1601 lines
59 KiB
Java
1601 lines
59 KiB
Java
/*
|
|
* Copyright (c) 2000, 2021, Oracle and/or its affiliates. All rights reserved.
|
|
* DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
|
|
*
|
|
* This code is free software; you can redistribute it and/or modify it
|
|
* under the terms of the GNU General Public License version 2 only, as
|
|
* published by the Free Software Foundation. Oracle designates this
|
|
* particular file as subject to the "Classpath" exception as provided
|
|
* by Oracle in the LICENSE file that accompanied this code.
|
|
*
|
|
* This code is distributed in the hope that it will be useful, but WITHOUT
|
|
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
|
|
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
|
|
* version 2 for more details (a copy is included in the LICENSE file that
|
|
* accompanied this code).
|
|
*
|
|
* You should have received a copy of the GNU General Public License version
|
|
* 2 along with this work; if not, write to the Free Software Foundation,
|
|
* Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
|
|
*
|
|
* Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
|
|
* or visit www.oracle.com if you need additional information or have any
|
|
* questions.
|
|
*/
|
|
|
|
package java.util;
|
|
|
|
import java.io.ObjectInputStream;
|
|
import java.io.ObjectOutputStream;
|
|
import java.lang.reflect.Array;
|
|
import java.util.function.BiConsumer;
|
|
import java.util.function.BiFunction;
|
|
import java.util.function.Consumer;
|
|
import jdk.internal.access.SharedSecrets;
|
|
|
|
/**
|
|
* This class implements the {@code Map} interface with a hash table, using
|
|
* reference-equality in place of object-equality when comparing keys (and
|
|
* values). In other words, in an {@code IdentityHashMap}, two keys
|
|
* {@code k1} and {@code k2} are considered equal if and only if
|
|
* {@code (k1==k2)}. (In normal {@code Map} implementations (like
|
|
* {@code HashMap}) two keys {@code k1} and {@code k2} are considered equal
|
|
* if and only if {@code (k1==null ? k2==null : k1.equals(k2))}.)
|
|
*
|
|
* <p><b>This class is <i>not</i> a general-purpose {@code Map}
|
|
* implementation! While this class implements the {@code Map} interface, it
|
|
* intentionally violates {@code Map's} general contract, which mandates the
|
|
* use of the {@code equals} method when comparing objects. This class is
|
|
* designed for use only in the rare cases wherein reference-equality
|
|
* semantics are required.</b>
|
|
*
|
|
* <p>A typical use of this class is <i>topology-preserving object graph
|
|
* transformations</i>, such as serialization or deep-copying. To perform such
|
|
* a transformation, a program must maintain a "node table" that keeps track
|
|
* of all the object references that have already been processed. The node
|
|
* table must not equate distinct objects even if they happen to be equal.
|
|
* Another typical use of this class is to maintain <i>proxy objects</i>. For
|
|
* example, a debugging facility might wish to maintain a proxy object for
|
|
* each object in the program being debugged.
|
|
*
|
|
* <p>This class provides all of the optional map operations, and permits
|
|
* {@code null} values and the {@code null} key. This class makes no
|
|
* guarantees as to the order of the map; in particular, it does not guarantee
|
|
* that the order will remain constant over time.
|
|
*
|
|
* <p>This class provides constant-time performance for the basic
|
|
* operations ({@code get} and {@code put}), assuming the system
|
|
* identity hash function ({@link System#identityHashCode(Object)})
|
|
* disperses elements properly among the buckets.
|
|
*
|
|
* <p>This class has one tuning parameter (which affects performance but not
|
|
* semantics): <i>expected maximum size</i>. This parameter is the maximum
|
|
* number of key-value mappings that the map is expected to hold. Internally,
|
|
* this parameter is used to determine the number of buckets initially
|
|
* comprising the hash table. The precise relationship between the expected
|
|
* maximum size and the number of buckets is unspecified.
|
|
*
|
|
* <p>If the size of the map (the number of key-value mappings) sufficiently
|
|
* exceeds the expected maximum size, the number of buckets is increased.
|
|
* Increasing the number of buckets ("rehashing") may be fairly expensive, so
|
|
* it pays to create identity hash maps with a sufficiently large expected
|
|
* maximum size. On the other hand, iteration over collection views requires
|
|
* time proportional to the number of buckets in the hash table, so it
|
|
* pays not to set the expected maximum size too high if you are especially
|
|
* concerned with iteration performance or memory usage.
