欢迎您访问程序员文章站本站旨在为大家提供分享程序员计算机编程知识!
您现在的位置是: 首页  >  IT编程

JDK8中的ConcurrentHashMap源码

程序员文章站 2022-06-05 17:42:06
背景 上文JDK8中的HashMap源码写了HashMap,这次写ConcurrentHashMap ConcurrentHashMap源码 /** * Maps the specified key to the specified value in this table. * Neither th ......

背景

上文jdk8中的hashmap源码写了hashmap,这次写concurrenthashmap

concurrenthashmap源码

/**
     * maps the specified key to the specified value in this table.
     * neither the key nor the value can be null.
     *
     * <p>the value can be retrieved by calling the {@code get} method
     * with a key that is equal to the original key.
     *
     * @param key key with which the specified value is to be associated
     * @param value value to be associated with the specified key
     * @return the previous value associated with {@code key}, or
     *         {@code null} if there was no mapping for {@code key}
     * @throws nullpointerexception if the specified key or value is null
     */
    public v put(k key, v value) {
        return putval(key, value, false);
    }

    /** implementation for put and putifabsent */
    final v putval(k key, v value, boolean onlyifabsent) {
        if (key == null || value == null) throw new nullpointerexception();
        int hash = spread(key.hashcode());
        int bincount = 0;
        for (node<k,v>[] tab = table;;) {
            node<k,v> f; int n, i, fh;
            //tab为空,则初始化
            if (tab == null || (n = tab.length) == 0)
                tab = inittable();
            else if ((f = tabat(tab, i = (n - 1) & hash)) == null) { //该槽为空,则尝试插入
                if (castabat(tab, i, null,
                             new node<k,v>(hash, key, value, null)))
                    break;                   // no lock when adding to empty bin
            }
            else if ((fh = f.hash) == moved) //正在移动,
                tab = helptransfer(tab, f);
            else {
                v oldval = null;
                synchronized (f) { //对该槽进行加锁
                    if (tabat(tab, i) == 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 && key.equals(ek)))) {
                                    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>(hash, key,
                                                              value, null);
                                    break;
                                }
                            }
                        }
                        else if (f instanceof treebin) {
                            node<k,v> p;
                            bincount = 2;
                            if ((p = ((treebin<k,v>)f).puttreeval(hash, key,
                                                           value)) != null) {
                                oldval = p.val;
                                if (!onlyifabsent)
                                    p.val = value;
                            }
                        }
                    }
                }
                if (bincount != 0) {
                    if (bincount >= treeify_threshold)
                        treeifybin(tab, i);
                    if (oldval != null)
                        return oldval;
                    break;
                }
            }
        }
        addcount(1l, bincount);
        return null;
    }
/**
     * returns the value to which the specified key is mapped,
     * or {@code null} if this map contains no mapping for the key.
     *
     * <p>more formally, if this map contains a mapping from a key
     * {@code k} to a value {@code v} such that {@code key.equals(k)},
     * then this method returns {@code v}; otherwise it returns
     * {@code null}.  (there can be at most one such mapping.)
     *
     * @throws nullpointerexception if the specified key is null
     */
    public v get(object key) {
        node<k,v>[] tab; node<k,v> e, p; int n, eh; k ek;
        //获得hash值
        int h = spread(key.hashcode());
        //表非空,且该处不为空
        if ((tab = table) != null && (n = tab.length) > 0 &&
            (e = tabat(tab, (n - 1) & h)) != null) {
            if ((eh = e.hash) == h) { //判断第1个
                if ((ek = e.key) == key || (ek != null && key.equals(ek)))
                    return e.val;
            }
            else if (eh < 0) //eh<0,找其他的
                return (p = e.find(h, key)) != null ? p.val : null;
            while ((e = e.next) != null) { //遍历
                if (e.hash == h &&
                    ((ek = e.key) == key || (ek != null && key.equals(ek))))
                    return e.val;
            }
        }
        return null;
    }

concurrenthashmap代码太多了,粘了好几次粘不上来。只粘几个方法吧。

阅后感

concurrenthashmap通过几个原子操作尽量减少加锁操作。

扩容部分没有看太明白,尤其时扩容时进行get操作。后续再继续学习。

 

 

 

 

 

/* * oracle proprietary/confidential. use is subject to license terms. * * * * * * * * * * * * * * * * * * * * */
/* * * * * * * 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 java.util.concurrent;
import java.io.objectstreamfield;import java.io.serializable;import java.lang.reflect.parameterizedtype;import java.lang.reflect.type;import java.util.abstractmap;import java.util.arrays;import java.util.collection;import java.util.comparator;import java.util.enumeration;import java.util.hashmap;import java.util.hashtable;import java.util.iterator;import java.util.map;import java.util.nosuchelementexception;import java.util.set;import java.util.spliterator;import java.util.concurrent.concurrentmap;import java.util.concurrent.forkjoinpool;import java.util.concurrent.atomic.atomicreference;import java.util.concurrent.locks.locksupport;import java.util.concurrent.locks.reentrantlock;import java.util.function.biconsumer;import java.util.function.bifunction;import java.util.function.binaryoperator;import java.util.function.consumer;import java.util.function.doublebinaryoperator;import java.util.function.function;import java.util.function.intbinaryoperator;import java.util.function.longbinaryoperator;import java.util.function.todoublebifunction;import java.util.function.todoublefunction;import java.util.function.tointbifunction;import java.util.function.tointfunction;import java.util.function.tolongbifunction;import java.util.function.tolongfunction;import java.util.stream.stream;
/** * a hash table supporting full concurrency of retrievals and * high expected concurrency for updates. this class obeys the * same functional specification as {@link java.util.hashtable}, and * includes versions of methods corresponding to each method of * {@code hashtable}. however, even though all operations are * thread-safe, retrieval operations do <em>not</em> entail locking, * and there is <em>not</em> any support for locking the entire table * in a way that prevents all access.  this class is fully * interoperable with {@code hashtable} in programs that rely on its * thread safety but not on its synchronization details. * * <p>retrieval operations (including {@code get}) generally do not * block, so may overlap with update operations (including {@code put} * and {@code remove}). retrievals reflect the results of the most * recently <em>completed</em> update operations holding upon their * onset. (more formally, an update operation for a given key bears a * <em>happens-before</em> relation with any (non-null) retrieval for * that key reporting the updated value.)  for aggregate operations * such as {@code putall} and {@code clear}, concurrent retrievals may * reflect insertion or removal of only some entries.  similarly, * iterators, spliterators 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 <em>not</em> throw {@link * java.util.concurrentmodificationexception 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 * {@code size}, {@code isempty}, and {@code 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. * * <p>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 {@code * initialcapacity} constructor argument. an additional optional * {@code 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 {@code concurrencylevel} as an additional hint for * internal sizing.  note that using many keys with exactly the same * {@code hashcode()} is a sure way to slow down performance of any * hash table. to ameliorate impact, when keys are {@link comparable}, * this class may use comparison order among keys to help break ties. * * <p>a {@link set} projection of a concurrenthashmap may be created * (using {@link #newkeyset()} or {@link #newkeyset(int)}), or viewed * (using {@link #keyset(object)} when only keys are of interest, and the * mapped values are (perhaps transiently) not used or all take the * same mapping value. * * <p>a concurrenthashmap can be used as scalable frequency map (a * form of histogram or multiset) by using {@link * java.util.concurrent.atomic.longadder} values and initializing via * {@link #computeifabsent computeifabsent}. for example, to add a count * to a {@code concurrenthashmap<string,longadder> freqs}, you can use * {@code freqs.computeifabsent(k -> new longadder()).increment();} * * <p>this class and its views and iterators implement all of the * <em>optional</em> methods of the {@link map} and {@link iterator} * interfaces. * * <p>like {@link hashtable} but unlike {@link hashmap}, this class * does <em>not</em> allow {@code null} to be used as a key or value. * * <p>concurrenthashmaps support a set of sequential and parallel bulk * operations that, unlike most {@link stream} methods, 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 concurrenthashmap * 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 {@link java.util.map.entry} * objects do not support method {@code setvalue}. * * <ul> * <li> foreach: perform a given action on each element. * a variant form applies a given transformation on each element * before performing the action.</li> * * <li> search: return the first available non-null result of * applying a given function on each element; skipping further * search when a result is found.</li> * * <li> 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: * * <ul> * * <li> plain reductions. (there is not a form of this method for * (key, value) function arguments since there is no corresponding * return type.)</li> * * <li> mapped reductions that accumulate the results of a given * function applied to each element.</li> * * <li> reductions to scalar doubles, longs, and ints, using a * given basis value.</li> * * </ul> * </li> * </ul> * * <p>these bulk operations accept a {@code parallelismthreshold} * argument. methods proceed sequentially if the current map size is * estimated to be less than the given threshold. using a value of * {@code long.max_value} suppresses all parallelism.  using a value * of {@code 1} results in maximal parallelism by partitioning into * enough subtasks to fully utilize the {@link * 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. * * <p>the concurrency properties of bulk operations follow * from those of concurrenthashmap: any non-null result returned * from {@code 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. * * <p>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. * * <p>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 {@code new * abstractmap.simpleentry(k,v)}. * * <p>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. * * <p>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. * * <p>all arguments to all task methods must be non-null. * * <p>this class is a member of the * <a href="{@docroot}/../technotes/guides/collections/index.html"> * java collections framework</a>. * * @since 1.5 * @author doug lea * @param <k> the type of keys maintained by this map * @param <v> the type of mapped values */public class concurrenthashmap<k,v> extends abstractmap<k,v>    implements concurrentmap<k,v>, serializable {    private static final long serialversionuid = 7249069246763182397l;
    /*     * 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 = integer.max_value - 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 {@code 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 - resize_stamp_bits)) - 1;
    /**     * the bit shift for recording size stamp in sizectl.     */    private static final int resize_stamp_shift = 32 - resize_stamp_bits;
    /*     * 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", integer.type),        new objectstreamfield("segmentshift", integer.type)    };
    /* ---------------- nodes -------------- */
    /**     * 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;        volatile v val;        volatile node<k,v> next;
        node(int hash, k key, v val, node<k,v> next) {            this.hash = hash;            this.key = key;            this.val = val;            this.next = next;        }
        public final k getkey()       { return key; }        public final v getvalue()     { return val; }        public final int hashcode()   { return key.hashcode() ^ val.hashcode(); }        public final string tostring(){ return key + "=" + val; }        public final v setvalue(v value) {            throw new unsupportedoperationexception();        }
        public final boolean equals(object o) {            object k, v, u; map.entry<?,?> e;            return ((o instanceof map.entry) &&                    (k = (e = (map.entry<?,?>)o).getkey()) != null &&                    (v = e.getvalue()) != null &&                    (k == key || k.equals(key)) &&                    (v == (u = val) || v.equals(u)));        }
        /**         * virtualized support for map.get(); overridden in subclasses.         */        node<k,v> find(int h, object k) {            node<k,v> e = this;            if (k != null) {                do {                    k ek;                    if (e.hash == h &&                        ((ek = e.key) == k || (ek != null && k.equals(ek))))                        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)) & hash_bits;    }
    /**     * 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 >= maximum_capacity) ? maximum_capacity : n + 1;    }
    /**     * returns x's class if it is of the form "class c implements     * comparable<c>", else null.     */    static class<?> comparableclassfor(object x) {        if (x instanceof comparable) {            class<?> c; type[] ts, as; type t; parameterizedtype 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<?> kc, object k, object x) {        return (x == null || x.getclass() != kc ? 0 :                ((comparable)k).compareto(x));    }
    /* ---------------- 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>[] tab, int i) {        return (node<k,v>)u.getobjectvolatile(tab, ((long)i << ashift) + abase);    }
    static final <k,v> boolean castabat(node<k,v>[] tab, int i,                                        node<k,v> c, node<k,v> v) {        return u.compareandswapobject(tab, ((long)i << ashift) + abase, c, v);    }
    static final <k,v> void settabat(node<k,v>[] tab, int i, node<k,v> v) {        u.putobjectvolatile(tab, ((long)i << ashift) + abase, 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;

