (Node<K,V> p; (p = it.advance()) != null; )
            h += .hashCode(p.key) ^ .hashCode(p.val); // EQUIVALENCE_MOD
      }
      return h;
   }

   
Returns a string representation of this map. The string representation consists of a list of key-value mappings (in no particular order) enclosed in braces ("{}"). Adjacent mappings are separated by the characters ", " (comma and space). Each key-value mapping is rendered as the key followed by an equals sign ("=") followed by the associated value.

Returns:
a string representation of this map
   public String toString() {
      Node<K,V>[] t;
      int f = (t = ) == null ? 0 : t.length;
      Traverser<K,V> it = new Traverser<K,V>(tf, 0, f);
      StringBuilder sb = new StringBuilder();
      sb.append('{');
      Node<K,V> p;
      if ((p = it.advance()) != null) {
         for (;;) {
            K k = p.key;
            V v = p.val;
            sb.append(k == this ? "(this Map)" : .toString(k)); // EQUIVALENCE_MOD
            sb.append('=');
            sb.append(v == this ? "(this Map)" : .toString(v)); // EQUIVALENCE_MOD
            if ((p = it.advance()) == null)
               break;
            sb.append(',').append(' ');
         }
      }
      return sb.append('}').toString();
   }

   
Compares the specified object with this map for equality. Returns true if the given object is a map with the same mappings as this map. This operation may return misleading results if either map is concurrently modified during execution of this method.

Parameters:
o object to be compared for equality with this map
Returns:
true if the specified object is equal to this map
   @SuppressWarnings("unchecked")
   public boolean equals(Object o) {
      if (o != this) {
         if (!(o instanceof Map))
            return false;
         Map<?,?> m = (Map<?,?>) o;
         Node<K,V>[] t;
         int f = (t = ) == null ? 0 : t.length;
         Traverser<K,V> it = new Traverser<K,V>(tf, 0, f);
         for (Node<K,V> p; (p = it.advance()) != null; ) {
            V val = p.val;
            Object v = m.get(p.key);
            if (v == null || (v != val && !.equals(valv))) // EQUIVALENCE_MOD
               return false;
         }
         for (Map.Entry<?,?> e : m.entrySet()) {
            Object mkmvv;
            if ((mk = e.getKey()) == null ||
                  (mv = e.getValue()) == null ||
                  (v = get(mk)) == null ||
                  (mv != v && !.equals((V) vmv))) // EQUIVALENCE_MOD
               return false;
         }
      }
      return true;
   }

   
Stripped-down version of helper class used in previous version, declared for the sake of serialization compatibility
   static class Segment<K,V> extends ReentrantLock implements Serializable {
      private static final long serialVersionUID = 2249069246763182397L;
      final float loadFactor;
      Segment(float lf) { this. = lf; }
   }

   
Saves the state of the EquivalentConcurrentHashMapV8 instance to a stream (i.e., serializes it).

Parameters:
s the stream
Throws:
java.io.IOException if an I/O error occurs
SerialData:
the key (Object) and value (Object) for each key-value mapping, followed by a null pair. The key-value mappings are emitted in no particular order.
   @SuppressWarnings("unchecked")
   private void writeObject(java.io.ObjectOutputStream s)
         throws java.io.IOException {
      s.writeObject(); // EQUIVALENCE_MOD
      s.writeObject(); // EQUIVALENCE_MOD
      // For serialization compatibility
      // Emulate segment calculation from previous version of this class
      int sshift = 0;
      int ssize = 1;
      while (ssize < ) {
         ++sshift;
         ssize <<= 1;
      }
      int segmentShift = 32 - sshift;
      int segmentMask = ssize - 1;
      Segment<K,V>[] segments = (Segment<K,V>[])
            new Segment<?,?>[];
      for (int i = 0; i < segments.length; ++i)
         segments[i] = new Segment<K,V>();
      s.putFields().put("segments"segments);
      s.putFields().put("segmentShift"segmentShift);
      s.putFields().put("segmentMask"segmentMask);
      s.writeFields();
      Node<K,V>[] t;
      if ((t = ) != null) {
         Traverser<K,V> it = new Traverser<K,V>(tt.length, 0, t.length);
         for (Node<K,V> p; (p = it.advance()) != null; ) {
            s.writeObject(p.key);
            s.writeObject(p.val);
         }
      }
      s.writeObject(null);
      s.writeObject(null);
      segments = null// throw away
   }

   
Reconstitutes the instance from a stream (that is, deserializes it).

Parameters:
s the stream
Throws:
java.lang.ClassNotFoundException if the class of a serialized object could not be found
java.io.IOException if an I/O error occurs
   @SuppressWarnings("unchecked")
   private void readObject(java.io.ObjectInputStream s)
         throws java.io.IOExceptionClassNotFoundException {
       = (Equivalence<K>) s.readObject();
       = (Equivalence<V>) s.readObject();
       = new NodeEquivalence<K, V>();
      /*
       * To improve performance in typical cases, we create nodes
       * while reading, then place in table once size is known.
       * However, we must also validate uniqueness and deal with
       * overpopulated bins while doing so, which requires
       * specialized versions of putVal mechanics.
       */
       = -1; // force exclusion for table construction
      s.defaultReadObject();
      long size = 0L;
      Node<K,V> p = null;
      for (;;) {
         K k = (K) s.readObject();
         V v = (V) s.readObject();
         if (k != null && v != null) {
            p = new Node<K,V>(spread(.hashCode(k)), kvp); // EQUIVALENCE_MOD
            ++size;
         }
         else
            break;
      }
      if (size == 0L)
          = 0;
      else {
         int n;
         if (size >= (long)( >>> 1))
            n = ;
         else {
            int sz = (int)size;
            n = tableSizeFor(sz + (sz >>> 1) + 1);
         }
         @SuppressWarnings("unchecked")
         Node<K,V>[] tab = (Node<K,V>[])new Node<?,?>[n];
         int mask = n - 1;
         long added = 0L;
         while (p != null) {
            boolean insertAtFront;
            Node<K,V> next = p.nextfirst;
            int h = p.hashj = h & mask;
            if ((first = tabAt(tabj)) == null)
               insertAtFront = true;
            else {
               K k = p.key;
               if (first.hash < 0) {
                  TreeBin<K,V> t = (TreeBin<K,V>)first;
                  if (t.putTreeVal(hkp.val) == null)
                     ++added;
        &nbs