|
|
*
|
|
* <p><strong>Note that this implementation is not synchronized.</strong>
|
|
* If multiple threads access an identity hash map concurrently, and at
|
|
* least one of the threads modifies the map structurally, it <i>must</i>
|
|
* be synchronized externally. (A structural modification is any operation
|
|
* that adds or deletes one or more mappings; merely changing the value
|
|
* associated with a key that an instance already contains is not a
|
|
* structural modification.) This is typically accomplished by
|
|
* synchronizing on some object that naturally encapsulates the map.
|
|
*
|
|
* If no such object exists, the map should be "wrapped" using the
|
|
* {@link Collections#synchronizedMap Collections.synchronizedMap}
|
|
* method. This is best done at creation time, to prevent accidental
|
|
* unsynchronized access to the map:<pre>
|
|
* Map m = Collections.synchronizedMap(new IdentityHashMap(...));</pre>
|
|
*
|
|
* <p>The iterators returned by the {@code iterator} method of the
|
|
* collections returned by all of this class's "collection view
|
|
* methods" are <i>fail-fast</i>: if the map is structurally modified
|
|
* at any time after the iterator is created, in any way except
|
|
* through the iterator's own {@code remove} method, the iterator
|
|
* will throw a {@link ConcurrentModificationException}. Thus, in the
|
|
* face of concurrent modification, the iterator fails quickly and
|
|
* cleanly, rather than risking arbitrary, non-deterministic behavior
|
|
* at an undetermined time in the future.
|
|
*
|
|
* <p>Note that the fail-fast behavior of an iterator cannot be guaranteed
|
|
* as it is, generally speaking, impossible to make any hard guarantees in the
|
|
* presence of unsynchronized concurrent modification. Fail-fast iterators
|
|
* throw {@code ConcurrentModificationException} on a best-effort basis.
|
|
* Therefore, it would be wrong to write a program that depended on this
|
|
* exception for its correctness: <i>fail-fast iterators should be used only
|
|
* to detect bugs.</i>
|
|
*
|
|
* <p>This class is a member of the
|
|
* <a href="{@docRoot}/java.base/java/util/package-summary.html#CollectionsFramework">
|
|
* Java Collections Framework</a>.
|
|
*
|
|
* @implNote
|
|
* <p>This is a simple <i>linear-probe</i> hash table,
|
|
* as described for example in texts by Sedgewick and Knuth. The array
|
|
* contains alternating keys and values, with keys at even indexes and values
|
|
* at odd indexes. (This arrangement has better locality for large
|
|
* tables than does using separate arrays.) For many Java implementations
|
|
* and operation mixes, this class will yield better performance than
|
|
* {@link HashMap}, which uses <i>chaining</i> rather than linear-probing.
|
|
*
|
|
* @see System#identityHashCode(Object)
|
|
* @see Object#hashCode()
|
|
* @see Collection
|
|
* @see Map
|
|
* @see HashMap
|
|
* @see TreeMap
|
|
* @author Doug Lea and Josh Bloch
|
|
* @since 1.4
|
|
*/
|
|
|
|
public class IdentityHashMap<K,V>
|
|
extends AbstractMap<K,V>
|
|
implements Map<K,V>, java.io.Serializable, Cloneable
|
|
{
|
|
/**
|
|
* The initial capacity used by the no-args constructor.
|
|
* MUST be a power of two. The value 32 corresponds to the
|
|
* (specified) expected maximum size of 21, given a load factor
|
|
* of 2/3.
|
|
*/
|
|
private static final int DEFAULT_CAPACITY = 32;
|
|
|
|
/**
|
|
* The minimum capacity, used if a lower value is implicitly specified
|
|
* by either of the constructors with arguments. The value 4 corresponds
|
|
* to an expected maximum size of 2, given a load factor of 2/3.
|
|
* MUST be a power of two.
|
|
*/
|
|
private static final int MINIMUM_CAPACITY = 4;
|
|
|
|
/**
|
|
* The maximum capacity, used if a higher value is implicitly specified
|
|
* by either of the constructors with arguments.