    /* ---------------- public operations -------------- */
    /**     * creates a new, empty map with the default initial table size (16).     */    public concurrenthashmap() {    }
    /**     * creates a new, empty map with an initial table size     * accommodating the specified number of elements without the need     * to dynamically resize.     *     * @param initialcapacity the implementation performs internal     * sizing to accommodate this many elements.     * @throws illegalargumentexception if the initial capacity of     * elements is negative     */    public concurrenthashmap(int initialcapacity) {        if (initialcapacity < 0)            throw new illegalargumentexception();        int cap = ((initialcapacity >= (maximum_capacity >>> 1)) ?                   maximum_capacity :                   tablesizefor(initialcapacity + (initialcapacity >>> 1) + 1));        this.sizectl = cap;    }
    /**     * creates a new map with the same mappings as the given map.     *     * @param m the map     */    public concurrenthashmap(map<? extends k, ? extends v> m) {        this.sizectl = default_capacity;        putall(m);    }
    /**     * creates a new, empty map with an initial table size based on     * the given number of elements ({@code initialcapacity}) and     * initial table density ({@code loadfactor}).     *     * @param initialcapacity the initial capacity. the implementation     * performs internal sizing to accommodate this many elements,     * given the specified load factor.     * @param loadfactor the load factor (table density) for     * establishing the initial table size     * @throws illegalargumentexception if the initial capacity of     * elements is negative or the load factor is nonpositive     *     * @since 1.6     */    public concurrenthashmap(int initialcapacity, float loadfactor) {        this(initialcapacity, loadfactor, 1);    }
    /**     * creates a new, empty map with an initial table size based on     * the given number of elements ({@code initialcapacity}), table     * density ({@code loadfactor}), and number of concurrently     * updating threads ({@code concurrencylevel}).     *     * @param initialcapacity the initial capacity. the implementation     * performs internal sizing to accommodate this many elements,     * given the specified load factor.     * @param loadfactor the load factor (table density) for     * establishing the initial table size     * @param concurrencylevel the estimated number of concurrently     * updating threads. the implementation may use this value as     * a sizing hint.     * @throws illegalargumentexception if the initial capacity is     * negative or the load factor or concurrencylevel are     * nonpositive     */    public concurrenthashmap(int initialcapacity,                             float loadfactor, int concurrencylevel) {        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)maximum_capacity) ?            maximum_capacity : tablesizefor((int)size);        this.sizectl = cap;    }
    // original (since jdk1.2) map methods
    /**     * {@inheritdoc}     */    public int size() {        long n = sumcount();        return ((n < 0l) ? 0 :                (n > (long)integer.max_value) ? integer.max_value :                (int)n);    }
    /**     * {@inheritdoc}     */    public boolean isempty() {        return sumcount() <= 0l; // ignore transient negative values    }
    /**     * returns the value to which the specified key is mapped,     * or {@code null} if this map contains no mapping for the key.     *     * <p>more formally, if this map contains a mapping from a key     * {@code k} to a value {@code v} such that {@code key.equals(k)},     * then this method returns {@code v}; otherwise it returns     * {@code null}.  (there can be at most one such mapping.)     *     * @throws nullpointerexception if the specified key is null     */    public v get(object key) {        node<k,v>[] tab; node<k,v> e, p; int n, eh; k ek;        //获得hash值        int h = spread(key.hashcode());        //表非空,且该处不为空        if ((tab = table) != null && (n = tab.length) > 0 &&            (e = tabat(tab, (n - 1) & h)) != null) {            if ((eh = e.hash) == h) { //判断第1个                if ((ek = e.key) == key || (ek != null && key.equals(ek)))                    return e.val;            }            else if (eh < 0) //eh<0,找其他的                return (p = e.find(h, key)) != null ? p.val : null;            while ((e = e.next) != null) { //遍历                if (e.hash == h &&                    ((ek = e.key) == key || (ek != null && key.equals(ek))))                    return e.val;            }        }        return null;    }
    /**     * tests if the specified object is a key in this table.     *     * @param  key possible key     * @return {@code true} if and only if the specified object     *         is a key in this table, as determined by the     *         {@code equals} method; {@code false} otherwise     * @throws nullpointerexception if the specified key is null     */    public boolean containskey(object key) {        return get(key) != null;    }
    /**     * returns {@code 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 {@code containskey}.     *     * @param value value whose presence in this map is to be tested     * @return {@code true} if this map maps one or more keys to the     *         specified value     * @throws nullpointerexception if the specified value is null     */    public boolean containsvalue(object value) {        if (value == null)            throw new nullpointerexception();        node<k,v>[] t;        if ((t = table) != null) {            traverser<k,v> it = new traverser<k,v>(t, t.length, 0, t.length);            for (node<k,v> p; (p = it.advance()) != null; ) {                v v;                if ((v = p.val) == value || (v != null && value.equals(v)))                    return true;            }        }        return false;    }
    /**     * maps the specified key to the specified value in this table.     * neither the key nor the value can be null.     *     * <p>the value can be retrieved by calling the {@code get} method     * with a key that is equal to the original key.     *     * @param key key with which the specified value is to be associated     * @param value value to be associated with the specified key     * @return the previous value associated with {@code key}, or     *         {@code null} if there was no mapping for {@code key}     * @throws nullpointerexception if the specified key or value is null     */    public v put(k key, v value) {        return putval(key, value, false);    }
    /** implementation for put and putifabsent */    final v putval(k key, v value, boolean onlyifabsent) {        if (key == null || value == null) throw new nullpointerexception();        int hash = spread(key.hashcode());        int bincount = 0;        for (node<k,v>[] tab = table;;) {            node<k,v> f; int n, i, fh;            //tab为空,则初始化            if (tab == null || (n = tab.length) == 0)                tab = inittable();            else if ((f = tabat(tab, i = (n - 1) & hash)) == null) { //该槽为空,则尝试插入                if (castabat(tab, i, null,                             new node<k,v>(hash, key, value, null)))                    break;                   // no lock when adding to empty bin            }            else if ((fh = f.hash) == moved) //正在移动,                tab = helptransfer(tab, f);            else {                v oldval = null;                synchronized (f) { //对该槽进行加锁                    if (tabat(tab, i) == 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 && key.equals(ek)))) {                                    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>(hash, key,                                                              value, null);                                    break;                                }                            }                        }                        else if (f instanceof treebin) {                            node<k,v> p;                            bincount = 2;                            if ((p = ((treebin<k,v>)f).puttreeval(hash, key,                                                           value)) != null) {                                oldval = p.val;                                if (!onlyifabsent)                                    p.val = value;                            }                        }                    }                }                if (bincount != 0) {                    if (bincount >= treeify_threshold)                        treeifybin(tab, i);                    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.     *     * @param 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.     *     * @param  key the key that needs to be removed     * @return the previous value associated with {@code key}, or     *         {@code null} if there was no mapping for {@code key}     * @throws nullpointerexception if the specified key is null     */    public v remove(object key) {        return replacenode(key, null, null);    }
    /**     * 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 value, object cv) {        int hash = spread(key.hashcode());        for (node<k,v>[] tab = table;;) {            node<k,v> f; int n, i, fh;            if (tab == null || (n = tab.length) == 0 ||                (f = tabat(tab, i = (n - 1) & hash)) == null)                break;            else if ((fh = f.hash) == moved)                tab = helptransfer(tab, f);            else {                v oldval = null;                boolean validated = false;                synchronized (f) {                    if (tabat(tab, i) == f) {                        if (fh >= 0) {                            validated = true;                            for (node<k,v> e = f, pred = null;;) {                                k ek;                                if (e.hash == hash &&                                    ((ek = e.key) == key ||                                     (ek != null && key.equals(ek)))) {                                    v ev = e.val;                                    if (cv == null || cv == ev ||                                        (ev != null && cv.equals(ev))) {                                        oldval = ev;                                        if (value != null)                                            e.val = value;                                        else if (pred != null)                                            pred.next = e.next;                                        else                                            settabat(tab, i, e.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> r, p;                            if ((r = t.root) != null &&                                (p = r.findtreenode(hash, key, null)) != null) {                                v pv = p.val;                                if (cv == null || cv == pv ||                                    (pv != null && cv.equals(pv))) {                                    oldval = pv;                                    if (value != null)                                        p.val = value;                                    else if (t.removetreenode(p))                                        settabat(tab, i, untreeify(t.first));                                }                            }                        }                    }                }                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 = table;        while (tab != null && i < tab.length) {            int fh;            node<k,v> f = tabat(tab, i);            if (f == null)                ++i;            else if ((fh = f.hash) == moved) {                tab = helptransfer(tab, f);                i = 0; // restart            }            else {                synchronized (f) {                    if (tabat(tab, i) == f) {                        node<k,v> p = (fh >= 0 ? f :                                       (f instanceof treebin) ?                                       ((treebin<k,v>)f).first : null);                        while (p != null) {                            --delta;                            p = p.next;                        }                        settabat(tab, i++, null);                    }                }            }        }        if (delta != 0l)            addcount(delta, -1);    }
    /**     * returns a {@link 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 {@code iterator.remove}, {@code set.remove},     * {@code removeall}, {@code retainall}, and {@code clear}     * operations.  it does not support the {@code add} or     * {@code addall} operations.     *     * <p>the view's iterators and spliterators are     * <a href="package-summary.html#weakly"><i>weakly consistent</i></a>.     *     * <p>the view's {@code spliterator} reports {@link spliterator#concurrent},     * {@link spliterator#distinct}, and {@link spliterator#nonnull}.     *     * @return the set view     */    public keysetview<k,v> keyset() {        keysetview<k,v> ks;        return (ks = keyset) != null ? ks : (keyset = new keysetview<k,v>(this, null));    }
    /**     * returns a {@link 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 {@code iterator.remove},     * {@code collection.remove}, {@code removeall},     * {@code retainall}, and {@code clear} operations.  it does not     * support the {@code add} or {@code addall} operations.     *     * <p>the view's iterators and spliterators are     * <a href="package-summary.html#weakly"><i>weakly consistent</i></a>.     *     * <p>the view's {@code spliterator} reports {@link spliterator#concurrent}     * and {@link spliterator#nonnull}.     *     * @return the collection view     */    public collection<v> values() {        valuesview<k,v> vs;        return (vs = values) != null ? vs : (values = new valuesview<k,v>(this));    }
    /**     * returns a {@link 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 {@code iterator.remove}, {@code set.remove},     * {@code removeall}, {@code retainall}, and {@code clear}     * operations.     *     * <p>the view's iterators and spliterators are     * <a href="package-summary.html#weakly"><i>weakly consistent</i></a>.     *     * <p>the view's {@code spliterator} reports {@link spliterator#concurrent},     * {@link spliterator#distinct}, and {@link spliterator#nonnull}.     *     * @return the set view     */    public set<map.entry<k,v>> entryset() {        entrysetview<k,v> es;        return (es = entryset) != null ? es : (entryset = new entrysetview<k,v>(this));    }
    /**     * returns the hash code value for this {@link map}, i.e.,     * the sum of, for each key-value pair in the map,     * {@code key.hashcode() ^ value.hashcode()}.     *     * @return the hash code value for this map     */    public int hashcode() {        int h = 0;        node<k,v>[] t;        if ((t = table) != null) {            traverser<k,v> it = new traverser<k,v>(t, t.length, 0, t.length);            for (node<k,v> p; (p = it.advance()) != null; )                h += p.key.hashcode() ^ p.val.hashcode();        }        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 ("{@code {}}").  adjacent     * mappings are separated by the characters {@code ", "} (comma     * and space).  each key-value mapping is rendered as the key     * followed by an equals sign ("{@code =}") followed by the     * associated value.     *     * @return a string representation of this map     */    public string tostring() {        node<k,v>[] t;        int f = (t = table) == null ? 0 : t.length;        traverser<k,v> it = new traverser<k,v>(t, f, 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)" : k);                sb.append('=');                sb.append(v == this ? "(this map)" : v);                if ((p = it.advance()) == null)                    break;                sb.append(',').append(' ');            }        }        return sb.append('}').tostring();    }
    /**     * compares the specified object with this map for equality.     * returns {@code 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.     *     * @param o object to be compared for equality with this map     * @return {@code true} if the specified object is equal to this map     */    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 = table) == null ? 0 : t.length;            traverser<k,v> it = new traverser<k,v>(t, f, 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 && !v.equals(val)))                    return false;            }            for (map.entry<?,?> e : m.entryset()) {                object mk, mv, v;                if ((mk = e.getkey()) == null ||                    (mv = e.getvalue()) == null ||                    (v = get(mk)) == null ||                    (mv != v && !mv.equals(v)))                    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.loadfactor = lf; }    }
    /**     * saves the state of the {@code concurrenthashmap} instance to a     * stream (i.e., serializes it).     * @param 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.     */    private void writeobject(java.io.objectoutputstream s)        throws java.io.ioexception {        // for serialization compatibility        // emulate segment calculation from previous version of this class        int sshift = 0;        int ssize = 1;        while (ssize < default_concurrency_level) {            ++sshift;            ssize <<= 1;        }        int segmentshift = 32 - sshift;        int segmentmask = ssize - 1;        @suppresswarnings("unchecked")        segment<k,v>[] segments = (segment<k,v>[])            new segment<?,?>[default_concurrency_level];        for (int i = 0; i < segments.length; ++i)            segments[i] = new segment<k,v>(load_factor);        s.putfields().put("segments", segments);        s.putfields().put("segmentshift", segmentshift);        s.putfields().put("segmentmask", segmentmask);        s.writefields();
        node<k,v>[] t;        if ((t = table) != null) {            traverser<k,v> it = new traverser<k,v>(t, t.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).     * @param s the stream     * @throws classnotfoundexception if the class of a serialized object     *         could not be found     * @throws java.io.ioexception if an i/o error occurs     */    private void readobject(java.io.objectinputstream s)        throws java.io.ioexception, classnotfoundexception {        /*         * 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.         */        sizectl = -1; // force exclusion for table construction        s.defaultreadobject();        long size = 0l;        node<k,v> p = null;        for (;;) {            @suppresswarnings("unchecked")            k k = (k) s.readobject();            @suppresswarnings("unchecked")            v v = (v) s.readobject();            if (k != null && v != null) {                p = new node<k,v>(spread(k.hashcode()), k, v, p);                ++size;            }            else                break;        }        if (size == 0l)            sizectl = 0;        else {            int n;            if (size >= (long)(maximum_capacity >>> 1))                n = maximum_capacity;            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.next, first;                int h = p.hash, j = h & mask;                if ((first = tabat(tab, j)) == null)                    insertatfront = true;                else {                    k k = p.key;                    if (first.hash < 0) {                        treebin<k,v> t = (treebin<k,v>)first;                        if (t.puttreeval(h, k, p.val) == null)                            ++added;                        insertatfront = false;                    }                    else {                        int bincount = 0;                        insertatfront = true;                        node<k,v> q; k qk;                        for (q = first; q != null; q = q.next) {                            if (q.hash == h &&                                ((qk = q.key) == k ||                                 (qk != null && k.equals(qk)))) {                                insertatfront = false;                                break;                            }                            ++bincount;                        }                        if (insertatfront && bincount >= treeify_threshold) {                            insertatfront = false;                            ++added;                            p.next = first;                            treenode<k,v> hd = null, tl = null;                            for (q = p; q != null; q = q.next) {                                treenode<k,v> t = new treenode<k,v>                                    (q.hash, q.key, q.val, null, null);                                if ((t.prev = tl) == null)                                    hd = t;                                else                                    tl.next = t;                                tl = t;                            }                            settabat(tab, j, new treebin<k,v>(hd));                        }                    }                }                if (insertatfront) {                    ++added;                    p.next = first;                    settabat(tab, j, p);                }                p = next;            }            table = tab;            sizectl = n - (n >>> 2);            basecount = added;        }    }
    // concurrentmap methods
    /**     * {@inheritdoc}     *     * @return the previous value associated with the specified key,     *         or {@code null} if there was no mapping for the key     * @throws nullpointerexception if the specified key or value is null     */    public v putifabsent(k key, v value) {        return putval(key, value, true);    }
    /**     * {@inheritdoc}     *     * @throws nullpointerexception if the specified key is null     */    public boolean remove(object key, object value) {        if (key == null)            throw new nullpointerexception();        return value != null && replacenode(key, null, value) != null;    }
    /**     * {@inheritdoc}     *     * @throws nullpointerexception if any of the arguments are null     */    public boolean replace(k key, v oldvalue, v newvalue) {        if (key == null || oldvalue == null || newvalue == null)            throw new nullpointerexception();        return replacenode(key, newvalue, oldvalue) != null;    }
    /**     * {@inheritdoc}     *     * @return the previous value associated with the specified key,     *         or {@code null} if there was no mapping for the key     * @throws nullpointerexception if the specified key or value is null     */    public v replace(k key, v value) {        if (key == null || value == null)            throw new nullpointerexception();        return replacenode(key, value, null);    }
    // overrides of jdk8+ map extension method defaults
    /**     * returns the value to which the specified key is mapped, or the     * given default value if this map contains no mapping for the     * key.     *     * @param key the key whose associated value is to be returned     * @param defaultvalue the value to return if this map contains     * no mapping for the given key     * @return the mapping for the key, if present; else the default value     * @throws nullpointerexception if the specified key is null     */    public v getordefault(object key, v defaultvalue) {        v v;        return (v = get(key)) == null ? defaultvalue : v;    }
    public void foreach(biconsumer<? super k, ? super v> action) {        if (action == null) throw new nullpointerexception();        node<k,v>[] t;        if ((t = table) != null) {            traverser<k,v> it = new traverser<k,v>(t, t.length, 0, t.length);            for (node<k,v> p; (p = it.advance()) != null; ) {                action.accept(p.key, p.val);            }        }    }
    public void replaceall(bifunction<? super k, ? super v, ? extends v> function) {        if (function == null) throw new nullpointerexception();        node<k,v>[] t;        if ((t = table) != null) {            traverser<k,v> it = new traverser<k,v>(t, t.length, 0, t.length);            for (node<k,v> p; (p = it.advance()) != null; ) {                v oldvalue = p.val;                for (k key = p.key;;) {                    v newvalue = function.apply(key, oldvalue);                    if (newvalue == null)                        throw new nullpointerexception();                    if (replacenode(key, newvalue, oldvalue) != null ||                        (oldvalue = get(key)) == null)                        break;                }            }        }    }
    /**     * if the specified key is not already associated with a value,     * attempts to compute its value using the given mapping function     * and enters it into this map unless {@code null}.  the entire     * method invocation is performed atomically, so the function is     * applied at most once per key.  some attempted update operations     * on this map by other threads may be blocked while computation     * is in progress, so the computation should be short and simple,     * and must not attempt to update any other mappings of this map.     *     * @param key key with which the specified value is to be associated     * @param mappingfunction the function to compute a value     * @return the current (existing or computed) value associated with     *         the specified key, or null if the computed value is null     * @throws nullpointerexception if the specified key or mappingfunction     *         is null     * @throws illegalstateexception if the computation detectably     *         attempts a recursive update to this map that would     *         otherwise never complete     * @throws runtimeexception or error if the mappingfunction does so,     *         in which case the mapping is left unestablished     */    public v computeifabsent(k key, function<? super k, ? extends v> mappingfunction) {        if (key == null || mappingfunction == null)            throw new nullpointerexception();        int h = spread(key.hashcode());        v val = null;        int bincount = 0;        for (node<k,v>[] tab = table;;) {            node<k,v> f; int n, i, fh;            if (tab == null || (n = tab.length) == 0)                tab = inittable();            else if ((f = tabat(tab, i = (n - 1) & h)) == null) {                node<k,v> r = new reservationnode<k,v>();                synchronized (r) {                    if (castabat(tab, i, null, r)) {                        bincount = 1;                        node<k,v> node = null;                        try {                            if ((val = mappingfunction.apply(key)) != null)                                node = new node<k,v>(h, key, val, null);                        } finally {                            settabat(tab, i, node);                        }                    }                }                if (bincount != 0)                    break;            }            else if ((fh = f.hash) == moved)                tab = helptransfer(tab, f);            else {                boolean added = false;                synchronized (f) {                    if (tabat(tab, i) == f) {                        if (fh >= 0) {                            bincount = 1;                            for (node<k,v> e = f;; ++bincount) {                                k ek; v ev;                                if (e.hash == h &&                                    ((ek = e.key) == key ||                                     (ek != null && key.equals(ek)))) {                                    val = e.val;                                    break;                                }                                node<k,v> pred = e;                                if ((e = e.next) == null) {                                    if ((val = mappingfunction.apply(key)) != null) {                                        added = true;                                        pred.next = new node<k,v>(h, key, val, null);                                    }                                    break;                                }                            }                        }                        else if (f instanceof treebin) {                            bincount = 2;                            treebin<k,v> t = (treebin<k,v>)f;                            treenode<k,v> r, p;                            if ((r = t.root) != null &&                                (p = r.findtreenode(h, key, null)) != null)                                val = p.val;                            else if ((val = mappingfunction.apply(key)) != null) {                                added = true;                                t.puttreeval(h, key, val);                            }                        }                    }                }                if (bincount != 0) {                    if (bincount >= treeify_threshold)                        treeifybin(tab, i);                    if (!added)                        return val;                    break;                }            }        }        if (val != null)            addcount(1l, bincount);        return val;    }
    /**     * if the value for the specified key is present, attempts to     * compute a new mapping given the key and its current mapped     * value.  the entire method invocation is performed atomically.     * some attempted update operations on this map by other threads     * may be blocked while computation is in progress, so the     * computation should be short and simple, and must not attempt to     * update any other mappings of this map.     *     * @param key key with which a value may be associated     * @param remappingfunction the function to compute a value     * @return the new value associated with the specified key, or null if none     * @throws nullpointerexception if the specified key or remappingfunction     *         is null     * @throws illegalstateexception if the computation detectably     *         attempts a recursive update to this map that would     *         otherwise never complete     * @throws runtimeexception or error if the remappingfunction does so,     *         in which case the mapping is unchanged     */    public v computeifpresent(k key, bifunction<? super k, ? super v, ? extends v> remappingfunction) {        if (key == null || remappingfunction == null)            throw new nullpointerexception();        int h = spread(key.hashcode());        v val = null;        int delta = 0;        int bincount = 0;        for (node<k,v>[] tab = table;;) {            node<k,v> f; int n, i, fh;            if (tab == null || (n = tab.length) == 0)                tab = inittable();            else if ((f = tabat(tab, i = (n - 1) & h)) == null)                break;            else if ((fh = f.hash) == moved)                tab = helptransfer(tab, f);            else {                synchronized (f) {                    if (tabat(tab, i) == f) {                        if (fh >= 0) {                            bincount = 1;                            for (node<k,v> e = f, pred = null;; ++bincount) {                                k ek;                                if (e.hash == h &&                                    ((ek = e.key) == key ||                                     (ek != null && key.equals(ek)))) {                                    val = remappingfunction.apply(key, e.val);                                    if (val != null)                                        e.val = val;                                    else {                                        delta = -1;                                        node<k,v> en = e.next;                                        if (pred != null)                                            pred.next = en;                                        else                                            settabat(tab, i, en);                                    }                                    break;                                }                                pred = e;                                if ((e = e.next) == null)                                    break;                            }                        }                        else if (f instanceof treebin) {                            bincount = 2;                            treebin<k,v> t = (treebin<k,v>)f;                            treenode<k,v> r, p;                            if ((r = t.root) != null &&                                (p = r.findtreenode(h, key, null)) != null) {                                val = remappingfunction.apply(key, p.val);                                if (val != null)                                    p.val = val;                                else {                                    delta = -1;                                    if (t.removetreenode(p))                                        settabat(tab, i, untreeify(t.first));                                }                            }                        }                    }                }                if (bincount != 0)                    break;            }        }        if (delta != 0)            addcount((long)delta, bincount);        return val;    }