|
|
* MUST be a power of two <= 1<<29.
|
|
*
|
|
* In fact, the map can hold no more than MAXIMUM_CAPACITY-1 items
|
|
* because it has to have at least one slot with the key == null
|
|
* in order to avoid infinite loops in get(), put(), remove()
|
|
*/
|
|
private static final int MAXIMUM_CAPACITY = 1 << 29;
|
|
|
|
/**
|
|
* The table, resized as necessary. Length MUST always be a power of two.
|
|
*/
|
|
transient Object[] table; // non-private to simplify nested class access
|
|
|
|
/**
|
|
* The number of key-value mappings contained in this identity hash map.
|
|
*
|
|
* @serial
|
|
*/
|
|
int size;
|
|
|
|
/**
|
|
* The number of modifications, to support fast-fail iterators
|
|
*/
|
|
transient int modCount;
|
|
|
|
/**
|
|
* Value representing null keys inside tables.
|
|
*/
|
|
static final Object NULL_KEY = new Object();
|
|
|
|
/**
|
|
* Use NULL_KEY for key if it is null.
|
|
*/
|
|
private static Object maskNull(Object key) {
|
|
return (key == null ? NULL_KEY : key);
|
|
}
|
|
|
|
/**
|
|
* Returns internal representation of null key back to caller as null.
|
|
*/
|
|
static final Object unmaskNull(Object key) {
|
|
return (key == NULL_KEY ? null : key);
|
|
}
|
|
|
|
/**
|
|
* Constructs a new, empty identity hash map with a default expected
|
|
* maximum size (21).
|
|
*/
|
|
public IdentityHashMap() {
|
|
init(DEFAULT_CAPACITY);
|
|
}
|
|
|
|
/**
|
|
* Constructs a new, empty map with the specified expected maximum size.
|
|
* Putting more than the expected number of key-value mappings into
|
|
* the map may cause the internal data structure to grow, which may be
|
|
* somewhat time-consuming.
|
|
*
|
|
* @param expectedMaxSize the expected maximum size of the map
|
|
* @throws IllegalArgumentException if {@code expectedMaxSize} is negative
|
|
*/
|
|
public IdentityHashMap(int expectedMaxSize) {
|
|
if (expectedMaxSize < 0)
|
|
throw new IllegalArgumentException("expectedMaxSize is negative: "
|
|
+ expectedMaxSize);
|
|
init(capacity(expectedMaxSize));
|
|
}
|
|
|
|
/**
|
|
* Returns the appropriate capacity for the given expected maximum size.
|
|
* Returns the smallest power of two between MINIMUM_CAPACITY and
|
|
* MAXIMUM_CAPACITY, inclusive, that is greater than (3 *
|
|
* expectedMaxSize)/2, if such a number exists. Otherwise returns
|
|
* MAXIMUM_CAPACITY.
|
|
*/
|
|
private static int capacity(int expectedMaxSize) {
|
|
// assert expectedMaxSize >= 0;
|
|
return
|
|
(expectedMaxSize > MAXIMUM_CAPACITY / 3) ? MAXIMUM_CAPACITY :
|
|
(expectedMaxSize <= 2 * MINIMUM_CAPACITY / 3) ? MINIMUM_CAPACITY :
|
|
Integer.highestOneBit(expectedMaxSize + (expectedMaxSize << 1));
|
|
}
|
|
|
|
/**
|
|
* Initializes object to be an empty map with the specified initial
|
|
* capacity, which is assumed to be a power of two between
|
|
* MINIMUM_CAPACITY and MAXIMUM_CAPACITY inclusive.
|
|
*/
|
|
private void init(int initCapacity) {
|
|
// assert (initCapacity & -initCapacity) == initCapacity; // power of 2
|
|
// assert initCapacity >= MINIMUM_CAPACITY;
|
|
// assert initCapacity <= MAXIMUM_CAPACITY;
|
|
|
|
table = new Object[2 * initCapacity];
|
|
}
|
|
|
|
/**
|
|
* Constructs a new identity hash map containing the keys-value mappings
|
|
* in the specified map.