    /**     * attempts to compute a mapping for the specified key and its     * current mapped value (or {@code null} if there is no current     * mapping). the entire method invocation is performed atomically.     * some attempted update operations on this map by other threads     * may be blocked while computation is in progress, so the     * computation should be short and simple, and must not attempt to     * update any other mappings of this map.     *     * @param key key with which the specified value is to be associated     * @param remappingfunction the function to compute a value     * @return the new value associated with the specified key, or null if none     * @throws nullpointerexception if the specified key or remappingfunction     *         is null     * @throws illegalstateexception if the computation detectably     *         attempts a recursive update to this map that would     *         otherwise never complete     * @throws runtimeexception or error if the remappingfunction does so,     *         in which case the mapping is unchanged     */    public v compute(k key,                     bifunction<? super k, ? super v, ? extends v> remappingfunction) {        if (key == null || remappingfunction == null)            throw new nullpointerexception();        int h = spread(key.hashcode());        v val = null;        int delta = 0;        int bincount = 0;        for (node<k,v>[] tab = table;;) {            node<k,v> f; int n, i, fh;            if (tab == null || (n = tab.length) == 0)                tab = inittable();            else if ((f = tabat(tab, i = (n - 1) & h)) == null) {                node<k,v> r = new reservationnode<k,v>();                synchronized (r) {                    if (castabat(tab, i, null, r)) {                        bincount = 1;                        node<k,v> node = null;                        try {                            if ((val = remappingfunction.apply(key, null)) != null) {                                delta = 1;                                node = new node<k,v>(h, key, val, null);                            }                        } finally {                            settabat(tab, i, node);                        }                    }                }                if (bincount != 0)                    break;            }            else if ((fh = f.hash) == moved)                tab = helptransfer(tab, f);            else {                synchronized (f) {                    if (tabat(tab, i) == f) {                        if (fh >= 0) {                            bincount = 1;                            for (node<k,v> e = f, pred = null;; ++bincount) {                                k ek;                                if (e.hash == h &&                                    ((ek = e.key) == key ||                                     (ek != null && key.equals(ek)))) {                                    val = remappingfunction.apply(key, e.val);                                    if (val != null)                                        e.val = val;                                    else {                                        delta = -1;                                        node<k,v> en = e.next;                                        if (pred != null)                                            pred.next = en;                                        else                                            settabat(tab, i, en);                                    }                                    break;                                }                                pred = e;                                if ((e = e.next) == null) {                                    val = remappingfunction.apply(key, null);                                    if (val != null) {                                        delta = 1;                                        pred.next =                                            new node<k,v>(h, key, val, null);                                    }                                    break;                                }                            }                        }                        else if (f instanceof treebin) {                            bincount = 1;                            treebin<k,v> t = (treebin<k,v>)f;                            treenode<k,v> r, p;                            if ((r = t.root) != null)                                p = r.findtreenode(h, key, null);                            else                                p = null;                            v pv = (p == null) ? null : p.val;                            val = remappingfunction.apply(key, pv);                            if (val != null) {                                if (p != null)                                    p.val = val;                                else {                                    delta = 1;                                    t.puttreeval(h, key, val);                                }                            }                            else if (p != null) {                                delta = -1;                                if (t.removetreenode(p))                                    settabat(tab, i, untreeify(t.first));                            }                        }                    }                }                if (bincount != 0) {                    if (bincount >= treeify_threshold)                        treeifybin(tab, i);                    break;                }            }        }        if (delta != 0)            addcount((long)delta, bincount);        return val;    }
    /**     * if the specified key is not already associated with a     * (non-null) value, associates it with the given value.     * otherwise, replaces the value with the results of the given     * remapping function, or removes if {@code null}. the entire     * method invocation is performed atomically.  some attempted     * update operations on this map by other threads may be blocked     * while computation is in progress, so the computation should be     * short and simple, and must not attempt to update any other     * mappings of this map.     *     * @param key key with which the specified value is to be associated     * @param value the value to use if absent     * @param remappingfunction the function to recompute a value if present     * @return the new value associated with the specified key, or null if none     * @throws nullpointerexception if the specified key or the     *         remappingfunction is null     * @throws runtimeexception or error if the remappingfunction does so,     *         in which case the mapping is unchanged     */    public v merge(k key, v value, bifunction<? super v, ? super v, ? extends v> remappingfunction) {        if (key == null || value == null || remappingfunction == null)            throw new nullpointerexception();        int h = spread(key.hashcode());        v val = null;        int delta = 0;        int bincount = 0;        for (node<k,v>[] tab = table;;) {            node<k,v> f; int n, i, fh;            if (tab == null || (n = tab.length) == 0)                tab = inittable();            else if ((f = tabat(tab, i = (n - 1) & h)) == null) {                if (castabat(tab, i, null, new node<k,v>(h, key, value, null))) {                    delta = 1;                    val = value;                    break;                }            }            else if ((fh = f.hash) == moved)                tab = helptransfer(tab, f);            else {                synchronized (f) {                    if (tabat(tab, i) == f) {                        if (fh >= 0) {                            bincount = 1;                            for (node<k,v> e = f, pred = null;; ++bincount) {                                k ek;                                if (e.hash == h &&                                    ((ek = e.key) == key ||                                     (ek != null && key.equals(ek)))) {                                    val = remappingfunction.apply(e.val, value);                                    if (val != null)                                        e.val = val;                                    else {                                        delta = -1;                                        node<k,v> en = e.next;                                        if (pred != null)                                            pred.next = en;                                        else                                            settabat(tab, i, en);                                    }                                    break;                                }                                pred = e;                                if ((e = e.next) == null) {                                    delta = 1;                                    val = value;                                    pred.next =                                        new node<k,v>(h, key, val, null);                                    break;                                }                            }                        }                        else if (f instanceof treebin) {                            bincount = 2;                            treebin<k,v> t = (treebin<k,v>)f;                            treenode<k,v> r = t.root;                            treenode<k,v> p = (r == null) ? null :                                r.findtreenode(h, key, null);                            val = (p == null) ? value :                                remappingfunction.apply(p.val, value);                            if (val != null) {                                if (p != null)                                    p.val = val;                                else {                                    delta = 1;                                    t.puttreeval(h, key, val);                                }                            }                            else if (p != null) {                                delta = -1;                                if (t.removetreenode(p))                                    settabat(tab, i, untreeify(t.first));                            }                        }                    }                }                if (bincount != 0) {                    if (bincount >= treeify_threshold)                        treeifybin(tab, i);                    break;                }            }        }        if (delta != 0)            addcount((long)delta, bincount);        return val;    }
    // hashtable legacy methods
    /**     * legacy method testing if some key maps into the specified value     * in this table.  this method is identical in functionality to     * {@link #containsvalue(object)}, and exists solely to ensure     * full compatibility with class {@link java.util.hashtable},     * which supported this method prior to introduction of the     * java collections framework.     *     * @param  value a value to search for     * @return {@code true} if and only if some key maps to the     *         {@code value} argument in this table as     *         determined by the {@code equals} method;     *         {@code false} otherwise     * @throws nullpointerexception if the specified value is null     */    public boolean contains(object value) {        return containsvalue(value);    }
    /**     * returns an enumeration of the keys in this table.     *     * @return an enumeration of the keys in this table     * @see #keyset()     */    public enumeration<k> keys() {        node<k,v>[] t;        int f = (t = table) == null ? 0 : t.length;        return new keyiterator<k,v>(t, f, 0, f, this);    }
    /**     * returns an enumeration of the values in this table.     *     * @return an enumeration of the values in this table     * @see #values()     */    public enumeration<v> elements() {        node<k,v>[] t;        int f = (t = table) == null ? 0 : t.length;        return new valueiterator<k,v>(t, f, 0, f, this);    }
    // concurrenthashmap-only methods
    /**     * returns the number of mappings. this method should be used     * instead of {@link #size} because a concurrenthashmap may     * contain more mappings than can be represented as an int. the     * value returned is an estimate; the actual count may differ if     * there are concurrent insertions or removals.     *     * @return the number of mappings     * @since 1.8     */    public long mappingcount() {        long n = sumcount();        return (n < 0l) ? 0l : n; // ignore transient negative values    }
    /**     * creates a new {@link set} backed by a concurrenthashmap     * from the given type to {@code boolean.true}.     *     * @param <k> the element type of the returned set     * @return the new set     * @since 1.8     */    public static <k> keysetview<k,boolean> newkeyset() {        return new keysetview<k,boolean>            (new concurrenthashmap<k,boolean>(), boolean.true);    }
    /**     * creates a new {@link set} backed by a concurrenthashmap     * from the given type to {@code boolean.true}.     *     * @param initialcapacity the implementation performs internal     * sizing to accommodate this many elements.     * @param <k> the element type of the returned set     * @return the new set     * @throws illegalargumentexception if the initial capacity of     * elements is negative     * @since 1.8     */    public static <k> keysetview<k,boolean> newkeyset(int initialcapacity) {        return new keysetview<k,boolean>            (new concurrenthashmap<k,boolean>(initialcapacity), boolean.true);    }
    /**     * returns a {@link set} view of the keys in this map, using the     * given common mapped value for any additions (i.e., {@link     * collection#add} and {@link collection#addall(collection)}).     * this is of course only appropriate if it is acceptable to use     * the same value for all additions from this view.     *     * @param mappedvalue the mapped value to use for any additions     * @return the set view     * @throws nullpointerexception if the mappedvalue is null     */    public keysetview<k,v> keyset(v mappedvalue) {        if (mappedvalue == null)            throw new nullpointerexception();        return new keysetview<k,v>(this, mappedvalue);    }
    /* ---------------- special nodes -------------- */
    /**     * a node inserted at head of bins during transfer operations.     */    static final class forwardingnode<k,v> extends node<k,v> {        final node<k,v>[] nexttable;        forwardingnode(node<k,v>[] tab) {            super(moved, null, null, null);            this.nexttable = tab;        }
        node<k,v> find(int h, object k) {            // loop to avoid arbitrarily deep recursion on forwarding nodes            outer: for (node<k,v>[] tab = nexttable;;) {                node<k,v> e; int n;                if (k == null || tab == null || (n = tab.length) == 0 ||                    (e = tabat(tab, (n - 1) & h)) == null)                    return null;                for (;;) {                    int eh; k ek;                    if ((eh = e.hash) == h &&                        ((ek = e.key) == k || (ek != null && k.equals(ek))))                        return e;                    if (eh < 0) {                        if (e instanceof forwardingnode) {                            tab = ((forwardingnode<k,v>)e).nexttable;                            continue outer;                        }                        else                            return e.find(h, k);                    }                    if ((e = e.next) == null)                        return null;                }            }        }    }
    /**     * a place-holder node used in computeifabsent and compute     */    static final class reservationnode<k,v> extends node<k,v> {        reservationnode() {            super(reserved, null, null, null);        }
        node<k,v> find(int h, object k) {            return null;        }    }
    /* ---------------- table initialization and resizing -------------- */
    /**     * returns the stamp bits for resizing a table of size n.     * must be negative when shifted left by resize_stamp_shift.     */    static final int resizestamp(int n) {        return integer.numberofleadingzeros(n) | (1 << (resize_stamp_bits - 1));    }
    /**     * initializes table, using the size recorded in sizectl.     */    private final node<k,v>[] inittable() {        node<k,v>[] tab; int sc;        while ((tab = table) == null || tab.length == 0) {            if ((sc = sizectl) < 0)                thread.yield(); // lost initialization race; just spin            else if (u.compareandswapint(this, sizectl, sc, -1)) {                try {                    if ((tab = table) == null || tab.length == 0) {                        int n = (sc > 0) ? sc : default_capacity;                        @suppresswarnings("unchecked")                        node<k,v>[] nt = (node<k,v>[])new node<?,?>[n];                        table = tab = nt;                        sc = n - (n >>> 2);                    }                } finally {                    sizectl = sc;                }                break;            }        }        return tab;    }
    /**     * adds to count, and if table is too small and not already     * resizing, initiates transfer. if already resizing, helps     * perform transfer if work is available.  rechecks occupancy     * after a transfer to see if another resize is already needed     * because resizings are lagging additions.     *     * @param x the count to add     * @param check if <0, don't check resize, if <= 1 only check if uncontended     */    private final void addcount(long x, int check) {        countercell[] as; long b, s;        if ((as = countercells) != null ||            !u.compareandswaplong(this, basecount, b = basecount, s = b + x)) {            countercell a; long v; int m;            boolean uncontended = true;            if (as == null || (m = as.length - 1) < 0 ||                (a = as[threadlocalrandom.getprobe() & m]) == null ||                !(uncontended =                  u.compareandswaplong(a, cellvalue, v = a.value, v + x))) {                fulladdcount(x, uncontended);                return;            }            if (check <= 1)                return;            s = sumcount();        }        if (check >= 0) {            node<k,v>[] tab, nt; int n, sc;            while (s >= (long)(sc = sizectl) && (tab = table) != null &&                   (n = tab.length) < maximum_capacity) {                int rs = resizestamp(n);                if (sc < 0) {                    if ((sc >>> resize_stamp_shift) != rs || sc == rs + 1 ||                        sc == rs + max_resizers || (nt = nexttable) == null ||                        transferindex <= 0)                        break;                    if (u.compareandswapint(this, sizectl, sc, sc + 1))                        transfer(tab, nt);                }                else if (u.compareandswapint(this, sizectl, sc,                                             (rs << resize_stamp_shift) + 2))                    transfer(tab, null);                s = sumcount();            }        }    }
    /**     * helps transfer if a resize is in progress.     */    final node<k,v>[] helptransfer(node<k,v>[] tab, node<k,v> f) {        node<k,v>[] nexttab; int sc;        if (tab != null && (f instanceof forwardingnode) &&            (nexttab = ((forwardingnode<k,v>)f).nexttable) != null) {            int rs = resizestamp(tab.length);            while (nexttab == nexttable && table == tab &&                   (sc = sizectl) < 0) {                if ((sc >>> resize_stamp_shift) != rs || sc == rs + 1 ||                    sc == rs + max_resizers || transferindex <= 0)                    break;                if (u.compareandswapint(this, sizectl, sc, sc + 1)) {                    transfer(tab, nexttab);                    break;                }            }            return nexttab;        }        return table;    }
    /**     * tries to presize table to accommodate the given number of elements.     *     * @param size number of elements (doesn't need to be perfectly accurate)     */    private final void trypresize(int size) {        int c = (size >= (maximum_capacity >>> 1)) ? maximum_capacity :            tablesizefor(size + (size >>> 1) + 1);        int sc;        while ((sc = sizectl) >= 0) {            node<k,v>[] tab = table; int n;            if (tab == null || (n = tab.length) == 0) {                n = (sc > c) ? sc : c;                if (u.compareandswapint(this, sizectl, sc, -1)) {                    try {                        if (table == tab) {                            @suppresswarnings("unchecked")                            node<k,v>[] nt = (node<k,v>[])new node<?,?>[n];                            table = nt;                            sc = n - (n >>> 2);                        }                    } finally {                        sizectl = sc;                    }                }            }            else if (c <= sc || n >= maximum_capacity)                break;            else if (tab == table) {                int rs = resizestamp(n);                if (sc < 0) {                    node<k,v>[] nt;                    if ((sc >>> resize_stamp_shift) != rs || sc == rs + 1 ||                        sc == rs + max_resizers || (nt = nexttable) == null ||                        transferindex <= 0)                        break;                    if (u.compareandswapint(this, sizectl, sc, sc + 1))                        transfer(tab, nt);                }                else if (u.compareandswapint(this, sizectl, sc,                                             (rs << resize_stamp_shift) + 2))                    transfer(tab, null);            }        }    }
    /**     * moves and/or copies the nodes in each bin to new table. see     * above for explanation.     */    private final void transfer(node<k,v>[] tab, node<k,v>[] nexttab) {        int n = tab.length, stride;        if ((stride = (ncpu > 1) ? (n >>> 3) / ncpu : n) < min_transfer_stride)            stride = min_transfer_stride; // subdivide range        if (nexttab == null) {            // initiating            try {                @suppresswarnings("unchecked")                node<k,v>[] nt = (node<k,v>[])new node<?,?>[n << 1];                nexttab = nt;            } catch (throwable ex) {      // try to cope with oome                sizectl = integer.max_value;                return;            }            nexttable = nexttab;            transferindex = n;        }        int nextn = nexttab.length;        forwardingnode<k,v> fwd = new forwardingnode<k,v>(nexttab);        boolean advance = true;        boolean finishing = false; // to ensure sweep before committing nexttab        for (int i = 0, bound = 0;;) {            node<k,v> f; int fh;            while (advance) {                int nextindex, nextbound;                if (--i >= bound || finishing)                    advance = false;                else if ((nextindex = transferindex) <= 0) {                    i = -1;                    advance = false;                }                else if (u.compareandswapint                         (this, transferindex, nextindex,                          nextbound = (nextindex > stride ?                                       nextindex - stride : 0))) {                    bound = nextbound;                    i = nextindex - 1;                    advance = false;                }            }            if (i < 0 || i >= n || i + n >= nextn) {                int sc;                if (finishing) {                    nexttable = null;                    table = nexttab;                    sizectl = (n << 1) - (n >>> 1);                    return;                }                if (u.compareandswapint(this, sizectl, sc = sizectl, sc - 1)) {                    if ((sc - 2) != resizestamp(n) << resize_stamp_shift)                        return;                    finishing = advance = true;                    i = n; // recheck before commit                }            }            else if ((f = tabat(tab, i)) == null)                advance = castabat(tab, i, null, fwd);            else if ((fh = f.hash) == moved)                advance = true; // already processed            else {                synchronized (f) {                    if (tabat(tab, i) == f) {                        node<k,v> ln, hn;                        if (fh >= 0) {                            int runbit = fh & n;                            node<k,v> lastrun = f;                            for (node<k,v> p = f.next; p != null; p = p.next) {                                int b = p.hash & n;                                if (b != runbit) {                                    runbit = b;                                    lastrun = p;                                }                            }                            if (runbit == 0) {                                ln = lastrun;                                hn = null;                            }                            else {                                hn = lastrun;                                ln = null;                            }                            for (node<k,v> p = f; p != lastrun; p = p.next) {                                int ph = p.hash; k pk = p.key; v pv = p.val;                                if ((ph & n) == 0)                                    ln = new node<k,v>(ph, pk, pv, ln);                                else                                    hn = new node<k,v>(ph, pk, pv, hn);                            }                            settabat(nexttab, i, ln);                            settabat(nexttab, i + n, hn);                            settabat(tab, i, fwd);                            advance = true;                        }                        else if (f instanceof treebin) {                            treebin<k,v> t = (treebin<k,v>)f;                            treenode<k,v> lo = null, lotail = null;                            treenode<k,v> hi = null, hitail = null;                            int lc = 0, hc = 0;                            for (node<k,v> e = t.first; e != null; e = e.next) {                                int h = e.hash;                                treenode<k,v> p = new treenode<k,v>                                    (h, e.key, e.val, null, null);                                if ((h & n) == 0) {                                    if ((p.prev = lotail) == null)                                        lo = p;                                    else                                        lotail.next = p;                                    lotail = p;                                    ++lc;                                }                                else {                                    if ((p.prev = hitail) == null)                                        hi = p;                                    else                                        hitail.next = p;                                    hitail = p;                                    ++hc;                                }                            }                            ln = (lc <= untreeify_threshold) ? untreeify(lo) :                                (hc != 0) ? new treebin<k,v>(lo) : t;                            hn = (hc <= untreeify_threshold) ? untreeify(hi) :                                (lc != 0) ? new treebin<k,v>(hi) : t;                            settabat(nexttab, i, ln);                            settabat(nexttab, i + n, hn);                            settabat(tab, i, fwd);                            advance = true;                        }                    }                }            }        }    }
    /* ---------------- counter support -------------- */
    /**     * a padded cell for distributing counts.  adapted from longadder     * and striped64.  see their internal docs for explanation.     */    @sun.misc.contended static final class countercell {        volatile long value;        countercell(long x) { value = x; }    }
    final long sumcount() {        countercell[] as = countercells; countercell a;        long sum = basecount;        if (as != null) {            for (int i = 0; i < as.length; ++i) {                if ((a = as[i]) != null)                    sum += a.value;            }        }        return sum;    }
    // see longadder version for explanation    private final void fulladdcount(long x, boolean wasuncontended) {        int h;        if ((h = threadlocalrandom.getprobe()) == 0) {            threadlocalrandom.localinit();      // force initialization            h = threadlocalrandom.getprobe();            wasuncontended = true;        }        boolean collide = false;                // true if last slot nonempty        for (;;) {            countercell[] as; countercell a; int n; long v;            if ((as = countercells) != null && (n = as.length) > 0) {                if ((a = as[(n - 1) & h]) == null) {                    if (cellsbusy == 0) {            // try to attach new cell                        countercell r = new countercell(x); // optimistic create                        if (cellsbusy == 0 &&                            u.compareandswapint(this, cellsbusy, 0, 1)) {                            boolean created = false;                            try {               // recheck under lock                                countercell[] rs; int m, j;                                if ((rs = countercells) != null &&                                    (m = rs.length) > 0 &&                                    rs[j = (m - 1) & h] == null) {                                    rs[j] = r;                                    created = true;                                }                            } finally {                                cellsbusy = 0;                            }                            if (created)                                break;                            continue;           // slot is now non-empty                        }                    }                    collide = false;                }                else if (!wasuncontended)       // cas already known to fail                    wasuncontended = true;      // continue after rehash                else if (u.compareandswaplong(a, cellvalue, v = a.value, v + x))                    break;                else if (countercells != as || n >= ncpu)                    collide = false;            // at max size or stale                else if (!collide)                    collide = true;                else if (cellsbusy == 0 &&                         u.compareandswapint(this, cellsbusy, 0, 1)) {                    try {                        if (countercells == as) {// expand table unless stale                            countercell[] rs = new countercell[n << 1];                            for (int i = 0; i < n; ++i)                                rs[i] = as[i];                            countercells = rs;                        }                    } finally {                        cellsbusy = 0;                    }                    collide = false;                    continue;                   // retry with expanded table                }                h = threadlocalrandom.advanceprobe(h);            }            else if (cellsbusy == 0 && countercells == as &&                     u.compareandswapint(this, cellsbusy, 0, 1)) {                boolean init = false;                try {                           // initialize table                    if (countercells == as) {                        countercell[] rs = new countercell[2];                        rs[h & 1] = new countercell(x);                        countercells = rs;                        init = true;                    }                } finally {                    cellsbusy = 0;                }                if (init)                    break;            }            else if (u.compareandswaplong(this, basecount, v = basecount, v + x))                break;                          // fall back on using base        }    }
    /* ---------------- conversion from/to treebins -------------- */
    /**     * replaces all linked nodes in bin at given index unless table is     * too small, in which case resizes instead.     */    private final void treeifybin(node<k,v>[] tab, int index) {        node<k,v> b; int n, sc;        if (tab != null) {            if ((n = tab.length) < min_treeify_capacity)                trypresize(n << 1);            else if ((b = tabat(tab, index)) != null && b.hash >= 0) {                synchronized (b) {                    if (tabat(tab, index) == b) {                        treenode<k,v> hd = null, tl = null;                        for (node<k,v> e = b; e != null; e = e.next) {                            treenode<k,v> p =                                new treenode<k,v>(e.hash, e.key, e.val,                                                  null, null);                            if ((p.prev = tl) == null)                                hd = p;                            else                                tl.next = p;                            tl = p;                        }                        settabat(tab, index, new treebin<k,v>(hd));                    }                }            }        }    }
    /**     * returns a list on non-treenodes replacing those in given list.     */    static <k,v> node<k,v> untreeify(node<k,v> b) {        node<k,v> hd = null, tl = null;        for (node<k,v> q = b; q != null; q = q.next) {            node<k,v> p = new node<k,v>(q.hash, q.key, q.val, null);            if (tl == null)                hd = p;            else                tl.next = p;            tl = p;        }        return hd;    }
    /* ---------------- treenodes -------------- */
    /**     * nodes for use in treebins     */    static final class treenode<k,v> extends node<k,v> {        treenode<k,v> parent;  // red-black tree links        treenode<k,v> left;        treenode<k,v> right;        treenode<k,v> prev;    // needed to unlink next upon deletion        boolean red;
        treenode(int hash, k key, v val, node<k,v> next,                 treenode<k,v> parent) {            super(hash, key, val, next);            this.parent = parent;        }
        node<k,v> find(int h, object k) {            return findtreenode(h, k, null);        }
        /**         * returns the treenode (or null if not found) for the given key         * starting at given root.         */        final treenode<k,v> findtreenode(int h, object k, class<?> kc) {            if (k != null) {                treenode<k,v> p = this;                do  {                    int ph, dir; k pk; treenode<k,v> q;                    treenode<k,v> pl = p.left, pr = p.right;                    if ((ph = p.hash) > h)                        p = pl;                    else if (ph < h)                        p = pr;                    else if ((pk = p.key) == k || (pk != null && k.equals(pk)))                        return p;                    else if (pl == null)                        p = pr;                    else if (pr == null)                        p = pl;                    else if ((kc != null ||                              (kc = comparableclassfor(k)) != null) &&                             (dir = comparecomparables(kc, k, pk)) != 0)                        p = (dir < 0) ? pl : pr;                    else if ((q = pr.findtreenode(h, k, kc)) != null)                        return q;                    else                        p = pl;                } while (p != null);            }            return null;        }    }
    /* ---------------- treebins -------------- */
    /**     * treenodes used at the heads of bins. treebins do not hold user     * keys or values, but instead point to list of treenodes and     * their root. they also maintain a parasitic read-write lock     * forcing writers (who hold bin lock) to wait for readers (who do     * not) to complete before tree restructuring operations.     */    static final class treebin<k,v> extends node<k,v> {        treenode<k,v> root;        volatile treenode<k,v> first;        volatile thread waiter;        volatile int lockstate;        // values for lockstate        static final int writer = 1; // set while holding write lock        static final int waiter = 2; // set when waiting for write lock        static final int reader = 4; // increment value for setting read lock
        /**         * tie-breaking utility for ordering insertions when equal         * hashcodes and non-comparable. we don't require a total         * order, just a consistent insertion rule to maintain         * equivalence across rebalancings. tie-breaking further than         * necessary simplifies testing a bit.         */        static int tiebreakorder(object a, object b) {            int d;            if (a == null || b == null ||                (d = a.getclass().getname().                 compareto(b.getclass().getname())) == 0)                d = (system.identityhashcode(a) <= system.identityhashcode(b) ?                     -1 : 1);            return d;        }
        /**         * creates bin with initial set of nodes headed by b.         */        treebin(treenode<k,v> b) {            super(treebin, null, null, null);            this.first = b;            treenode<k,v> r = null;            for (treenode<k,v> x = b, next; x != null; x = next) {                next = (treenode<k,v>)x.next;                x.left = x.right = null;                if (r == null) {                    x.parent = null;                    x.red = false;                    r = x;                }                else {                    k k = x.key;                    int h = x.hash;                    class<?> kc = null;                    for (treenode<k,v> p = r;;) {                        int dir, ph;                        k pk = p.key;                        if ((ph = p.hash) > h)                            dir = -1;                        else if (ph < h)                            dir = 1;                        else if ((kc == null &&                                  (kc = comparableclassfor(k)) == null) ||                                 (dir = comparecomparables(kc, k, pk)) == 0)                            dir = tiebreakorder(k, pk);                            treenode<k,v> xp = p;                        if ((p = (dir <= 0) ? p.left : p.right) == null) {                            x.parent = xp;                            if (dir <= 0)                                xp.left = x;                            else                                xp.right = x;                            r = balanceinsertion(r, x);                            break;                        }                    }                }            }            this.root = r;            assert checkinvariants(root);        }
        /**         * acquires write lock for tree restructuring.         */        private final void lockroot() {            if (!u.compareandswapint(this, lockstate, 0, writer))                contendedlock(); // offload to separate method        }
        /**         * releases write lock for tree restructuring.         */        private final void unlockroot() {            lockstate = 0;        }
        /**         * possibly blocks awaiting root lock.         */        private final void contendedlock() {            boolean waiting = false;            for (int s;;) {                if (((s = lockstate) & ~waiter) == 0) {                    if (u.compareandswapint(this, lockstate, s, writer)) {                        if (waiting)                            waiter = null;                        return;                    }                }                else if ((s & waiter) == 0) {                    if (u.compareandswapint(this, lockstate, s, s | waiter)) {                        waiting = true;                        waiter = thread.currentthread();                    }                }                else if (waiting)                    locksupport.park(this);            }        }
        /**         * returns matching node or null if none. tries to search         * using tree comparisons from root, but continues linear         * search when lock not available.         */        final node<k,v> find(int h, object k) {            if (k != null) {                for (node<k,v> e = first; e != null; ) {                    int s; k ek;                    if (((s = lockstate) & (waiter|writer)) != 0) {                        if (e.hash == h &&                            ((ek = e.key) == k || (ek != null && k.equals(ek))))                            return e;                        e = e.next;                    }                    else if (u.compareandswapint(this, lockstate, s,                                                 s + reader)) {                        treenode<k,v> r, p;                        try {                            p = ((r = root) == null ? null :                                 r.findtreenode(h, k, null));                        } finally {                            thread w;                            if (u.getandaddint(this, lockstate, -reader) ==                                (reader|waiter) && (w = waiter) != null)                                locksupport.unpark(w);                        }                        return p;                    }                }            }            return null;        }
        /**         * finds or adds a node.         * @return null if added         */        final treenode<k,v> puttreeval(int h, k k, v v) {            class<?> kc = null;            boolean searched = false;            for (treenode<k,v> p = root;;) {                int dir, ph; k pk;                if (p == null) {                    first = root = new treenode<k,v>(h, k, v, null, null);                    break;                }                else if ((ph = p.hash) > h)                    dir = -1;                else if (ph < h)                    dir = 1;                else if ((pk = p.key) == k || (pk != null && k.equals(pk)))                    return p;                else if ((kc == null &&                          (kc = comparableclassfor(k)) == null) ||                         (dir = comparecomparables(kc, k, pk)) == 0) {                    if (!searched) {                        treenode<k,v> q, ch;                        searched = true;                        if (((ch = p.left) != null &&                             (q = ch.findtreenode(h, k, kc)) != null) ||                            ((ch = p.right) != null &&                             (q = ch.findtreenode(h, k, kc)) != null))                            return q;                    }                    dir = tiebreakorder(k, pk);                }
                treenode<k,v> xp = p;                if ((p = (dir <= 0) ? p.left : p.right) == null) {                    treenode<k,v> x, f = first;                    first = x = new treenode<k,v>(h, k, v, f, xp);                    if (f != null)                        f.prev = x;                    if (dir <= 0)                        xp.left = x;                    else                        xp.right = x;                    if (!xp.red)                        x.red = true;                    else {                        lockroot();                        try {                            root = balanceinsertion(root, x);                        } finally {                            unlockroot();                        }                    }                    break;                }            }            assert checkinvariants(root);            return null;        }
        /**         * removes the given node, that must be present before this         * call.  this is messier than typical red-black deletion code         * because we cannot swap the contents of an interior node         * with a leaf successor that is pinned by "next" pointers         * that are accessible independently of lock. so instead we         * swap the tree linkages.         *         * @return true if now too small, so should be untreeified         */        final boolean removetreenode(treenode<k,v> p) {            treenode<k,v> next = (treenode<k,v>)p.next;            treenode<k,v> pred = p.prev;  // unlink traversal pointers            treenode<k,v> r, rl;            if (pred == null)                first = next;            else                pred.next = next;            if (next != null)                next.prev = pred;            if (first == null) {                root = null;                return true;            }            if ((r = root) == null || r.right == null || // too small                (rl = r.left) == null || rl.left == null)                return true;            lockroot();            try {                treenode<k,v> replacement;                treenode<k,v> pl = p.left;                treenode<k,v> pr = p.right;                if (pl != null && pr != null) {                    treenode<k,v> s = pr, sl;                    while ((sl = s.left) != null) // find successor                        s = sl;                    boolean c = s.red; s.red = p.red; p.red = c; // swap colors                    treenode<k,v> sr = s.right;                    treenode<k,v> pp = p.parent;                    if (s == pr) { // p was s's direct parent                        p.parent = s;                        s.right = p;                    }                    else {                        treenode<k,v> sp = s.parent;                        if ((p.parent = sp) != null) {                            if (s == sp.left)                                sp.left = p;                            else                                sp.right = p;                        }                        if ((s.right = pr) != null)                            pr.parent = s;                    }                    p.left = null;                    if ((p.right = sr) != null)                        sr.parent = p;                    if ((s.left = pl) != null)                        pl.parent = s;                    if ((s.parent = pp) == null)                        r = s;                    else if (p == pp.left)                        pp.left = s;                    else                        pp.right = s;                    if (sr != null)                        replacement = sr;                    else                        replacement = p;                }                else if (pl != null)                    replacement = pl;                else if (pr != null)                    replacement = pr;                else                    replacement = p;                if (replacement != p) {                    treenode<k,v> pp = replacement.parent = p.parent;                    if (pp == null)                        r = replacement;                    else if (p == pp.left)                        pp.left = replacement;                    else                        pp.right = replacement;                    p.left = p.right = p.parent = null;                }
                root = (p.red) ? r : balancedeletion(r, replacement);
                if (p == replacement) {  // detach pointers                    treenode<k,v> pp;                    if ((pp = p.parent) != null) {                        if (p == pp.left)                            pp.left = null;                        else if (p == pp.right)                            pp.right = null;                        p.parent = null;                    }                }            } finally {                unlockroot();            }            assert checkinvariants(root);            return false;        }
        /* ------------------------------------------------------------ */        // red-black tree methods, all adapted from clr
        static <k,v> treenode<k,v> rotateleft(treenode<k,v> root,                                              treenode<k,v> p) {            treenode<k,v> r, pp, rl;            if (p != null && (r = p.right) != null) {                if ((rl = p.right = r.left) != null)                    rl.parent = p;                if ((pp = r.parent = p.parent) == null)                    (root = r).red = false;                else if (pp.left == p)                    pp.left = r;                else                    pp.right = r;                r.left = p;                p.parent = r;            }            return root;        }
        static <k,v> treenode<k,v> rotateright(treenode<k,v> root,                                               treenode<k,v> p) {            treenode<k,v> l, pp, lr;            if (p != null && (l = p.left) != null) {                if ((lr = p.left = l.right) != null)                    lr.parent = p;                if ((pp = l.parent = p.parent) == null)                    (root = l).red = false;                else if (pp.right == p)                    pp.right = l;                else                    pp.left = l;                l.right = p;                p.parent = l;            }            return root;        }
        static <k,v> treenode<k,v> balanceinsertion(treenode<k,v> root,                                                    treenode<k,v> x) {            x.red = true;            for (treenode<k,v> xp, xpp, xppl, xppr;;) {                if ((xp = x.parent) == null) {                    x.red = false;                    return x;                }                else if (!xp.red || (xpp = xp.parent) == null)                    return root;                if (xp == (xppl = xpp.left)) {                    if ((xppr = xpp.right) != null && xppr.red) {                        xppr.red = false;                        xp.red = false;                        xpp.red = true;                        x = xpp;                    }                    else {                        if (x == xp.right) {                            root = rotateleft(root, x = xp);                            xpp = (xp = x.parent) == null ? null : xp.parent;                        }                        if (xp != null) {                            xp.red = false;                            if (xpp != null) {                                xpp.red = true;                                root = rotateright(root, xpp);                            }                        }                    }                }                else {                    if (xppl != null && xppl.red) {                        xppl.red = false;                        xp.red = false;                        xpp.red = true;                        x = xpp;                    }                    else {                        if (x == xp.left) {                            root = rotateright(root, x = xp);                            xpp = (xp = x.parent) == null ? null : xp.parent;                        }                        if (xp != null) {                            xp.red = false;                            if (xpp != null) {                                xpp.red = true;                                root = rotateleft(root, xpp);                            }                        }                    }                }            }        }
        static <k,v> treenode<k,v> balancedeletion(treenode<k,v> root,                                                   treenode<k,v> x) {            for (treenode<k,v> xp, xpl, xpr;;)  {                if (x == null || x == root)                    return root;                else if ((xp = x.parent) == null) {                    x.red = false;                    return x;                }                else if (x.red) {                    x.red = false;                    return root;                }                else if ((xpl = xp.left) == x) {                    if ((xpr = xp.right) != null && xpr.red) {                        xpr.red = false;                        xp.red = true;                        root = rotateleft(root, xp);                        xpr = (xp = x.parent) == null ? null : xp.right;                    }                    if (xpr == null)                        x = xp;                    else {                        treenode<k,v> sl = xpr.left, sr = xpr.right;                        if ((sr == null || !sr.red) &&                            (sl == null || !sl.red)) {                            xpr.red = true;                            x = xp;                        }                        else {                            if (sr == null || !sr.red) {                                if (sl != null)                                    sl.red = false;                                xpr.red = true;                                root = rotateright(root, xpr);                                xpr = (xp = x.parent) == null ?                                    null : xp.right;                            }                            if (xpr != null) {                                xpr.red = (xp == null) ? false : xp.red;                                if ((sr = xpr.right) != null)                                    sr.red = false;                            }                            if (xp != null) {                                xp.red = false;                                root = rotateleft(root, xp);                            }                            x = root;                        }                    }                }                else { // symmetric                    if (xpl != null && xpl.red) {                        xpl.red = false;                        xp.red = true;                        root = rotateright(root, xp);                        xpl = (xp = x.parent) == null ? null : xp.left;                    }                    if (xpl == null)                        x = xp;                    else {                        treenode<k,v> sl = xpl.left, sr = xpl.right;                        if ((sl == null || !sl.red) &&                            (sr == null || !sr.red)) {                            xpl.red = true;                            x = xp;                        }                        else {                            if (sl == null || !sl.red) {                                if (sr != null)                                    sr.red = false;                                xpl.red = true;                                root = rotateleft(root, xpl);                                xpl = (xp = x.parent) == null ?                                    null : xp.left;                            }                            if (xpl != null) {                                xpl.red = (xp == null) ? false : xp.red;                                if ((sl = xpl.left) != null)                                    sl.red = false;                            }                            if (xp != null) {                                xp.red = false;                                root = rotateright(root, xp);                            }                            x = root;                        }                    }                }            }        }
        /**         * recursive invariant check         */        static <k,v> boolean checkinvariants(treenode<k,v> t) {            treenode<k,v> tp = t.parent, tl = t.left, tr = t.right,                tb = t.prev, tn = (treenode<k,v>)t.next;            if (tb != null && tb.next != t)                return false;            if (tn != null && tn.prev != t)                return false;            if (tp != null && t != tp.left && t != tp.right)                return false;            if (tl != null && (tl.parent != t || tl.hash > t.hash))                return false;            if (tr != null && (tr.parent != t || tr.hash < t.hash))                return false;            if (t.red && tl != null && tl.red && tr != null && tr.red)                return false;            if (tl != null && !checkinvariants(tl))                return false;            if (tr != null && !checkinvariants(tr))                return false;            return true;        }
        private static final sun.misc.unsafe u;        private static final long lockstate;        static {            try {                u = sun.misc.unsafe.getunsafe();                class<?> k = treebin.class;                lockstate = u.objectfieldoffset                    (k.getdeclaredfield("lockstate"));            } catch (exception e) {                throw new error(e);            }        }    }
    /* ----------------table traversal -------------- */
    /**     * records the table, its length, and current traversal index for a     * traverser that must process a region of a forwarded table before     * proceeding with current table.     */    static final class tablestack<k,v> {        int length;        int index;        node<k,v>[] tab;        tablestack<k,v> next;    }
    /**     * encapsulates traversal for methods such as containsvalue; also     * serves as a base class for other iterators and spliterators.     *     * method advance visits once each still-valid node that was     * reachable upon iterator construction. it might miss some that     * were added to a bin after the bin was visited, which is ok wrt     * consistency guarantees. maintaining this property in the face     * of possible ongoing resizes requires a fair amount of     * bookkeeping state that is difficult to optimize away amidst     * volatile accesses.  even so, traversal maintains reasonable     * throughput.     *     * normally, iteration proceeds bin-by-bin traversing lists.     * however, if the table has been resized, then all future steps     * must traverse both the bin at the current index as well as at     * (index + basesize); and so on for further resizings. to     * paranoically cope with potential sharing by users of iterators     * across threads, iteration terminates if a bounds checks fails     * for a table read.     */    static class traverser<k,v> {        node<k,v>[] tab;        // current table; updated if resized        node<k,v> next;         // the next entry to use        tablestack<k,v> stack, spare; // to save/restore on forwardingnodes        int index;              // index of bin to use next        int baseindex;          // current index of initial table        int baselimit;          // index bound for initial table        final int basesize;     // initial table size
        traverser(node<k,v>[] tab, int size, int index, int limit) {            this.tab = tab;            this.basesize = size;            this.baseindex = this.index = index;            this.baselimit = limit;            this.next = null;        }
        /**         * advances if possible, returning next valid node, or null if none.         */        final node<k,v> advance() {            node<k,v> e;            if ((e = next) != null)                e = e.next;            for (;;) {                node<k,v>[] t; int i, n;  // must use locals in checks                if (e != null)                    return next = e;                if (baseindex >= baselimit || (t = tab) == null ||                    (n = t.length) <= (i = index) || i < 0)                    return next = null;                if ((e = tabat(t, i)) != null && e.hash < 0) {                    if (e instanceof forwardingnode) {                        tab = ((forwardingnode<k,v>)e).nexttable;                        e = null;                        pushstate(t, i, n);                        continue;                    }                    else if (e instanceof treebin)                        e = ((treebin<k,v>)e).first;                    else                        e = null;                }                if (stack != null)                    recoverstate(n);                else if ((index = i + basesize) >= n)                    index = ++baseindex; // visit upper slots if present            }        }
        /**         * saves traversal state upon encountering a forwarding node.         */        private void pushstate(node<k,v>[] t, int i, int n) {            tablestack<k,v> s = spare;  // reuse if possible            if (s != null)                spare = s.next;            else                s = new tablestack<k,v>();            s.tab = t;            s.length = n;            s.index = i;            s.next = stack;            stack = s;        }
        /**         * possibly pops traversal state.         *         * @param n length of current table         */        private void recoverstate(int n) {            tablestack<k,v> s; int len;            while ((s = stack) != null && (index += (len = s.length)) >= n) {                n = len;                index = s.index;                tab = s.tab;                s.tab = null;                tablestack<k,v> next = s.next;                s.next = spare; // save for reuse                stack = next;                spare = s;            }            if (s == null && (index += basesize) >= n)                index = ++baseindex;        }    }
    /**     * base of key, value, and entry iterators. adds fields to     * traverser to support iterator.remove.     */    static class baseiterator<k,v> extends traverser<k,v> {        final concurrenthashmap<k,v> map;        node<k,v> lastreturned;        baseiterator(node<k,v>[] tab, int size, int index, int limit,                    concurrenthashmap<k,v> map) {            super(tab, size, index, limit);            this.map = map;            advance();        }
        public final boolean hasnext() { return next != null; }        public final boolean hasmoreelements() { return next != null; }
        public final void remove() {            node<k,v> p;            if ((p = lastreturned) == null)                throw new illegalstateexception();            lastreturned = null;            map.replacenode(p.key, null, null);        }    }
    static final class keyiterator<k,v> extends baseiterator<k,v>        implements iterator<k>, enumeration<k> {        keyiterator(node<k,v>[] tab, int index, int size, int limit,                    concurrenthashmap<k,v> map) {            super(tab, index, size, limit, map);        }
        public final k next() {            node<k,v> p;            if ((p = next) == null)                throw new nosuchelementexception();            k k = p.key;            lastreturned = p;            advance();            return k;        }
        public final k nextelement() { return next(); }    }
    static final class valueiterator<k,v> extends baseiterator<k,v>        implements iterator<v>, enumeration<v> {        valueiterator(node<k,v>[] tab, int index, int size, int limit,                      concurrenthashmap<k,v> map) {            super(tab, index, size, limit, map);        }
        public final v next() {            node<k,v> p;            if ((p = next) == null)                throw new nosuchelementexception();            v v = p.val;            lastreturned = p;            advance();            return v;        }
        public final v nextelement() { return next(); }    }
    static final class entryiterator<k,v> extends baseiterator<k,v>        implements iterator<map.entry<k,v>> {        entryiterator(node<k,v>[] tab, int index, int size, int limit,                      concurrenthashmap<k,v> map) {            super(tab, index, size, limit, map);        }
        public final map.entry<k,v> next() {            node<k,v> p;            if ((p = next) == null)                throw new nosuchelementexception();            k k = p.key;            v v = p.val;            lastreturned = p;            advance();            return new mapentry<k,v>(k, v, map);        }    }
    /**     * exported entry for entryiterator     */    static final class mapentry<k,v> implements map.entry<k,v> {        final k key; // non-null        v val;       // non-null        final concurrenthashmap<k,v> map;        mapentry(k key, v val, concurrenthashmap<k,v> map) {            this.key = key;            this.val = val;            this.map = map;        }        public k getkey()        { return key; }        public v getvalue()      { return val; }        public int hashcode()    { return key.hashcode() ^ val.hashcode(); }        public string tostring() { return key + "=" + val; }
        public boolean equals(object o) {            object k, v; map.entry<?,?> e;            return ((o instanceof map.entry) &&                    (k = (e = (map.entry<?,?>)o).getkey()) != null &&                    (v = e.getvalue()) != null &&                    (k == key || k.equals(key)) &&                    (v == val || v.equals(val)));        }
        /**         * sets our entry's value and writes through to the map. the         * value to return is somewhat arbitrary here. since we do not         * necessarily track asynchronous changes, the most recent         * "previous" value could be different from what we return (or         * could even have been removed, in which case the put will         * re-establish). we do not and cannot guarantee more.         */        public v setvalue(v value) {            if (value == null) throw new nullpointerexception();            v v = val;            val = value;            map.put(key, value);            return v;        }    }
    static final class keyspliterator<k,v> extends traverser<k,v>        implements spliterator<k> {        long est;               // size estimate        keyspliterator(node<k,v>[] tab, int size, int index, int limit,                       long est) {            super(tab, size, index, limit);            this.est = est;        }
        public spliterator<k> trysplit() {            int i, f, h;            return (h = ((i = baseindex) + (f = baselimit)) >>> 1) <= i ? null :                new keyspliterator<k,v>(tab, basesize, baselimit = h,                                        f, est >>>= 1);        }