|
|
*
|
|
* @param m the map whose mappings are to be placed into this map
|
|
* @throws NullPointerException if the specified map is null
|
|
*/
|
|
public IdentityHashMap(Map<? extends K, ? extends V> m) {
|
|
// Allow for a bit of growth
|
|
this((int) ((1 + m.size()) * 1.1));
|
|
putAll(m);
|
|
}
|
|
|
|
/**
|
|
* Returns the number of key-value mappings in this identity hash map.
|
|
*
|
|
* @return the number of key-value mappings in this map
|
|
*/
|
|
public int size() {
|
|
return size;
|
|
}
|
|
|
|
/**
|
|
* Returns {@code true} if this identity hash map contains no key-value
|
|
* mappings.
|
|
*
|
|
* @return {@code true} if this identity hash map contains no key-value
|
|
* mappings
|
|
*/
|
|
public boolean isEmpty() {
|
|
return size == 0;
|
|
}
|
|
|
|
/**
|
|
* Returns index for Object x.
|
|
*/
|
|
private static int hash(Object x, int length) {
|
|
int h = System.identityHashCode(x);
|
|
// Multiply by -254 to use the hash LSB and to ensure index is even
|
|
return ((h << 1) - (h << 8)) & (length - 1);
|
|
}
|
|
|
|
/**
|
|
* Circularly traverses table of size len.
|
|
*/
|
|
private static int nextKeyIndex(int i, int len) {
|
|
return (i + 2 < len ? i + 2 : 0);
|
|
}
|
|
|
|
/**
|
|
* 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 == k)},
|
|
* then this method returns {@code v}; otherwise it returns
|
|
* {@code null}. (There can be at most one such mapping.)
|
|
*
|
|
* <p>A return value of {@code null} does not <i>necessarily</i>
|
|
* indicate that the map contains no mapping for the key; it's also
|
|
* possible that the map explicitly maps the key to {@code null}.
|
|
* The {@link #containsKey containsKey} operation may be used to
|
|
* distinguish these two cases.
|
|
*
|
|
* @see #put(Object, Object)
|
|
*/
|
|
@SuppressWarnings("unchecked")
|
|
public V get(Object key) {
|
|
Object k = maskNull(key);
|
|
Object[] tab = table;
|
|
int len = tab.length;
|
|
int i = hash(k, len);
|
|
while (true) {
|
|
Object item = tab[i];
|
|
if (item == k)
|
|
return (V) tab[i + 1];
|
|
if (item == null)
|
|
return null;
|
|
i = nextKeyIndex(i, len);
|
|
}
|
|
}
|
|
|
|
/**
|
|
* Tests whether the specified object reference is a key in this identity
|
|
* hash map.
|
|
*
|
|
* @param key possible key
|
|
* @return {@code true} if the specified object reference is a key
|
|
* in this map
|
|
* @see #containsValue(Object)
|
|
*/
|
|
public boolean containsKey(Object key) {
|
|
Object k = maskNull(key);
|
|
Object[] tab = table;
|
|
int len = tab.length;
|
|
int i = hash(k, len);
|
|
while (true) {
|
|
Object item = tab[i];
|
|
if (item == k)
|
|
return true;
|
|
if (item == null)
|
|
return false;
|
|
i = nextKeyIndex(i, len);
|
|
}
|
|
}
|
|
|
|
/**
|
|
* Tests whether the specified object reference is a value in this identity
|
|
* hash map.
|
|
*
|
|
* @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 object reference
|
|
* @see #containsKey(Object)
|
|
*/
|
|
public boolean containsValue(Object value) {
|
|
Object[] tab = table;
|
|
for (int i = 1; i < tab.length; i += 2)
|
|
if (tab[i] == value && tab[i - 1] != null)
|
|
return true;
|
|
|
|
return false;
|
|
}
|
|
|
|
/**
|
|
* Tests if the specified key-value mapping is in the map.
|
|
*
|
|
* @param key possible key
|
|
* @param value possible value
|
|
* @return {@code true} if and only if the specified key-value
|
|
* mapping is in the map
|
|
*/
|
|
private boolean containsMapping(Object key, Object value) {
|
|
Object k = maskNull(key);
|
|
Object[] tab = table;
|
|
int len = tab.length;
|
|
int i = hash(k, len);
|
|
while (true) {
|
|
Object item = tab[i];
|
|
if (item == k)
|
|
return tab[i + 1] == value;
|
|
if (item == null)
|
|
return false;
|
|
i = nextKeyIndex(i, len);
|
|
}
|
|
}
|
|
|
|
/**
|
|
* Associates the specified value with the specified key in this identity
|
|
* hash map. If the map previously contained a mapping for the key, the
|
|
* old value is replaced.