        public void foreachremaining(consumer<? super k> action) {            if (action == null) throw new nullpointerexception();            for (node<k,v> p; (p = advance()) != null;)                action.accept(p.key);        }
        public boolean tryadvance(consumer<? super k> action) {            if (action == null) throw new nullpointerexception();            node<k,v> p;            if ((p = advance()) == null)                return false;            action.accept(p.key);            return true;        }
        public long estimatesize() { return est; }
        public int characteristics() {            return spliterator.distinct | spliterator.concurrent |                spliterator.nonnull;        }    }
    static final class valuespliterator<k,v> extends traverser<k,v>        implements spliterator<v> {        long est;               // size estimate        valuespliterator(node<k,v>[] tab, int size, int index, int limit,                         long est) {            super(tab, size, index, limit);            this.est = est;        }
        public spliterator<v> trysplit() {            int i, f, h;            return (h = ((i = baseindex) + (f = baselimit)) >>> 1) <= i ? null :                new valuespliterator<k,v>(tab, basesize, baselimit = h,                                          f, est >>>= 1);        }
        public void foreachremaining(consumer<? super v> action) {            if (action == null) throw new nullpointerexception();            for (node<k,v> p; (p = advance()) != null;)                action.accept(p.val);        }
        public boolean tryadvance(consumer<? super v> action) {            if (action == null) throw new nullpointerexception();            node<k,v> p;            if ((p = advance()) == null)                return false;            action.accept(p.val);            return true;        }
        public long estimatesize() { return est; }
        public int characteristics() {            return spliterator.concurrent | spliterator.nonnull;        }    }
    static final class entryspliterator<k,v> extends traverser<k,v>        implements spliterator<map.entry<k,v>> {        final concurrenthashmap<k,v> map; // to export mapentry        long est;               // size estimate        entryspliterator(node<k,v>[] tab, int size, int index, int limit,                         long est, concurrenthashmap<k,v> map) {            super(tab, size, index, limit);            this.map = map;            this.est = est;        }
        public spliterator<map.entry<k,v>> trysplit() {            int i, f, h;            return (h = ((i = baseindex) + (f = baselimit)) >>> 1) <= i ? null :                new entryspliterator<k,v>(tab, basesize, baselimit = h,                                          f, est >>>= 1, map);        }
        public void foreachremaining(consumer<? super map.entry<k,v>> action) {            if (action == null) throw new nullpointerexception();            for (node<k,v> p; (p = advance()) != null; )                action.accept(new mapentry<k,v>(p.key, p.val, map));        }
        public boolean tryadvance(consumer<? super map.entry<k,v>> action) {            if (action == null) throw new nullpointerexception();            node<k,v> p;            if ((p = advance()) == null)                return false;            action.accept(new mapentry<k,v>(p.key, p.val, map));            return true;        }
        public long estimatesize() { return est; }
        public int characteristics() {            return spliterator.distinct | spliterator.concurrent |                spliterator.nonnull;        }    }
    // parallel bulk operations
    /**     * computes initial batch value for bulk tasks. the returned value     * is approximately exp2 of the number of times (minus one) to     * split task by two before executing leaf action. this value is     * faster to compute and more convenient to use as a guide to     * splitting than is the depth, since it is used while dividing by     * two anyway.     */    final int batchfor(long b) {        long n;        if (b == long.max_value || (n = sumcount()) <= 1l || n < b)            return 0;        int sp = forkjoinpool.getcommonpoolparallelism() << 2; // slack of 4        return (b <= 0l || (n /= b) >= sp) ? sp : (int)n;    }
    /**     * performs the given action for each (key, value).     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param action the action     * @since 1.8     */    public void foreach(long parallelismthreshold,                        biconsumer<? super k,? super v> action) {        if (action == null) throw new nullpointerexception();        new foreachmappingtask<k,v>            (null, batchfor(parallelismthreshold), 0, 0, table,             action).invoke();    }
    /**     * performs the given action for each non-null transformation     * of each (key, value).     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param transformer a function returning the transformation     * for an element, or null if there is no transformation (in     * which case the action is not applied)     * @param action the action     * @param <u> the return type of the transformer     * @since 1.8     */    public <u> void foreach(long parallelismthreshold,                            bifunction<? super k, ? super v, ? extends u> transformer,                            consumer<? super u> action) {        if (transformer == null || action == null)            throw new nullpointerexception();        new foreachtransformedmappingtask<k,v,u>            (null, batchfor(parallelismthreshold), 0, 0, table,             transformer, action).invoke();    }
    /**     * returns a non-null result from applying the given search     * function on each (key, value), or null if none.  upon     * success, further element processing is suppressed and the     * results of any other parallel invocations of the search     * function are ignored.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param searchfunction a function returning a non-null     * result on success, else null     * @param <u> the return type of the search function     * @return a non-null result from applying the given search     * function on each (key, value), or null if none     * @since 1.8     */    public <u> u search(long parallelismthreshold,                        bifunction<? super k, ? super v, ? extends u> searchfunction) {        if (searchfunction == null) throw new nullpointerexception();        return new searchmappingstask<k,v,u>            (null, batchfor(parallelismthreshold), 0, 0, table,             searchfunction, new atomicreference<u>()).invoke();    }
    /**     * returns the result of accumulating the given transformation     * of all (key, value) pairs using the given reducer to     * combine values, or null if none.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param transformer a function returning the transformation     * for an element, or null if there is no transformation (in     * which case it is not combined)     * @param reducer a commutative associative combining function     * @param <u> the return type of the transformer     * @return the result of accumulating the given transformation     * of all (key, value) pairs     * @since 1.8     */    public <u> u reduce(long parallelismthreshold,                        bifunction<? super k, ? super v, ? extends u> transformer,                        bifunction<? super u, ? super u, ? extends u> reducer) {        if (transformer == null || reducer == null)            throw new nullpointerexception();        return new mapreducemappingstask<k,v,u>            (null, batchfor(parallelismthreshold), 0, 0, table,             null, transformer, reducer).invoke();    }
    /**     * returns the result of accumulating the given transformation     * of all (key, value) pairs using the given reducer to     * combine values, and the given basis as an identity value.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param transformer a function returning the transformation     * for an element     * @param basis the identity (initial default value) for the reduction     * @param reducer a commutative associative combining function     * @return the result of accumulating the given transformation     * of all (key, value) pairs     * @since 1.8     */    public double reducetodouble(long parallelismthreshold,                                 todoublebifunction<? super k, ? super v> transformer,                                 double basis,                                 doublebinaryoperator reducer) {        if (transformer == null || reducer == null)            throw new nullpointerexception();        return new mapreducemappingstodoubletask<k,v>            (null, batchfor(parallelismthreshold), 0, 0, table,             null, transformer, basis, reducer).invoke();    }
    /**     * returns the result of accumulating the given transformation     * of all (key, value) pairs using the given reducer to     * combine values, and the given basis as an identity value.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param transformer a function returning the transformation     * for an element     * @param basis the identity (initial default value) for the reduction     * @param reducer a commutative associative combining function     * @return the result of accumulating the given transformation     * of all (key, value) pairs     * @since 1.8     */    public long reducetolong(long parallelismthreshold,                             tolongbifunction<? super k, ? super v> transformer,                             long basis,                             longbinaryoperator reducer) {        if (transformer == null || reducer == null)            throw new nullpointerexception();        return new mapreducemappingstolongtask<k,v>            (null, batchfor(parallelismthreshold), 0, 0, table,             null, transformer, basis, reducer).invoke();    }
    /**     * returns the result of accumulating the given transformation     * of all (key, value) pairs using the given reducer to     * combine values, and the given basis as an identity value.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param transformer a function returning the transformation     * for an element     * @param basis the identity (initial default value) for the reduction     * @param reducer a commutative associative combining function     * @return the result of accumulating the given transformation     * of all (key, value) pairs     * @since 1.8     */    public int reducetoint(long parallelismthreshold,                           tointbifunction<? super k, ? super v> transformer,                           int basis,                           intbinaryoperator reducer) {        if (transformer == null || reducer == null)            throw new nullpointerexception();        return new mapreducemappingstointtask<k,v>            (null, batchfor(parallelismthreshold), 0, 0, table,             null, transformer, basis, reducer).invoke();    }
    /**     * performs the given action for each key.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param action the action     * @since 1.8     */    public void foreachkey(long parallelismthreshold,                           consumer<? super k> action) {        if (action == null) throw new nullpointerexception();        new foreachkeytask<k,v>            (null, batchfor(parallelismthreshold), 0, 0, table,             action).invoke();    }
    /**     * performs the given action for each non-null transformation     * of each key.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param transformer a function returning the transformation     * for an element, or null if there is no transformation (in     * which case the action is not applied)     * @param action the action     * @param <u> the return type of the transformer     * @since 1.8     */    public <u> void foreachkey(long parallelismthreshold,                               function<? super k, ? extends u> transformer,                               consumer<? super u> action) {        if (transformer == null || action == null)            throw new nullpointerexception();        new foreachtransformedkeytask<k,v,u>            (null, batchfor(parallelismthreshold), 0, 0, table,             transformer, action).invoke();    }
    /**     * returns a non-null result from applying the given search     * function on each key, or null if none. upon success,     * further element processing is suppressed and the results of     * any other parallel invocations of the search function are     * ignored.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param searchfunction a function returning a non-null     * result on success, else null     * @param <u> the return type of the search function     * @return a non-null result from applying the given search     * function on each key, or null if none     * @since 1.8     */    public <u> u searchkeys(long parallelismthreshold,                            function<? super k, ? extends u> searchfunction) {        if (searchfunction == null) throw new nullpointerexception();        return new searchkeystask<k,v,u>            (null, batchfor(parallelismthreshold), 0, 0, table,             searchfunction, new atomicreference<u>()).invoke();    }
    /**     * returns the result of accumulating all keys using the given     * reducer to combine values, or null if none.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param reducer a commutative associative combining function     * @return the result of accumulating all keys using the given     * reducer to combine values, or null if none     * @since 1.8     */    public k reducekeys(long parallelismthreshold,                        bifunction<? super k, ? super k, ? extends k> reducer) {        if (reducer == null) throw new nullpointerexception();        return new reducekeystask<k,v>            (null, batchfor(parallelismthreshold), 0, 0, table,             null, reducer).invoke();    }
    /**     * returns the result of accumulating the given transformation     * of all keys using the given reducer to combine values, or     * null if none.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param transformer a function returning the transformation     * for an element, or null if there is no transformation (in     * which case it is not combined)     * @param reducer a commutative associative combining function     * @param <u> the return type of the transformer     * @return the result of accumulating the given transformation     * of all keys     * @since 1.8     */    public <u> u reducekeys(long parallelismthreshold,                            function<? super k, ? extends u> transformer,         bifunction<? super u, ? super u, ? extends u> reducer) {        if (transformer == null || reducer == null)            throw new nullpointerexception();        return new mapreducekeystask<k,v,u>            (null, batchfor(parallelismthreshold), 0, 0, table,             null, transformer, reducer).invoke();    }
    /**     * returns the result of accumulating the given transformation     * of all keys using the given reducer to combine values, and     * the given basis as an identity value.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param transformer a function returning the transformation     * for an element     * @param basis the identity (initial default value) for the reduction     * @param reducer a commutative associative combining function     * @return the result of accumulating the given transformation     * of all keys     * @since 1.8     */    public double reducekeystodouble(long parallelismthreshold,                                     todoublefunction<? super k> transformer,                                     double basis,                                     doublebinaryoperator reducer) {        if (transformer == null || reducer == null)            throw new nullpointerexception();        return new mapreducekeystodoubletask<k,v>            (null, batchfor(parallelismthreshold), 0, 0, table,             null, transformer, basis, reducer).invoke();    }
    /**     * returns the result of accumulating the given transformation     * of all keys using the given reducer to combine values, and     * the given basis as an identity value.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param transformer a function returning the transformation     * for an element     * @param basis the identity (initial default value) for the reduction     * @param reducer a commutative associative combining function     * @return the result of accumulating the given transformation     * of all keys     * @since 1.8     */    public long reducekeystolong(long parallelismthreshold,                                 tolongfunction<? super k> transformer,                                 long basis,                                 longbinaryoperator reducer) {        if (transformer == null || reducer == null)            throw new nullpointerexception();        return new mapreducekeystolongtask<k,v>            (null, batchfor(parallelismthreshold), 0, 0, table,             null, transformer, basis, reducer).invoke();    }
    /**     * returns the result of accumulating the given transformation     * of all keys using the given reducer to combine values, and     * the given basis as an identity value.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param transformer a function returning the transformation     * for an element     * @param basis the identity (initial default value) for the reduction     * @param reducer a commutative associative combining function     * @return the result of accumulating the given transformation     * of all keys     * @since 1.8     */    public int reducekeystoint(long parallelismthreshold,                               tointfunction<? super k> transformer,                               int basis,                               intbinaryoperator reducer) {        if (transformer == null || reducer == null)            throw new nullpointerexception();        return new mapreducekeystointtask<k,v>            (null, batchfor(parallelismthreshold), 0, 0, table,             null, transformer, basis, reducer).invoke();    }
    /**     * performs the given action for each value.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param action the action     * @since 1.8     */    public void foreachvalue(long parallelismthreshold,                             consumer<? super v> action) {        if (action == null)            throw new nullpointerexception();        new foreachvaluetask<k,v>            (null, batchfor(parallelismthreshold), 0, 0, table,             action).invoke();    }
    /**     * performs the given action for each non-null transformation     * of each value.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param transformer a function returning the transformation     * for an element, or null if there is no transformation (in     * which case the action is not applied)     * @param action the action     * @param <u> the return type of the transformer     * @since 1.8     */    public <u> void foreachvalue(long parallelismthreshold,                                 function<? super v, ? extends u> transformer,                                 consumer<? super u> action) {        if (transformer == null || action == null)            throw new nullpointerexception();        new foreachtransformedvaluetask<k,v,u>            (null, batchfor(parallelismthreshold), 0, 0, table,             transformer, action).invoke();    }
    /**     * returns a non-null result from applying the given search     * function on each value, or null if none.  upon success,     * further element processing is suppressed and the results of     * any other parallel invocations of the search function are     * ignored.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param searchfunction a function returning a non-null     * result on success, else null     * @param <u> the return type of the search function     * @return a non-null result from applying the given search     * function on each value, or null if none     * @since 1.8     */    public <u> u searchvalues(long parallelismthreshold,                              function<? super v, ? extends u> searchfunction) {        if (searchfunction == null) throw new nullpointerexception();        return new searchvaluestask<k,v,u>            (null, batchfor(parallelismthreshold), 0, 0, table,             searchfunction, new atomicreference<u>()).invoke();    }
    /**     * returns the result of accumulating all values using the     * given reducer to combine values, or null if none.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param reducer a commutative associative combining function     * @return the result of accumulating all values     * @since 1.8     */    public v reducevalues(long parallelismthreshold,                          bifunction<? super v, ? super v, ? extends v> reducer) {        if (reducer == null) throw new nullpointerexception();        return new reducevaluestask<k,v>            (null, batchfor(parallelismthreshold), 0, 0, table,             null, reducer).invoke();    }
    /**     * returns the result of accumulating the given transformation     * of all values using the given reducer to combine values, or     * null if none.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param transformer a function returning the transformation     * for an element, or null if there is no transformation (in     * which case it is not combined)     * @param reducer a commutative associative combining function     * @param <u> the return type of the transformer     * @return the result of accumulating the given transformation     * of all values     * @since 1.8     */    public <u> u reducevalues(long parallelismthreshold,                              function<? super v, ? extends u> transformer,                              bifunction<? super u, ? super u, ? extends u> reducer) {        if (transformer == null || reducer == null)            throw new nullpointerexception();        return new mapreducevaluestask<k,v,u>            (null, batchfor(parallelismthreshold), 0, 0, table,             null, transformer, reducer).invoke();    }
    /**     * returns the result of accumulating the given transformation     * of all values using the given reducer to combine values,     * and the given basis as an identity value.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param transformer a function returning the transformation     * for an element     * @param basis the identity (initial default value) for the reduction     * @param reducer a commutative associative combining function     * @return the result of accumulating the given transformation     * of all values     * @since 1.8     */    public double reducevaluestodouble(long parallelismthreshold,                                       todoublefunction<? super v> transformer,                                       double basis,                                       doublebinaryoperator reducer) {        if (transformer == null || reducer == null)            throw new nullpointerexception();        return new mapreducevaluestodoubletask<k,v>            (null, batchfor(parallelismthreshold), 0, 0, table,             null, transformer, basis, reducer).invoke();    }
    /**     * returns the result of accumulating the given transformation     * of all values using the given reducer to combine values,     * and the given basis as an identity value.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param transformer a function returning the transformation     * for an element     * @param basis the identity (initial default value) for the reduction     * @param reducer a commutative associative combining function     * @return the result of accumulating the given transformation     * of all values     * @since 1.8     */    public long reducevaluestolong(long parallelismthreshold,                                   tolongfunction<? super v> transformer,                                   long basis,                                   longbinaryoperator reducer) {        if (transformer == null || reducer == null)            throw new nullpointerexception();        return new mapreducevaluestolongtask<k,v>            (null, batchfor(parallelismthreshold), 0, 0, table,             null, transformer, basis, reducer).invoke();    }
    /**     * returns the result of accumulating the given transformation     * of all values using the given reducer to combine values,     * and the given basis as an identity value.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param transformer a function returning the transformation     * for an element     * @param basis the identity (initial default value) for the reduction     * @param reducer a commutative associative combining function     * @return the result of accumulating the given transformation     * of all values     * @since 1.8     */    public int reducevaluestoint(long parallelismthreshold,                                 tointfunction<? super v> transformer,                                 int basis,                                 intbinaryoperator reducer) {        if (transformer == null || reducer == null)            throw new nullpointerexception();        return new mapreducevaluestointtask<k,v>            (null, batchfor(parallelismthreshold), 0, 0, table,             null, transformer, basis, reducer).invoke();    }
    /**     * performs the given action for each entry.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param action the action     * @since 1.8     */    public void foreachentry(long parallelismthreshold,                             consumer<? super map.entry<k,v>> action) {        if (action == null) throw new nullpointerexception();        new foreachentrytask<k,v>(null, batchfor(parallelismthreshold), 0, 0, table,                                  action).invoke();    }
    /**     * performs the given action for each non-null transformation     * of each entry.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param transformer a function returning the transformation     * for an element, or null if there is no transformation (in     * which case the action is not applied)     * @param action the action     * @param <u> the return type of the transformer     * @since 1.8     */    public <u> void foreachentry(long parallelismthreshold,                                 function<map.entry<k,v>, ? extends u> transformer,                                 consumer<? super u> action) {        if (transformer == null || action == null)            throw new nullpointerexception();        new foreachtransformedentrytask<k,v,u>            (null, batchfor(parallelismthreshold), 0, 0, table,             transformer, action).invoke();    }
    /**     * returns a non-null result from applying the given search     * function on each entry, or null if none.  upon success,     * further element processing is suppressed and the results of     * any other parallel invocations of the search function are     * ignored.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param searchfunction a function returning a non-null     * result on success, else null     * @param <u> the return type of the search function     * @return a non-null result from applying the given search     * function on each entry, or null if none     * @since 1.8     */    public <u> u searchentries(long parallelismthreshold,                               function<map.entry<k,v>, ? extends u> searchfunction) {        if (searchfunction == null) throw new nullpointerexception();        return new searchentriestask<k,v,u>            (null, batchfor(parallelismthreshold), 0, 0, table,             searchfunction, new atomicreference<u>()).invoke();    }
    /**     * returns the result of accumulating all entries using the     * given reducer to combine values, or null if none.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param reducer a commutative associative combining function     * @return the result of accumulating all entries     * @since 1.8     */    public map.entry<k,v> reduceentries(long parallelismthreshold,                                        bifunction<map.entry<k,v>, map.entry<k,v>, ? extends map.entry<k,v>> reducer) {        if (reducer == null) throw new nullpointerexception();        return new reduceentriestask<k,v>            (null, batchfor(parallelismthreshold), 0, 0, table,             null, reducer).invoke();    }
    /**     * returns the result of accumulating the given transformation     * of all entries using the given reducer to combine values,     * or null if none.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param transformer a function returning the transformation     * for an element, or null if there is no transformation (in     * which case it is not combined)     * @param reducer a commutative associative combining function     * @param <u> the return type of the transformer     * @return the result of accumulating the given transformation     * of all entries     * @since 1.8     */    public <u> u reduceentries(long parallelismthreshold,                               function<map.entry<k,v>, ? extends u> transformer,                               bifunction<? super u, ? super u, ? extends u> reducer) {        if (transformer == null || reducer == null)            throw new nullpointerexception();        return new mapreduceentriestask<k,v,u>            (null, batchfor(parallelismthreshold), 0, 0, table,             null, transformer, reducer).invoke();    }
    /**     * returns the result of accumulating the given transformation     * of all entries using the given reducer to combine values,     * and the given basis as an identity value.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param transformer a function returning the transformation     * for an element     * @param basis the identity (initial default value) for the reduction     * @param reducer a commutative associative combining function     * @return the result of accumulating the given transformation     * of all entries     * @since 1.8     */    public double reduceentriestodouble(long parallelismthreshold,                                        todoublefunction<map.entry<k,v>> transformer,                                        double basis,                                        doublebinaryoperator reducer) {        if (transformer == null || reducer == null)            throw new nullpointerexception();        return new mapreduceentriestodoubletask<k,v>            (null, batchfor(parallelismthreshold), 0, 0, table,             null, transformer, basis, reducer).invoke();    }
    /**     * returns the result of accumulating the given transformation     * of all entries using the given reducer to combine values,     * and the given basis as an identity value.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param transformer a function returning the transformation     * for an element     * @param basis the identity (initial default value) for the reduction     * @param reducer a commutative associative combining function     * @return the result of accumulating the given transformation     * of all entries     * @since 1.8     */    public long reduceentriestolong(long parallelismthreshold,                                    tolongfunction<map.entry<k,v>> transformer,                                    long basis,                                    longbinaryoperator reducer) {        if (transformer == null || reducer == null)            throw new nullpointerexception();        return new mapreduceentriestolongtask<k,v>            (null, batchfor(parallelismthreshold), 0, 0, table,             null, transformer, basis, reducer).invoke();    }
    /**     * returns the result of accumulating the given transformation     * of all entries using the given reducer to combine values,     * and the given basis as an identity value.     *     * @param parallelismthreshold the (estimated) number of elements     * needed for this operation to be executed in parallel     * @param transformer a function returning the transformation     * for an element     * @param basis the identity (initial default value) for the reduction     * @param reducer a commutative associative combining function     * @return the result of accumulating the given transformation     * of all entries     * @since 1.8     */    public int reduceentriestoint(long parallelismthreshold,                                  tointfunction<map.entry<k,v>> transformer,                                  int basis,                                  intbinaryoperator reducer) {        if (transformer == null || reducer == null)            throw new nullpointerexception();        return new mapreduceentriestointtask<k,v>            (null, batchfor(parallelismthreshold), 0, 0, table,             null, transformer, basis, reducer).invoke();    }