|
|
*
|
|
* @param key the key with which the specified value is to be associated
|
|
* @param value the 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}.
|
|
* (A {@code null} return can also indicate that the map
|
|
* previously associated {@code null} with {@code key}.)
|
|
* @see Object#equals(Object)
|
|
* @see #get(Object)
|
|
* @see #containsKey(Object)
|
|
*/
|
|
public V put(K key, V value) {
|
|
final Object k = maskNull(key);
|
|
|
|
retryAfterResize: for (;;) {
|
|
final Object[] tab = table;
|
|
final int len = tab.length;
|
|
int i = hash(k, len);
|
|
|
|
for (Object item; (item = tab[i]) != null;
|
|
i = nextKeyIndex(i, len)) {
|
|
if (item == k) {
|
|
@SuppressWarnings("unchecked")
|
|
V oldValue = (V) tab[i + 1];
|
|
tab[i + 1] = value;
|
|
return oldValue;
|
|
}
|
|
}
|
|
|
|
final int s = size + 1;
|
|
// Use optimized form of 3 * s.
|
|
// Next capacity is len, 2 * current capacity.
|
|
if (s + (s << 1) > len && resize(len))
|
|
continue retryAfterResize;
|
|
|
|
modCount++;
|
|
tab[i] = k;
|
|
tab[i + 1] = value;
|
|
size = s;
|
|
return null;
|
|
}
|
|
}
|
|
|
|
/**
|
|
* Resizes the table if necessary to hold given capacity.
|
|
*
|
|
* @param newCapacity the new capacity, must be a power of two.
|
|
* @return whether a resize did in fact take place
|
|
*/
|
|
private boolean resize(int newCapacity) {
|
|
// assert (newCapacity & -newCapacity) == newCapacity; // power of 2
|
|
int newLength = newCapacity * 2;
|
|
|
|
Object[] oldTable = table;
|
|
int oldLength = oldTable.length;
|
|
if (oldLength == 2 * MAXIMUM_CAPACITY) { // can't expand any further
|
|
if (size == MAXIMUM_CAPACITY - 1)
|
|
throw new IllegalStateException("Capacity exhausted.");
|
|
return false;
|
|
}
|
|
if (oldLength >= newLength)
|
|
return false;
|
|
|
|
Object[] newTable = new Object[newLength];
|
|
|
|
for (int j = 0; j < oldLength; j += 2) {
|
|
Object key = oldTable[j];
|
|
if (key != null) {
|
|
Object value = oldTable[j+1];
|
|
oldTable[j] = null;
|
|
oldTable[j+1] = null;
|
|
int i = hash(key, newLength);
|
|
while (newTable[i] != null)
|
|
i = nextKeyIndex(i, newLength);
|
|
newTable[i] = key;
|
|
newTable[i + 1] = value;
|
|
}
|
|
}
|
|
table = newTable;
|
|
return true;
|
|
}
|
|
|
|
/**
|
|
* Copies all of the mappings from the specified map to this map.
|
|
* These mappings will 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
|
|
* @throws NullPointerException if the specified map is null
|
|
*/
|
|
public void putAll(Map<? extends K, ? extends V> m) {
|
|
int n = m.size();
|
|
if (n == 0)
|
|
return;
|
|
if (n > size)
|
|
resize(capacity(n)); // conservatively pre-expand
|
|
|
|
for (Entry<? extends K, ? extends V> e : m.entrySet())
|
|
put(e.getKey(), e.getValue());
|
|
}
|
|
|
|
/**
|
|
* Removes the mapping for this key from this map if present.
|
|
*
|
|
* @param key key whose mapping is to be removed from the map
|
|
* @return the previous value associated with {@code key}, or
|
|
* {@code null} if there was no mapping for {@code key}.
|
|
* (A {@code null} return can also indicate that the map
|
|
* previously associated {@code null} with {@code key}.)
|
|
*/
|
|
public V remove(Object key) {
|
|
Object k = maskNull(key);
|
|
Object[] tab = table;
|
|
int len = tab.length;
|
|
int i = hash(k, len);
|
|
|
|
while (true) {
|
|
Object item = tab[i];
|
|
if (item == k) {
|
|
modCount++;
|
|
size--;
|
|
@SuppressWarnings("unchecked")
|
|
V oldValue = (V) tab[i + 1];
|
|
tab[i + 1] = null;
|
|
tab[i] = null;
|
|
closeDeletion(i);
|
|
return oldValue;
|
|
}
|
|
if (item == null)
|
|
return null;
|
|
i = nextKeyIndex(i, len);
|
|
}
|
|
}
|
|
|
|
/**
|
|
* Removes the specified key-value mapping from the map if it is present.
|
|
*
|
|
* @param key possible key
|
|
* @param value possible value
|
|
* @return {@code true} if and only if the specified key-value
|
|
* mapping was in the map
|
|
*/
|
|
private boolean removeMapping(Object key, Object value) {
|
|
Object k = maskNull(key);
|
|
Object[] tab = table;
|
|
int len = tab.length;
|
|
int i = hash(k, len);
|
|
|
|
while (true) {
|
|
Object item = tab[i];
|
|
if (item == k) {
|
|
if (tab[i + 1] != value)
|
|
return false;
|
|
modCount++;
|
|
size--;
|
|
tab[i] = null;
|
|
tab[i + 1] = null;
|
|
closeDeletion(i);
|
|
return true;
|
|
}
|
|
if (item == null)
|
|
return false;
|
|
i = nextKeyIndex(i, len);
|
|
}
|
|
}
|
|
|
|
/**
|
|
* Rehash all possibly-colliding entries following a
|
|
* deletion. This preserves the linear-probe
|
|
* collision properties required by get, put, etc.
|
|
*
|
|
* @param d the index of a newly empty deleted slot
|
|
*/
|
|
private void closeDeletion(int d) {
|
|
// Adapted from Knuth Section 6.4 Algorithm R
|
|
Object[] tab = table;
|
|
int len = tab.length;
|
|
|
|
// Look for items to swap into newly vacated slot
|
|
// starting at index immediately following deletion,
|
|
// and continuing until a null slot is seen, indicating
|
|
// the end of a run of possibly-colliding keys.
|
|
Object item;
|
|
for (int i = nextKeyIndex(d, len); (item = tab[i]) != null;
|
|
i = nextKeyIndex(i, len) ) {
|
|
// The following test triggers if the item at slot i (which
|
|
// hashes to be at slot r) should take the spot vacated by d.
|
|
// If so, we swap it in, and then continue with d now at the
|
|
// newly vacated i. This process will terminate when we hit
|
|
// the null slot at the end of this run.
|
|
// The test is messy because we are using a circular table.
|
|
int r = hash(item, len);
|
|
if ((i < r && (r <= d || d <= i)) || (r <= d && d <= i)) {
|
|
tab[d] = item;
|
|
tab[d + 1] = tab[i + 1];
|
|
tab[i] = null;
|
|
tab[i + 1] = null;
|
|
d = i;
|
|
}
|
|
}
|
|
}
|
|
|
|
/**
|
|
* Removes all of the mappings from this map.
|
|
* The map will be empty after this call returns.
|
|
*/
|
|
public void clear() {
|
|
modCount++;
|
|
Object[] tab = table;
|
|
for (int i = 0; i < tab.length; i++)
|
|
tab[i] = null;
|
|
size = 0;
|
|
}
|
|
|
|
/**
|
|
* Compares the specified object with this map for equality. Returns
|
|
* {@code true} if the given object is also a map and the two maps
|
|
* represent identical object-reference mappings. More formally, this
|
|
* map is equal to another map {@code m} if and only if
|
|
* {@code this.entrySet().equals(m.entrySet())}.