    /* ----------------views -------------- */
    /**     * base class for views.     */    abstract static class collectionview<k,v,e>        implements collection<e>, java.io.serializable {        private static final long serialversionuid = 7249069246763182397l;        final concurrenthashmap<k,v> map;        collectionview(concurrenthashmap<k,v> map)  { this.map = map; }
        /**         * returns the map backing this view.         *         * @return the map backing this view         */        public concurrenthashmap<k,v> getmap() { return map; }
        /**         * removes all of the elements from this view, by removing all         * the mappings from the map backing this view.         */        public final void clear()      { map.clear(); }        public final int size()        { return map.size(); }        public final boolean isempty() { return map.isempty(); }
        // implementations below rely on concrete classes supplying these        // abstract methods        /**         * returns an iterator over the elements in this collection.         *         * <p>the returned iterator is         * <a href="package-summary.html#weakly"><i>weakly consistent</i></a>.         *         * @return an iterator over the elements in this collection         */        public abstract iterator<e> iterator();        public abstract boolean contains(object o);        public abstract boolean remove(object o);
        private static final string oomemsg = "required array size too large";
        public final object[] toarray() {            long sz = map.mappingcount();            if (sz > max_array_size)                throw new outofmemoryerror(oomemsg);            int n = (int)sz;            object[] r = new object[n];            int i = 0;            for (e e : this) {                if (i == n) {                    if (n >= max_array_size)                        throw new outofmemoryerror(oomemsg);                    if (n >= max_array_size - (max_array_size >>> 1) - 1)                        n = max_array_size;                    else                        n += (n >>> 1) + 1;                    r = arrays.copyof(r, n);                }                r[i++] = e;            }            return (i == n) ? r : arrays.copyof(r, i);        }
        @suppresswarnings("unchecked")        public final <t> t[] toarray(t[] a) {            long sz = map.mappingcount();            if (sz > max_array_size)                throw new outofmemoryerror(oomemsg);            int m = (int)sz;            t[] r = (a.length >= m) ? a :                (t[])java.lang.reflect.array                .newinstance(a.getclass().getcomponenttype(), m);            int n = r.length;            int i = 0;            for (e e : this) {                if (i == n) {                    if (n >= max_array_size)                        throw new outofmemoryerror(oomemsg);                    if (n >= max_array_size - (max_array_size >>> 1) - 1)                        n = max_array_size;                    else                        n += (n >>> 1) + 1;                    r = arrays.copyof(r, n);                }                r[i++] = (t)e;            }            if (a == r && i < n) {                r[i] = null; // null-terminate                return r;            }            return (i == n) ? r : arrays.copyof(r, i);        }
        /**         * returns a string representation of this collection.         * the string representation consists of the string representations         * of the collection's elements in the order they are returned by         * its iterator, enclosed in square brackets ({@code "[]"}).         * adjacent elements are separated by the characters {@code ", "}         * (comma and space).  elements are converted to strings as by         * {@link string#valueof(object)}.         *         * @return a string representation of this collection         */        public final string tostring() {            stringbuilder sb = new stringbuilder();            sb.append('[');            iterator<e> it = iterator();            if (it.hasnext()) {                for (;;) {                    object e = it.next();                    sb.append(e == this ? "(this collection)" : e);                    if (!it.hasnext())                        break;                    sb.append(',').append(' ');                }            }            return sb.append(']').tostring();        }
        public final boolean containsall(collection<?> c) {            if (c != this) {                for (object e : c) {                    if (e == null || !contains(e))                        return false;                }            }            return true;        }
        public final boolean removeall(collection<?> c) {            if (c == null) throw new nullpointerexception();            boolean modified = false;            for (iterator<e> it = iterator(); it.hasnext();) {                if (c.contains(it.next())) {                    it.remove();                    modified = true;                }            }            return modified;        }
        public final boolean retainall(collection<?> c) {            if (c == null) throw new nullpointerexception();            boolean modified = false;            for (iterator<e> it = iterator(); it.hasnext();) {                if (!c.contains(it.next())) {                    it.remove();                    modified = true;                }            }            return modified;        }
    }
    /**     * a view of a concurrenthashmap as a {@link set} of keys, in     * which additions may optionally be enabled by mapping to a     * common value.  this class cannot be directly instantiated.     * see {@link #keyset() keyset()},     * {@link #keyset(object) keyset(v)},     * {@link #newkeyset() newkeyset()},     * {@link #newkeyset(int) newkeyset(int)}.     *     * @since 1.8     */    public static class keysetview<k,v> extends collectionview<k,v,k>        implements set<k>, java.io.serializable {        private static final long serialversionuid = 7249069246763182397l;        private final v value;        keysetview(concurrenthashmap<k,v> map, v value) {  // non-public            super(map);            this.value = value;        }
        /**         * returns the default mapped value for additions,         * or {@code null} if additions are not supported.         *         * @return the default mapped value for additions, or {@code null}         * if not supported         */        public v getmappedvalue() { return value; }
        /**         * {@inheritdoc}         * @throws nullpointerexception if the specified key is null         */        public boolean contains(object o) { return map.containskey(o); }
        /**         * removes the key from this map view, by removing the key (and its         * corresponding value) from the backing map.  this method does         * nothing if the key is not in the map.         *         * @param  o the key to be removed from the backing map         * @return {@code true} if the backing map contained the specified key         * @throws nullpointerexception if the specified key is null         */        public boolean remove(object o) { return map.remove(o) != null; }
        /**         * @return an iterator over the keys of the backing map         */        public iterator<k> iterator() {            node<k,v>[] t;            concurrenthashmap<k,v> m = map;            int f = (t = m.table) == null ? 0 : t.length;            return new keyiterator<k,v>(t, f, 0, f, m);        }
        /**         * adds the specified key to this set view by mapping the key to         * the default mapped value in the backing map, if defined.         *         * @param e key to be added         * @return {@code true} if this set changed as a result of the call         * @throws nullpointerexception if the specified key is null         * @throws unsupportedoperationexception if no default mapped value         * for additions was provided         */        public boolean add(k e) {            v v;            if ((v = value) == null)                throw new unsupportedoperationexception();            return map.putval(e, v, true) == null;        }
        /**         * adds all of the elements in the specified collection to this set,         * as if by calling {@link #add} on each one.         *         * @param c the elements to be inserted into this set         * @return {@code true} if this set changed as a result of the call         * @throws nullpointerexception if the collection or any of its         * elements are {@code null}         * @throws unsupportedoperationexception if no default mapped value         * for additions was provided         */        public boolean addall(collection<? extends k> c) {            boolean added = false;            v v;            if ((v = value) == null)                throw new unsupportedoperationexception();            for (k e : c) {                if (map.putval(e, v, true) == null)                    added = true;            }            return added;        }
        public int hashcode() {            int h = 0;            for (k e : this)                h += e.hashcode();            return h;        }
        public boolean equals(object o) {            set<?> c;            return ((o instanceof set) &&                    ((c = (set<?>)o) == this ||                     (containsall(c) && c.containsall(this))));        }
        public spliterator<k> spliterator() {            node<k,v>[] t;            concurrenthashmap<k,v> m = map;            long n = m.sumcount();            int f = (t = m.table) == null ? 0 : t.length;            return new keyspliterator<k,v>(t, f, 0, f, n < 0l ? 0l : n);        }
        public void foreach(consumer<? super k> action) {            if (action == null) throw new nullpointerexception();            node<k,v>[] t;            if ((t = map.table) != null) {                traverser<k,v> it = new traverser<k,v>(t, t.length, 0, t.length);                for (node<k,v> p; (p = it.advance()) != null; )                    action.accept(p.key);            }        }    }
    /**     * a view of a concurrenthashmap as a {@link collection} of     * values, in which additions are disabled. this class cannot be     * directly instantiated. see {@link #values()}.     */    static final class valuesview<k,v> extends collectionview<k,v,v>        implements collection<v>, java.io.serializable {        private static final long serialversionuid = 2249069246763182397l;        valuesview(concurrenthashmap<k,v> map) { super(map); }        public final boolean contains(object o) {            return map.containsvalue(o);        }
        public final boolean remove(object o) {            if (o != null) {                for (iterator<v> it = iterator(); it.hasnext();) {                    if (o.equals(it.next())) {                        it.remove();                        return true;                    }                }            }            return false;        }
        public final iterator<v> iterator() {            concurrenthashmap<k,v> m = map;            node<k,v>[] t;            int f = (t = m.table) == null ? 0 : t.length;            return new valueiterator<k,v>(t, f, 0, f, m);        }
        public final boolean add(v e) {            throw new unsupportedoperationexception();        }        public final boolean addall(collection<? extends v> c) {            throw new unsupportedoperationexception();        }
        public spliterator<v> spliterator() {            node<k,v>[] t;            concurrenthashmap<k,v> m = map;            long n = m.sumcount();            int f = (t = m.table) == null ? 0 : t.length;            return new valuespliterator<k,v>(t, f, 0, f, n < 0l ? 0l : n);        }
        public void foreach(consumer<? super v> action) {            if (action == null) throw new nullpointerexception();            node<k,v>[] t;            if ((t = map.table) != null) {                traverser<k,v> it = new traverser<k,v>(t, t.length, 0, t.length);                for (node<k,v> p; (p = it.advance()) != null; )                    action.accept(p.val);            }        }    }
    /**     * a view of a concurrenthashmap as a {@link set} of (key, value)     * entries.  this class cannot be directly instantiated. see     * {@link #entryset()}.     */    static final class entrysetview<k,v> extends collectionview<k,v,map.entry<k,v>>        implements set<map.entry<k,v>>, java.io.serializable {        private static final long serialversionuid = 2249069246763182397l;        entrysetview(concurrenthashmap<k,v> map) { super(map); }
        public boolean contains(object o) {            object k, v, r; map.entry<?,?> e;            return ((o instanceof map.entry) &&                    (k = (e = (map.entry<?,?>)o).getkey()) != null &&                    (r = map.get(k)) != null &&                    (v = e.getvalue()) != null &&                    (v == r || v.equals(r)));        }
        public boolean remove(object o) {            object k, v; map.entry<?,?> e;            return ((o instanceof map.entry) &&                    (k = (e = (map.entry<?,?>)o).getkey()) != null &&                    (v = e.getvalue()) != null &&                    map.remove(k, v));        }
        /**         * @return an iterator over the entries of the backing map         */        public iterator<map.entry<k,v>> iterator() {            concurrenthashmap<k,v> m = map;            node<k,v>[] t;            int f = (t = m.table) == null ? 0 : t.length;            return new entryiterator<k,v>(t, f, 0, f, m);        }
        public boolean add(entry<k,v> e) {            return map.putval(e.getkey(), e.getvalue(), false) == null;        }
        public boolean addall(collection<? extends entry<k,v>> c) {            boolean added = false;            for (entry<k,v> e : c) {                if (add(e))                    added = true;            }            return added;        }
        public final int hashcode() {            int h = 0;            node<k,v>[] t;            if ((t = map.table) != null) {                traverser<k,v> it = new traverser<k,v>(t, t.length, 0, t.length);                for (node<k,v> p; (p = it.advance()) != null; ) {                    h += p.hashcode();                }            }            return h;        }
        public final boolean equals(object o) {            set<?> c;            return ((o instanceof set) &&                    ((c = (set<?>)o) == this ||                     (containsall(c) && c.containsall(this))));        }
        public spliterator<map.entry<k,v>> spliterator() {            node<k,v>[] t;            concurrenthashmap<k,v> m = map;            long n = m.sumcount();            int f = (t = m.table) == null ? 0 : t.length;            return new entryspliterator<k,v>(t, f, 0, f, n < 0l ? 0l : n, m);        }
        public void foreach(consumer<? super map.entry<k,v>> action) {            if (action == null) throw new nullpointerexception();            node<k,v>[] t;            if ((t = map.table) != null) {                traverser<k,v> it = new traverser<k,v>(t, t.length, 0, t.length);                for (node<k,v> p; (p = it.advance()) != null; )                    action.accept(new mapentry<k,v>(p.key, p.val, map));            }        }
    }
    // -------------------------------------------------------
    /**     * base class for bulk tasks. repeats some fields and code from     * class traverser, because we need to subclass countedcompleter.     */    @suppresswarnings("serial")    abstract static class bulktask<k,v,r> extends countedcompleter<r> {        node<k,v>[] tab;        // same as traverser        node<k,v> next;        tablestack<k,v> stack, spare;        int index;        int baseindex;        int baselimit;        final int basesize;        int batch;              // split control
        bulktask(bulktask<k,v,?> par, int b, int i, int f, node<k,v>[] t) {            super(par);            this.batch = b;            this.index = this.baseindex = i;            if ((this.tab = t) == null)                this.basesize = this.baselimit = 0;            else if (par == null)                this.basesize = this.baselimit = t.length;            else {                this.baselimit = f;                this.basesize = par.basesize;            }        }
        /**         * same as traverser version         */        final node<k,v> advance() {            node<k,v> e;            if ((e = next) != null)                e = e.next;            for (;;) {                node<k,v>[] t; int i, n;                if (e != null)                    return next = e;                if (baseindex >= baselimit || (t = tab) == null ||                    (n = t.length) <= (i = index) || i < 0)                    return next = null;                if ((e = tabat(t, i)) != null && e.hash < 0) {                    if (e instanceof forwardingnode) {                        tab = ((forwardingnode<k,v>)e).nexttable;                        e = null;                        pushstate(t, i, n);                        continue;                    }                    else if (e instanceof treebin)                        e = ((treebin<k,v>)e).first;                    else                        e = null;                }                if (stack != null)                    recoverstate(n);                else if ((index = i + basesize) >= n)                    index = ++baseindex;            }        }
        private void pushstate(node<k,v>[] t, int i, int n) {            tablestack<k,v> s = spare;            if (s != null)                spare = s.next;            else                s = new tablestack<k,v>();            s.tab = t;            s.length = n;            s.index = i;            s.next = stack;            stack = s;        }
        private void recoverstate(int n) {            tablestack<k,v> s; int len;            while ((s = stack) != null && (index += (len = s.length)) >= n) {                n = len;                index = s.index;                tab = s.tab;                s.tab = null;                tablestack<k,v> next = s.next;                s.next = spare; // save for reuse                stack = next;                spare = s;            }            if (s == null && (index += basesize) >= n)                index = ++baseindex;        }    }
    /*     * task classes. coded in a regular but ugly format/style to     * simplify checks that each variant differs in the right way from     * others. the null screenings exist because compilers cannot tell     * that we've already null-checked task arguments, so we force     * simplest hoisted bypass to help avoid convoluted traps.     */    @suppresswarnings("serial")    static final class foreachkeytask<k,v>        extends bulktask<k,v,void> {        final consumer<? super k> action;        foreachkeytask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             consumer<? super k> action) {            super(p, b, i, f, t);            this.action = action;        }        public final void compute() {            final consumer<? super k> action;            if ((action = this.action) != null) {                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    new foreachkeytask<k,v>                        (this, batch >>>= 1, baselimit = h, f, tab,                         action).fork();                }                for (node<k,v> p; (p = advance()) != null;)                    action.accept(p.key);                propagatecompletion();            }        }    }
    @suppresswarnings("serial")    static final class foreachvaluetask<k,v>        extends bulktask<k,v,void> {        final consumer<? super v> action;        foreachvaluetask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             consumer<? super v> action) {            super(p, b, i, f, t);            this.action = action;        }        public final void compute() {            final consumer<? super v> action;            if ((action = this.action) != null) {                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    new foreachvaluetask<k,v>                        (this, batch >>>= 1, baselimit = h, f, tab,                         action).fork();                }                for (node<k,v> p; (p = advance()) != null;)                    action.accept(p.val);                propagatecompletion();            }        }    }
    @suppresswarnings("serial")    static final class foreachentrytask<k,v>        extends bulktask<k,v,void> {        final consumer<? super entry<k,v>> action;        foreachentrytask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             consumer<? super entry<k,v>> action) {            super(p, b, i, f, t);            this.action = action;        }        public final void compute() {            final consumer<? super entry<k,v>> action;            if ((action = this.action) != null) {                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    new foreachentrytask<k,v>                        (this, batch >>>= 1, baselimit = h, f, tab,                         action).fork();                }                for (node<k,v> p; (p = advance()) != null; )                    action.accept(p);                propagatecompletion();            }        }    }
    @suppresswarnings("serial")    static final class foreachmappingtask<k,v>        extends bulktask<k,v,void> {        final biconsumer<? super k, ? super v> action;        foreachmappingtask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             biconsumer<? super k,? super v> action) {            super(p, b, i, f, t);            this.action = action;        }        public final void compute() {            final biconsumer<? super k, ? super v> action;            if ((action = this.action) != null) {                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    new foreachmappingtask<k,v>                        (this, batch >>>= 1, baselimit = h, f, tab,                         action).fork();                }                for (node<k,v> p; (p = advance()) != null; )                    action.accept(p.key, p.val);                propagatecompletion();            }        }    }
    @suppresswarnings("serial")    static final class foreachtransformedkeytask<k,v,u>        extends bulktask<k,v,void> {        final function<? super k, ? extends u> transformer;        final consumer<? super u> action;        foreachtransformedkeytask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             function<? super k, ? extends u> transformer, consumer<? super u> action) {            super(p, b, i, f, t);            this.transformer = transformer; this.action = action;        }        public final void compute() {            final function<? super k, ? extends u> transformer;            final consumer<? super u> action;            if ((transformer = this.transformer) != null &&                (action = this.action) != null) {                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    new foreachtransformedkeytask<k,v,u>                        (this, batch >>>= 1, baselimit = h, f, tab,                         transformer, action).fork();                }                for (node<k,v> p; (p = advance()) != null; ) {                    u u;                    if ((u = transformer.apply(p.key)) != null)                        action.accept(u);                }                propagatecompletion();            }        }    }
    @suppresswarnings("serial")    static final class foreachtransformedvaluetask<k,v,u>        extends bulktask<k,v,void> {        final function<? super v, ? extends u> transformer;        final consumer<? super u> action;        foreachtransformedvaluetask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             function<? super v, ? extends u> transformer, consumer<? super u> action) {            super(p, b, i, f, t);            this.transformer = transformer; this.action = action;        }        public final void compute() {            final function<? super v, ? extends u> transformer;            final consumer<? super u> action;            if ((transformer = this.transformer) != null &&                (action = this.action) != null) {                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    new foreachtransformedvaluetask<k,v,u>                        (this, batch >>>= 1, baselimit = h, f, tab,                         transformer, action).fork();                }                for (node<k,v> p; (p = advance()) != null; ) {                    u u;                    if ((u = transformer.apply(p.val)) != null)                        action.accept(u);                }                propagatecompletion();            }        }    }