|
|
*
|
|
* <p><b>Owing to the reference-equality-based semantics of this map it is
|
|
* possible that the symmetry and transitivity requirements of the
|
|
* {@code Object.equals} contract may be violated if this map is compared
|
|
* to a normal map. However, the {@code Object.equals} contract is
|
|
* guaranteed to hold among {@code IdentityHashMap} instances.</b>
|
|
*
|
|
* @param o object to be compared for equality with this map
|
|
* @return {@code true} if the specified object is equal to this map
|
|
* @see Object#equals(Object)
|
|
*/
|
|
public boolean equals(Object o) {
|
|
if (o == this) {
|
|
return true;
|
|
} else if (o instanceof IdentityHashMap<?, ?> m) {
|
|
if (m.size() != size)
|
|
return false;
|
|
|
|
Object[] tab = m.table;
|
|
for (int i = 0; i < tab.length; i+=2) {
|
|
Object k = tab[i];
|
|
if (k != null && !containsMapping(k, tab[i + 1]))
|
|
return false;
|
|
}
|
|
return true;
|
|
} else if (o instanceof Map<?, ?> m) {
|
|
return entrySet().equals(m.entrySet());
|
|
} else {
|
|
return false; // o is not a Map
|
|
}
|
|
}
|
|
|
|
/**
|
|
* Returns the hash code value for this map. The hash code of a map is
|
|
* defined to be the sum of the hash codes of each entry in the map's
|
|
* {@code entrySet()} view. This ensures that {@code m1.equals(m2)}
|
|
* implies that {@code m1.hashCode()==m2.hashCode()} for any two
|
|
* {@code IdentityHashMap} instances {@code m1} and {@code m2}, as
|
|
* required by the general contract of {@link Object#hashCode}.
|
|
*
|
|
* <p><b>Owing to the reference-equality-based semantics of the
|
|
* {@code Map.Entry} instances in the set returned by this map's
|
|
* {@code entrySet} method, it is possible that the contractual
|
|
* requirement of {@code Object.hashCode} mentioned in the previous
|
|
* paragraph will be violated if one of the two objects being compared is
|
|
* an {@code IdentityHashMap} instance and the other is a normal map.</b>
|
|
*
|
|
* @return the hash code value for this map
|
|
* @see Object#equals(Object)
|
|
* @see #equals(Object)
|
|
*/
|
|
public int hashCode() {
|
|
int result = 0;
|
|
Object[] tab = table;
|
|
for (int i = 0; i < tab.length; i +=2) {
|
|
Object key = tab[i];
|
|
if (key != null) {
|
|
Object k = unmaskNull(key);
|
|
result += System.identityHashCode(k) ^
|
|
System.identityHashCode(tab[i + 1]);
|
|
}
|
|
}
|
|
return result;
|
|
}
|
|
|
|
/**
|
|
* Returns a shallow copy of this identity hash map: the keys and values
|
|
* themselves are not cloned.
|
|
*
|
|
* @return a shallow copy of this map
|
|
*/
|
|
public Object clone() {
|
|
try {
|
|
IdentityHashMap<?,?> m = (IdentityHashMap<?,?>) super.clone();
|
|
m.entrySet = null;
|
|
m.table = table.clone();
|
|
return m;
|
|
} catch (CloneNotSupportedException e) {
|
|
throw new InternalError(e);
|
|
}
|
|
}
|
|
|
|
private abstract class IdentityHashMapIterator<T> implements Iterator<T> {
|
|
int index = (size != 0 ? 0 : table.length); // current slot.
|
|
int expectedModCount = modCount; // to support fast-fail
|
|
int lastReturnedIndex = -1; // to allow remove()
|
|
boolean indexValid; // To avoid unnecessary next computation
|
|
Object[] traversalTable = table; // reference to main table or copy
|
|
|
|
public boolean hasNext() {
|
|
Object[] tab = traversalTable;
|
|
for (int i = index; i < tab.length; i+=2) {
|
|
Object key = tab[i];
|
|
if (key != null) {
|
|
index = i;
|
|
return indexValid = true;
|
|
}
|
|
}
|
|
index = tab.length;
|
|
return false;
|
|
}
|
|
|
|
protected int nextIndex() {
|
|
if (modCount != expectedModCount)
|
|
throw new ConcurrentModificationException();
|
|
if (!indexValid && !hasNext())
|
|
throw new NoSuchElementException();
|
|
|
|
indexValid = false;
|
|
lastReturnedIndex = index;
|
|
index += 2;
|
|
return lastReturnedIndex;
|
|
}
|
|
|
|
public void remove() {
|
|
if (lastReturnedIndex == -1)
|
|
throw new IllegalStateException();
|
|
if (modCount != expectedModCount)
|
|
throw new ConcurrentModificationException();
|
|
|
|
expectedModCount = ++modCount;
|
|
int deletedSlot = lastReturnedIndex;
|
|
lastReturnedIndex = -1;
|
|
// back up index to revisit new contents after deletion
|
|
index = deletedSlot;
|
|
indexValid = false;
|
|
|
|
// Removal code proceeds as in closeDeletion except that
|
|
// it must catch the rare case where an element already
|
|
// seen is swapped into a vacant slot that will be later
|
|
// traversed by this iterator. We cannot allow future
|
|
// next() calls to return it again. The likelihood of
|
|
// this occurring under 2/3 load factor is very slim, but
|
|
// when it does happen, we must make a copy of the rest of
|
|
// the table to use for the rest of the traversal. Since
|
|
// this can only happen when we are near the end of the table,
|
|
// even in these rare cases, this is not very expensive in
|
|
// time or space.