    @suppresswarnings("serial")    static final class foreachtransformedentrytask<k,v,u>        extends bulktask<k,v,void> {        final function<map.entry<k,v>, ? extends u> transformer;        final consumer<? super u> action;        foreachtransformedentrytask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             function<map.entry<k,v>, ? extends u> transformer, consumer<? super u> action) {            super(p, b, i, f, t);            this.transformer = transformer; this.action = action;        }        public final void compute() {            final function<map.entry<k,v>, ? extends u> transformer;            final consumer<? super u> action;            if ((transformer = this.transformer) != null &&                (action = this.action) != null) {                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    new foreachtransformedentrytask<k,v,u>                        (this, batch >>>= 1, baselimit = h, f, tab,                         transformer, action).fork();                }                for (node<k,v> p; (p = advance()) != null; ) {                    u u;                    if ((u = transformer.apply(p)) != null)                        action.accept(u);                }                propagatecompletion();            }        }    }
    @suppresswarnings("serial")    static final class foreachtransformedmappingtask<k,v,u>        extends bulktask<k,v,void> {        final bifunction<? super k, ? super v, ? extends u> transformer;        final consumer<? super u> action;        foreachtransformedmappingtask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             bifunction<? super k, ? super v, ? extends u> transformer,             consumer<? super u> action) {            super(p, b, i, f, t);            this.transformer = transformer; this.action = action;        }        public final void compute() {            final bifunction<? super k, ? super v, ? extends u> transformer;            final consumer<? super u> action;            if ((transformer = this.transformer) != null &&                (action = this.action) != null) {                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    new foreachtransformedmappingtask<k,v,u>                        (this, batch >>>= 1, baselimit = h, f, tab,                         transformer, action).fork();                }                for (node<k,v> p; (p = advance()) != null; ) {                    u u;                    if ((u = transformer.apply(p.key, p.val)) != null)                        action.accept(u);                }                propagatecompletion();            }        }    }
    @suppresswarnings("serial")    static final class searchkeystask<k,v,u>        extends bulktask<k,v,u> {        final function<? super k, ? extends u> searchfunction;        final atomicreference<u> result;        searchkeystask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             function<? super k, ? extends u> searchfunction,             atomicreference<u> result) {            super(p, b, i, f, t);            this.searchfunction = searchfunction; this.result = result;        }        public final u getrawresult() { return result.get(); }        public final void compute() {            final function<? super k, ? extends u> searchfunction;            final atomicreference<u> result;            if ((searchfunction = this.searchfunction) != null &&                (result = this.result) != null) {                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    if (result.get() != null)                        return;                    addtopendingcount(1);                    new searchkeystask<k,v,u>                        (this, batch >>>= 1, baselimit = h, f, tab,                         searchfunction, result).fork();                }                while (result.get() == null) {                    u u;                    node<k,v> p;                    if ((p = advance()) == null) {                        propagatecompletion();                        break;                    }                    if ((u = searchfunction.apply(p.key)) != null) {                        if (result.compareandset(null, u))                            quietlycompleteroot();                        break;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class searchvaluestask<k,v,u>        extends bulktask<k,v,u> {        final function<? super v, ? extends u> searchfunction;        final atomicreference<u> result;        searchvaluestask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             function<? super v, ? extends u> searchfunction,             atomicreference<u> result) {            super(p, b, i, f, t);            this.searchfunction = searchfunction; this.result = result;        }        public final u getrawresult() { return result.get(); }        public final void compute() {            final function<? super v, ? extends u> searchfunction;            final atomicreference<u> result;            if ((searchfunction = this.searchfunction) != null &&                (result = this.result) != null) {                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    if (result.get() != null)                        return;                    addtopendingcount(1);                    new searchvaluestask<k,v,u>                        (this, batch >>>= 1, baselimit = h, f, tab,                         searchfunction, result).fork();                }                while (result.get() == null) {                    u u;                    node<k,v> p;                    if ((p = advance()) == null) {                        propagatecompletion();                        break;                    }                    if ((u = searchfunction.apply(p.val)) != null) {                        if (result.compareandset(null, u))                            quietlycompleteroot();                        break;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class searchentriestask<k,v,u>        extends bulktask<k,v,u> {        final function<entry<k,v>, ? extends u> searchfunction;        final atomicreference<u> result;        searchentriestask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             function<entry<k,v>, ? extends u> searchfunction,             atomicreference<u> result) {            super(p, b, i, f, t);            this.searchfunction = searchfunction; this.result = result;        }        public final u getrawresult() { return result.get(); }        public final void compute() {            final function<entry<k,v>, ? extends u> searchfunction;            final atomicreference<u> result;            if ((searchfunction = this.searchfunction) != null &&                (result = this.result) != null) {                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    if (result.get() != null)                        return;                    addtopendingcount(1);                    new searchentriestask<k,v,u>                        (this, batch >>>= 1, baselimit = h, f, tab,                         searchfunction, result).fork();                }                while (result.get() == null) {                    u u;                    node<k,v> p;                    if ((p = advance()) == null) {                        propagatecompletion();                        break;                    }                    if ((u = searchfunction.apply(p)) != null) {                        if (result.compareandset(null, u))                            quietlycompleteroot();                        return;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class searchmappingstask<k,v,u>        extends bulktask<k,v,u> {        final bifunction<? super k, ? super v, ? extends u> searchfunction;        final atomicreference<u> result;        searchmappingstask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             bifunction<? super k, ? super v, ? extends u> searchfunction,             atomicreference<u> result) {            super(p, b, i, f, t);            this.searchfunction = searchfunction; this.result = result;        }        public final u getrawresult() { return result.get(); }        public final void compute() {            final bifunction<? super k, ? super v, ? extends u> searchfunction;            final atomicreference<u> result;            if ((searchfunction = this.searchfunction) != null &&                (result = this.result) != null) {                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    if (result.get() != null)                        return;                    addtopendingcount(1);                    new searchmappingstask<k,v,u>                        (this, batch >>>= 1, baselimit = h, f, tab,                         searchfunction, result).fork();                }                while (result.get() == null) {                    u u;                    node<k,v> p;                    if ((p = advance()) == null) {                        propagatecompletion();                        break;                    }                    if ((u = searchfunction.apply(p.key, p.val)) != null) {                        if (result.compareandset(null, u))                            quietlycompleteroot();                        break;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class reducekeystask<k,v>        extends bulktask<k,v,k> {        final bifunction<? super k, ? super k, ? extends k> reducer;        k result;        reducekeystask<k,v> rights, nextright;        reducekeystask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             reducekeystask<k,v> nextright,             bifunction<? super k, ? super k, ? extends k> reducer) {            super(p, b, i, f, t); this.nextright = nextright;            this.reducer = reducer;        }        public final k getrawresult() { return result; }        public final void compute() {            final bifunction<? super k, ? super k, ? extends k> reducer;            if ((reducer = this.reducer) != null) {                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    (rights = new reducekeystask<k,v>                     (this, batch >>>= 1, baselimit = h, f, tab,                      rights, reducer)).fork();                }                k r = null;                for (node<k,v> p; (p = advance()) != null; ) {                    k u = p.key;                    r = (r == null) ? u : u == null ? r : reducer.apply(r, u);                }                result = r;                countedcompleter<?> c;                for (c = firstcomplete(); c != null; c = c.nextcomplete()) {                    @suppresswarnings("unchecked")                    reducekeystask<k,v>                        t = (reducekeystask<k,v>)c,                        s = t.rights;                    while (s != null) {                        k tr, sr;                        if ((sr = s.result) != null)                            t.result = (((tr = t.result) == null) ? sr :                                        reducer.apply(tr, sr));                        s = t.rights = s.nextright;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class reducevaluestask<k,v>        extends bulktask<k,v,v> {        final bifunction<? super v, ? super v, ? extends v> reducer;        v result;        reducevaluestask<k,v> rights, nextright;        reducevaluestask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             reducevaluestask<k,v> nextright,             bifunction<? super v, ? super v, ? extends v> reducer) {            super(p, b, i, f, t); this.nextright = nextright;            this.reducer = reducer;        }        public final v getrawresult() { return result; }        public final void compute() {            final bifunction<? super v, ? super v, ? extends v> reducer;            if ((reducer = this.reducer) != null) {                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    (rights = new reducevaluestask<k,v>                     (this, batch >>>= 1, baselimit = h, f, tab,                      rights, reducer)).fork();                }                v r = null;                for (node<k,v> p; (p = advance()) != null; ) {                    v v = p.val;                    r = (r == null) ? v : reducer.apply(r, v);                }                result = r;                countedcompleter<?> c;                for (c = firstcomplete(); c != null; c = c.nextcomplete()) {                    @suppresswarnings("unchecked")                    reducevaluestask<k,v>                        t = (reducevaluestask<k,v>)c,                        s = t.rights;                    while (s != null) {                        v tr, sr;                        if ((sr = s.result) != null)                            t.result = (((tr = t.result) == null) ? sr :                                        reducer.apply(tr, sr));                        s = t.rights = s.nextright;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class reduceentriestask<k,v>        extends bulktask<k,v,map.entry<k,v>> {        final bifunction<map.entry<k,v>, map.entry<k,v>, ? extends map.entry<k,v>> reducer;        map.entry<k,v> result;        reduceentriestask<k,v> rights, nextright;        reduceentriestask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             reduceentriestask<k,v> nextright,             bifunction<entry<k,v>, map.entry<k,v>, ? extends map.entry<k,v>> reducer) {            super(p, b, i, f, t); this.nextright = nextright;            this.reducer = reducer;        }        public final map.entry<k,v> getrawresult() { return result; }        public final void compute() {            final bifunction<map.entry<k,v>, map.entry<k,v>, ? extends map.entry<k,v>> reducer;            if ((reducer = this.reducer) != null) {                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    (rights = new reduceentriestask<k,v>                     (this, batch >>>= 1, baselimit = h, f, tab,                      rights, reducer)).fork();                }                map.entry<k,v> r = null;                for (node<k,v> p; (p = advance()) != null; )                    r = (r == null) ? p : reducer.apply(r, p);                result = r;                countedcompleter<?> c;                for (c = firstcomplete(); c != null; c = c.nextcomplete()) {                    @suppresswarnings("unchecked")                    reduceentriestask<k,v>                        t = (reduceentriestask<k,v>)c,                        s = t.rights;                    while (s != null) {                        map.entry<k,v> tr, sr;                        if ((sr = s.result) != null)                            t.result = (((tr = t.result) == null) ? sr :                                        reducer.apply(tr, sr));                        s = t.rights = s.nextright;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class mapreducekeystask<k,v,u>        extends bulktask<k,v,u> {        final function<? super k, ? extends u> transformer;        final bifunction<? super u, ? super u, ? extends u> reducer;        u result;        mapreducekeystask<k,v,u> rights, nextright;        mapreducekeystask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             mapreducekeystask<k,v,u> nextright,             function<? super k, ? extends u> transformer,             bifunction<? super u, ? super u, ? extends u> reducer) {            super(p, b, i, f, t); this.nextright = nextright;            this.transformer = transformer;            this.reducer = reducer;        }        public final u getrawresult() { return result; }        public final void compute() {            final function<? super k, ? extends u> transformer;            final bifunction<? super u, ? super u, ? extends u> reducer;            if ((transformer = this.transformer) != null &&                (reducer = this.reducer) != null) {                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    (rights = new mapreducekeystask<k,v,u>                     (this, batch >>>= 1, baselimit = h, f, tab,                      rights, transformer, reducer)).fork();                }                u r = null;                for (node<k,v> p; (p = advance()) != null; ) {                    u u;                    if ((u = transformer.apply(p.key)) != null)                        r = (r == null) ? u : reducer.apply(r, u);                }                result = r;                countedcompleter<?> c;                for (c = firstcomplete(); c != null; c = c.nextcomplete()) {                    @suppresswarnings("unchecked")                    mapreducekeystask<k,v,u>                        t = (mapreducekeystask<k,v,u>)c,                        s = t.rights;                    while (s != null) {                        u tr, sr;                        if ((sr = s.result) != null)                            t.result = (((tr = t.result) == null) ? sr :                                        reducer.apply(tr, sr));                        s = t.rights = s.nextright;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class mapreducevaluestask<k,v,u>        extends bulktask<k,v,u> {        final function<? super v, ? extends u> transformer;        final bifunction<? super u, ? super u, ? extends u> reducer;        u result;        mapreducevaluestask<k,v,u> rights, nextright;        mapreducevaluestask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             mapreducevaluestask<k,v,u> nextright,             function<? super v, ? extends u> transformer,             bifunction<? super u, ? super u, ? extends u> reducer) {            super(p, b, i, f, t); this.nextright = nextright;            this.transformer = transformer;            this.reducer = reducer;        }        public final u getrawresult() { return result; }        public final void compute() {            final function<? super v, ? extends u> transformer;            final bifunction<? super u, ? super u, ? extends u> reducer;            if ((transformer = this.transformer) != null &&                (reducer = this.reducer) != null) {                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    (rights = new mapreducevaluestask<k,v,u>                     (this, batch >>>= 1, baselimit = h, f, tab,                      rights, transformer, reducer)).fork();                }                u r = null;                for (node<k,v> p; (p = advance()) != null; ) {                    u u;                    if ((u = transformer.apply(p.val)) != null)                        r = (r == null) ? u : reducer.apply(r, u);                }                result = r;                countedcompleter<?> c;                for (c = firstcomplete(); c != null; c = c.nextcomplete()) {                    @suppresswarnings("unchecked")                    mapreducevaluestask<k,v,u>                        t = (mapreducevaluestask<k,v,u>)c,                        s = t.rights;                    while (s != null) {                        u tr, sr;                        if ((sr = s.result) != null)                            t.result = (((tr = t.result) == null) ? sr :                                        reducer.apply(tr, sr));                        s = t.rights = s.nextright;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class mapreduceentriestask<k,v,u>        extends bulktask<k,v,u> {        final function<map.entry<k,v>, ? extends u> transformer;        final bifunction<? super u, ? super u, ? extends u> reducer;        u result;        mapreduceentriestask<k,v,u> rights, nextright;        mapreduceentriestask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             mapreduceentriestask<k,v,u> nextright,             function<map.entry<k,v>, ? extends u> transformer,             bifunction<? super u, ? super u, ? extends u> reducer) {            super(p, b, i, f, t); this.nextright = nextright;            this.transformer = transformer;            this.reducer = reducer;        }        public final u getrawresult() { return result; }        public final void compute() {            final function<map.entry<k,v>, ? extends u> transformer;            final bifunction<? super u, ? super u, ? extends u> reducer;            if ((transformer = this.transformer) != null &&                (reducer = this.reducer) != null) {                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    (rights = new mapreduceentriestask<k,v,u>                     (this, batch >>>= 1, baselimit = h, f, tab,                      rights, transformer, reducer)).fork();                }                u r = null;                for (node<k,v> p; (p = advance()) != null; ) {                    u u;                    if ((u = transformer.apply(p)) != null)                        r = (r == null) ? u : reducer.apply(r, u);                }                result = r;                countedcompleter<?> c;                for (c = firstcomplete(); c != null; c = c.nextcomplete()) {                    @suppresswarnings("unchecked")                    mapreduceentriestask<k,v,u>                        t = (mapreduceentriestask<k,v,u>)c,                        s = t.rights;                    while (s != null) {                        u tr, sr;                        if ((sr = s.result) != null)                            t.result = (((tr = t.result) == null) ? sr :                                        reducer.apply(tr, sr));                        s = t.rights = s.nextright;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class mapreducemappingstask<k,v,u>        extends bulktask<k,v,u> {        final bifunction<? super k, ? super v, ? extends u> transformer;        final bifunction<? super u, ? super u, ? extends u> reducer;        u result;        mapreducemappingstask<k,v,u> rights, nextright;        mapreducemappingstask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             mapreducemappingstask<k,v,u> nextright,             bifunction<? super k, ? super v, ? extends u> transformer,             bifunction<? super u, ? super u, ? extends u> reducer) {            super(p, b, i, f, t); this.nextright = nextright;            this.transformer = transformer;            this.reducer = reducer;        }        public final u getrawresult() { return result; }        public final void compute() {            final bifunction<? super k, ? super v, ? extends u> transformer;            final bifunction<? super u, ? super u, ? extends u> reducer;            if ((transformer = this.transformer) != null &&                (reducer = this.reducer) != null) {                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    (rights = new mapreducemappingstask<k,v,u>                     (this, batch >>>= 1, baselimit = h, f, tab,                      rights, transformer, reducer)).fork();                }                u r = null;                for (node<k,v> p; (p = advance()) != null; ) {                    u u;                    if ((u = transformer.apply(p.key, p.val)) != null)                        r = (r == null) ? u : reducer.apply(r, u);                }                result = r;                countedcompleter<?> c;                for (c = firstcomplete(); c != null; c = c.nextcomplete()) {                    @suppresswarnings("unchecked")                    mapreducemappingstask<k,v,u>                        t = (mapreducemappingstask<k,v,u>)c,                        s = t.rights;                    while (s != null) {                        u tr, sr;                        if ((sr = s.result) != null)                            t.result = (((tr = t.result) == null) ? sr :                                        reducer.apply(tr, sr));                        s = t.rights = s.nextright;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class mapreducekeystodoubletask<k,v>        extends bulktask<k,v,double> {        final todoublefunction<? super k> transformer;        final doublebinaryoperator reducer;        final double basis;        double result;        mapreducekeystodoubletask<k,v> rights, nextright;        mapreducekeystodoubletask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             mapreducekeystodoubletask<k,v> nextright,             todoublefunction<? super k> transformer,             double basis,             doublebinaryoperator reducer) {            super(p, b, i, f, t); this.nextright = nextright;            this.transformer = transformer;            this.basis = basis; this.reducer = reducer;        }        public final double getrawresult() { return result; }        public final void compute() {            final todoublefunction<? super k> transformer;            final doublebinaryoperator reducer;            if ((transformer = this.transformer) != null &&                (reducer = this.reducer) != null) {                double r = this.basis;                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    (rights = new mapreducekeystodoubletask<k,v>                     (this, batch >>>= 1, baselimit = h, f, tab,                      rights, transformer, r, reducer)).fork();                }                for (node<k,v> p; (p = advance()) != null; )                    r = reducer.applyasdouble(r, transformer.applyasdouble(p.key));                result = r;                countedcompleter<?> c;                for (c = firstcomplete(); c != null; c = c.nextcomplete()) {                    @suppresswarnings("unchecked")                    mapreducekeystodoubletask<k,v>                        t = (mapreducekeystodoubletask<k,v>)c,                        s = t.rights;                    while (s != null) {                        t.result = reducer.applyasdouble(t.result, s.result);                        s = t.rights = s.nextright;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class mapreducevaluestodoubletask<k,v>        extends bulktask<k,v,double> {        final todoublefunction<? super v> transformer;        final doublebinaryoperator reducer;        final double basis;        double result;        mapreducevaluestodoubletask<k,v> rights, nextright;        mapreducevaluestodoubletask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             mapreducevaluestodoubletask<k,v> nextright,             todoublefunction<? super v> transformer,             double basis,             doublebinaryoperator reducer) {            super(p, b, i, f, t); this.nextright = nextright;            this.transformer = transformer;            this.basis = basis; this.reducer = reducer;        }        public final double getrawresult() { return result; }        public final void compute() {            final todoublefunction<? super v> transformer;            final doublebinaryoperator reducer;            if ((transformer = this.transformer) != null &&                (reducer = this.reducer) != null) {                double r = this.basis;                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    (rights = new mapreducevaluestodoubletask<k,v>                     (this, batch >>>= 1, baselimit = h, f, tab,                      rights, transformer, r, reducer)).fork();                }                for (node<k,v> p; (p = advance()) != null; )                    r = reducer.applyasdouble(r, transformer.applyasdouble(p.val));                result = r;                countedcompleter<?> c;                for (c = firstcomplete(); c != null; c = c.nextcomplete()) {                    @suppresswarnings("unchecked")                    mapreducevaluestodoubletask<k,v>                        t = (mapreducevaluestodoubletask<k,v>)c,                        s = t.rights;                    while (s != null) {                        t.result = reducer.applyasdouble(t.result, s.result);                        s = t.rights = s.nextright;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class mapreduceentriestodoubletask<k,v>        extends bulktask<k,v,double> {        final todoublefunction<map.entry<k,v>> transformer;        final doublebinaryoperator reducer;        final double basis;        double result;        mapreduceentriestodoubletask<k,v> rights, nextright;        mapreduceentriestodoubletask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             mapreduceentriestodoubletask<k,v> nextright,             todoublefunction<map.entry<k,v>> transformer,             double basis,             doublebinaryoperator reducer) {            super(p, b, i, f, t); this.nextright = nextright;            this.transformer = transformer;            this.basis = basis; this.reducer = reducer;        }        public final double getrawresult() { return result; }        public final void compute() {            final todoublefunction<map.entry<k,v>> transformer;            final doublebinaryoperator reducer;            if ((transformer = this.transformer) != null &&                (reducer = this.reducer) != null) {                double r = this.basis;                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    (rights = new mapreduceentriestodoubletask<k,v>                     (this, batch >>>= 1, baselimit = h, f, tab,                      rights, transformer, r, reducer)).fork();                }                for (node<k,v> p; (p = advance()) != null; )                    r = reducer.applyasdouble(r, transformer.applyasdouble(p));                result = r;                countedcompleter<?> c;                for (c = firstcomplete(); c != null; c = c.nextcomplete()) {                    @suppresswarnings("unchecked")                    mapreduceentriestodoubletask<k,v>                        t = (mapreduceentriestodoubletask<k,v>)c,                        s = t.rights;                    while (s != null) {                        t.result = reducer.applyasdouble(t.result, s.result);                        s = t.rights = s.nextright;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class mapreducemappingstodoubletask<k,v>        extends bulktask<k,v,double> {        final todoublebifunction<? super k, ? super v> transformer;        final doublebinaryoperator reducer;        final double basis;        double result;        mapreducemappingstodoubletask<k,v> rights, nextright;        mapreducemappingstodoubletask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             mapreducemappingstodoubletask<k,v> nextright,             todoublebifunction<? super k, ? super v> transformer,             double basis,             doublebinaryoperator reducer) {            super(p, b, i, f, t); this.nextright = nextright;            this.transformer = transformer;            this.basis = basis; this.reducer = reducer;        }        public final double getrawresult() { return result; }        public final void compute() {            final todoublebifunction<? super k, ? super v> transformer;            final doublebinaryoperator reducer;            if ((transformer = this.transformer) != null &&                (reducer = this.reducer) != null) {                double r = this.basis;                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    (rights = new mapreducemappingstodoubletask<k,v>                     (this, batch >>>= 1, baselimit = h, f, tab,                      rights, transformer, r, reducer)).fork();                }                for (node<k,v> p; (p = advance()) != null; )                    r = reducer.applyasdouble(r, transformer.applyasdouble(p.key, p.val));                result = r;                countedcompleter<?> c;                for (c = firstcomplete(); c != null; c = c.nextcomplete()) {                    @suppresswarnings("unchecked")                    mapreducemappingstodoubletask<k,v>                        t = (mapreducemappingstodoubletask<k,v>)c,                        s = t.rights;                    while (s != null) {                        t.result = reducer.applyasdouble(t.result, s.result);                        s = t.rights = s.nextright;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class mapreducekeystolongtask<k,v>        extends bulktask<k,v,long> {        final tolongfunction<? super k> transformer;        final longbinaryoperator reducer;        final long basis;        long result;        mapreducekeystolongtask<k,v> rights, nextright;        mapreducekeystolongtask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             mapreducekeystolongtask<k,v> nextright,             tolongfunction<? super k> transformer,             long basis,             longbinaryoperator reducer) {            super(p, b, i, f, t); this.nextright = nextright;            this.transformer = transformer;            this.basis = basis; this.reducer = reducer;        }        public final long getrawresult() { return result; }        public final void compute() {            final tolongfunction<? super k> transformer;            final longbinaryoperator reducer;            if ((transformer = this.transformer) != null &&                (reducer = this.reducer) != null) {                long r = this.basis;                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    (rights = new mapreducekeystolongtask<k,v>                     (this, batch >>>= 1, baselimit = h, f, tab,                      rights, transformer, r, reducer)).fork();                }                for (node<k,v> p; (p = advance()) != null; )                    r = reducer.applyaslong(r, transformer.applyaslong(p.key));                result = r;                countedcompleter<?> c;                for (c = firstcomplete(); c != null; c = c.nextcomplete()) {                    @suppresswarnings("unchecked")                    mapreducekeystolongtask<k,v>                        t = (mapreducekeystolongtask<k,v>)c,                        s = t.rights;                    while (s != null) {                        t.result = reducer.applyaslong(t.result, s.result);                        s = t.rights = s.nextright;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class mapreducevaluestolongtask<k,v>        extends bulktask<k,v,long> {        final tolongfunction<? super v> transformer;        final longbinaryoperator reducer;        final long basis;        long result;        mapreducevaluestolongtask<k,v> rights, nextright;        mapreducevaluestolongtask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             mapreducevaluestolongtask<k,v> nextright,             tolongfunction<? super v> transformer,             long basis,             longbinaryoperator reducer) {            super(p, b, i, f, t); this.nextright = nextright;            this.transformer = transformer;            this.basis = basis; this.reducer = reducer;        }        public final long getrawresult() { return result; }        public final void compute() {            final tolongfunction<? super v> transformer;            final longbinaryoperator reducer;            if ((transformer = this.transformer) != null &&                (reducer = this.reducer) != null) {                long r = this.basis;                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    (rights = new mapreducevaluestolongtask<k,v>                     (this, batch >>>= 1, baselimit = h, f, tab,                      rights, transformer, r, reducer)).fork();                }                for (node<k,v> p; (p = advance()) != null; )                    r = reducer.applyaslong(r, transformer.applyaslong(p.val));                result = r;                countedcompleter<?> c;                for (c = firstcomplete(); c != null; c = c.nextcomplete()) {                    @suppresswarnings("unchecked")                    mapreducevaluestolongtask<k,v>                        t = (mapreducevaluestolongtask<k,v>)c,                        s = t.rights;                    while (s != null) {                        t.result = reducer.applyaslong(t.result, s.result);                        s = t.rights = s.nextright;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class mapreduceentriestolongtask<k,v>        extends bulktask<k,v,long> {        final tolongfunction<map.entry<k,v>> transformer;        final longbinaryoperator reducer;        final long basis;        long result;        mapreduceentriestolongtask<k,v> rights, nextright;        mapreduceentriestolongtask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             mapreduceentriestolongtask<k,v> nextright,             tolongfunction<map.entry<k,v>> transformer,             long basis,             longbinaryoperator reducer) {            super(p, b, i, f, t); this.nextright = nextright;            this.transformer = transformer;            this.basis = basis; this.reducer = reducer;        }        public final long getrawresult() { return result; }        public final void compute() {            final tolongfunction<map.entry<k,v>> transformer;            final longbinaryoperator reducer;            if ((transformer = this.transformer) != null &&                (reducer = this.reducer) != null) {                long r = this.basis;                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    (rights = new mapreduceentriestolongtask<k,v>                     (this, batch >>>= 1, baselimit = h, f, tab,                      rights, transformer, r, reducer)).fork();                }                for (node<k,v> p; (p = advance()) != null; )                    r = reducer.applyaslong(r, transformer.applyaslong(p));                result = r;                countedcompleter<?> c;                for (c = firstcomplete(); c != null; c = c.nextcomplete()) {                    @suppresswarnings("unchecked")                    mapreduceentriestolongtask<k,v>                        t = (mapreduceentriestolongtask<k,v>)c,                        s = t.rights;                    while (s != null) {                        t.result = reducer.applyaslong(t.result, s.result);                        s = t.rights = s.nextright;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class mapreducemappingstolongtask<k,v>        extends bulktask<k,v,long> {        final tolongbifunction<? super k, ? super v> transformer;        final longbinaryoperator reducer;        final long basis;        long result;        mapreducemappingstolongtask<k,v> rights, nextright;        mapreducemappingstolongtask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             mapreducemappingstolongtask<k,v> nextright,             tolongbifunction<? super k, ? super v> transformer,             long basis,             longbinaryoperator reducer) {            super(p, b, i, f, t); this.nextright = nextright;            this.transformer = transformer;            this.basis = basis; this.reducer = reducer;        }        public final long getrawresult() { return result; }        public final void compute() {            final tolongbifunction<? super k, ? super v> transformer;            final longbinaryoperator reducer;            if ((transformer = this.transformer) != null &&                (reducer = this.reducer) != null) {                long r = this.basis;                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    (rights = new mapreducemappingstolongtask<k,v>                     (this, batch >>>= 1, baselimit = h, f, tab,                      rights, transformer, r, reducer)).fork();                }                for (node<k,v> p; (p = advance()) != null; )                    r = reducer.applyaslong(r, transformer.applyaslong(p.key, p.val));                result = r;                countedcompleter<?> c;                for (c = firstcomplete(); c != null; c = c.nextcomplete()) {                    @suppresswarnings("unchecked")                    mapreducemappingstolongtask<k,v>                        t = (mapreducemappingstolongtask<k,v>)c,                        s = t.rights;                    while (s != null) {                        t.result = reducer.applyaslong(t.result, s.result);                        s = t.rights = s.nextright;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class mapreducekeystointtask<k,v>        extends bulktask<k,v,integer> {        final tointfunction<? super k> transformer;        final intbinaryoperator reducer;        final int basis;        int result;        mapreducekeystointtask<k,v> rights, nextright;        mapreducekeystointtask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             mapreducekeystointtask<k,v> nextright,             tointfunction<? super k> transformer,             int basis,             intbinaryoperator reducer) {            super(p, b, i, f, t); this.nextright = nextright;            this.transformer = transformer;            this.basis = basis; this.reducer = reducer;        }        public final integer getrawresult() { return result; }        public final void compute() {            final tointfunction<? super k> transformer;            final intbinaryoperator reducer;            if ((transformer = this.transformer) != null &&                (reducer = this.reducer) != null) {                int r = this.basis;                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    (rights = new mapreducekeystointtask<k,v>                     (this, batch >>>= 1, baselimit = h, f, tab,                      rights, transformer, r, reducer)).fork();                }                for (node<k,v> p; (p = advance()) != null; )                    r = reducer.applyasint(r, transformer.applyasint(p.key));                result = r;                countedcompleter<?> c;                for (c = firstcomplete(); c != null; c = c.nextcomplete()) {                    @suppresswarnings("unchecked")                    mapreducekeystointtask<k,v>                        t = (mapreducekeystointtask<k,v>)c,                        s = t.rights;                    while (s != null) {                        t.result = reducer.applyasint(t.result, s.result);                        s = t.rights = s.nextright;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class mapreducevaluestointtask<k,v>        extends bulktask<k,v,integer> {        final tointfunction<? super v> transformer;        final intbinaryoperator reducer;        final int basis;        int result;        mapreducevaluestointtask<k,v> rights, nextright;        mapreducevaluestointtask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             mapreducevaluestointtask<k,v> nextright,             tointfunction<? super v> transformer,             int basis,             intbinaryoperator reducer) {            super(p, b, i, f, t); this.nextright = nextright;            this.transformer = transformer;            this.basis = basis; this.reducer = reducer;        }        public final integer getrawresult() { return result; }        public final void compute() {            final tointfunction<? super v> transformer;            final intbinaryoperator reducer;            if ((transformer = this.transformer) != null &&                (reducer = this.reducer) != null) {                int r = this.basis;                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    (rights = new mapreducevaluestointtask<k,v>                     (this, batch >>>= 1, baselimit = h, f, tab,                      rights, transformer, r, reducer)).fork();                }                for (node<k,v> p; (p = advance()) != null; )                    r = reducer.applyasint(r, transformer.applyasint(p.val));                result = r;                countedcompleter<?> c;                for (c = firstcomplete(); c != null; c = c.nextcomplete()) {                    @suppresswarnings("unchecked")                    mapreducevaluestointtask<k,v>                        t = (mapreducevaluestointtask<k,v>)c,                        s = t.rights;                    while (s != null) {                        t.result = reducer.applyasint(t.result, s.result);                        s = t.rights = s.nextright;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class mapreduceentriestointtask<k,v>        extends bulktask<k,v,integer> {        final tointfunction<map.entry<k,v>> transformer;        final intbinaryoperator reducer;        final int basis;        int result;        mapreduceentriestointtask<k,v> rights, nextright;        mapreduceentriestointtask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             mapreduceentriestointtask<k,v> nextright,             tointfunction<map.entry<k,v>> transformer,             int basis,             intbinaryoperator reducer) {            super(p, b, i, f, t); this.nextright = nextright;            this.transformer = transformer;            this.basis = basis; this.reducer = reducer;        }        public final integer getrawresult() { return result; }        public final void compute() {            final tointfunction<map.entry<k,v>> transformer;            final intbinaryoperator reducer;            if ((transformer = this.transformer) != null &&                (reducer = this.reducer) != null) {                int r = this.basis;                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    (rights = new mapreduceentriestointtask<k,v>                     (this, batch >>>= 1, baselimit = h, f, tab,                      rights, transformer, r, reducer)).fork();                }                for (node<k,v> p; (p = advance()) != null; )                    r = reducer.applyasint(r, transformer.applyasint(p));                result = r;                countedcompleter<?> c;                for (c = firstcomplete(); c != null; c = c.nextcomplete()) {                    @suppresswarnings("unchecked")                    mapreduceentriestointtask<k,v>                        t = (mapreduceentriestointtask<k,v>)c,                        s = t.rights;                    while (s != null) {                        t.result = reducer.applyasint(t.result, s.result);                        s = t.rights = s.nextright;                    }                }            }        }    }
    @suppresswarnings("serial")    static final class mapreducemappingstointtask<k,v>        extends bulktask<k,v,integer> {        final tointbifunction<? super k, ? super v> transformer;        final intbinaryoperator reducer;        final int basis;        int result;        mapreducemappingstointtask<k,v> rights, nextright;        mapreducemappingstointtask            (bulktask<k,v,?> p, int b, int i, int f, node<k,v>[] t,             mapreducemappingstointtask<k,v> nextright,             tointbifunction<? super k, ? super v> transformer,             int basis,             intbinaryoperator reducer) {            super(p, b, i, f, t); this.nextright = nextright;            this.transformer = transformer;            this.basis = basis; this.reducer = reducer;        }        public final integer getrawresult() { return result; }        public final void compute() {            final tointbifunction<? super k, ? super v> transformer;            final intbinaryoperator reducer;            if ((transformer = this.transformer) != null &&                (reducer = this.reducer) != null) {                int r = this.basis;                for (int i = baseindex, f, h; batch > 0 &&                         (h = ((f = baselimit) + i) >>> 1) > i;) {                    addtopendingcount(1);                    (rights = new mapreducemappingstointtask<k,v>                     (this, batch >>>= 1, baselimit = h, f, tab,                      rights, transformer, r, reducer)).fork();                }                for (node<k,v> p; (p = advance()) != null; )                    r = reducer.applyasint(r, transformer.applyasint(p.key, p.val));                result = r;                countedcompleter<?> c;                for (c = firstcomplete(); c != null; c = c.nextcomplete()) {                    @suppresswarnings("unchecked")                    mapreducemappingstointtask<k,v>                        t = (mapreducemappingstointtask<k,v>)c,                        s = t.rights;                    while (s != null) {                        t.result = reducer.applyasint(t.result, s.result);                        s = t.rights = s.nextright;                    }                }            }        }    }
    // unsafe mechanics    private static final sun.misc.unsafe u;    private static final long sizectl;    private static final long transferindex;    private static final long basecount;    private static final long cellsbusy;    private static final long cellvalue;    private static final long abase;    private static final int ashift;
    static {        try {            u = sun.misc.unsafe.getunsafe();            class<?> k = concurrenthashmap.class;            sizectl = u.objectfieldoffset                (k.getdeclaredfield("sizectl"));            transferindex = u.objectfieldoffset                (k.getdeclaredfield("transferindex"));            basecount = u.objectfieldoffset                (k.getdeclaredfield("basecount"));            cellsbusy = u.objectfieldoffset                (k.getdeclaredfield("cellsbusy"));            class<?> ck = countercell.class;            cellvalue = u.objectfieldoffset                (ck.getdeclaredfield("value"));            class<?> ak = node[].class;            abase = u.arraybaseoffset(ak);            int scale = u.arrayindexscale(ak);            if ((scale & (scale - 1)) != 0)                throw new error("data type scale not a power of two");            ashift = 31 - integer.numberofleadingzeros(scale);        } catch (exception e) {            throw new error(e);        }    }}