|
|
|
|
Object[] tab = traversalTable;
|
|
int len = tab.length;
|
|
|
|
int d = deletedSlot;
|
|
Object key = tab[d];
|
|
tab[d] = null; // vacate the slot
|
|
tab[d + 1] = null;
|
|
|
|
// If traversing a copy, remove in real table.
|
|
// We can skip gap-closure on copy.
|
|
if (tab != IdentityHashMap.this.table) {
|
|
IdentityHashMap.this.remove(key);
|
|
expectedModCount = modCount;
|
|
return;
|
|
}
|
|
|
|
size--;
|
|
|
|
Object item;
|
|
for (int i = nextKeyIndex(d, len); (item = tab[i]) != null;
|
|
i = nextKeyIndex(i, len)) {
|
|
int r = hash(item, len);
|
|
// See closeDeletion for explanation of this conditional
|
|
if ((i < r && (r <= d || d <= i)) ||
|
|
(r <= d && d <= i)) {
|
|
|
|
// If we are about to swap an already-seen element
|
|
// into a slot that may later be returned by next(),
|
|
// then clone the rest of table for use in future
|
|
// next() calls. It is OK that our copy will have
|
|
// a gap in the "wrong" place, since it will never
|
|
// be used for searching anyway.
|
|
|
|
if (i < deletedSlot && d >= deletedSlot &&
|
|
traversalTable == IdentityHashMap.this.table) {
|
|
int remaining = len - deletedSlot;
|
|
Object[] newTable = new Object[remaining];
|
|
System.arraycopy(tab, deletedSlot,
|
|
newTable, 0, remaining);
|
|
traversalTable = newTable;
|
|
index = 0;
|
|
}
|
|
|
|
tab[d] = item;
|
|
tab[d + 1] = tab[i + 1];
|
|
tab[i] = null;
|
|
tab[i + 1] = null;
|
|
d = i;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
private class KeyIterator extends IdentityHashMapIterator<K> {
|
|
@SuppressWarnings("unchecked")
|
|
public K next() {
|
|
return (K) unmaskNull(traversalTable[nextIndex()]);
|
|
}
|
|
}
|
|
|
|
private class ValueIterator extends IdentityHashMapIterator<V> {
|
|
@SuppressWarnings("unchecked")
|
|
public V next() {
|
|
return (V) traversalTable[nextIndex() + 1];
|
|
}
|
|
}
|
|
|
|
private class EntryIterator
|
|
extends IdentityHashMapIterator<Map.Entry<K,V>>
|
|
{
|
|
private Entry lastReturnedEntry;
|
|
|
|
public Map.Entry<K,V> next() {
|
|
lastReturnedEntry = new Entry(nextIndex());
|
|
return lastReturnedEntry;
|
|
}
|
|
|
|
public void remove() {
|
|
lastReturnedIndex =
|
|
((null == lastReturnedEntry) ? -1 : lastReturnedEntry.index);
|
|
super.remove();
|
|
lastReturnedEntry.index = lastReturnedIndex;
|
|
lastReturnedEntry = null;
|
|
}
|
|
|
|
private class Entry implements Map.Entry<K,V> {
|
|
private int index;
|
|
|
|
private Entry(int index) {
|
|
this.index = index;
|
|
}
|
|
|
|
@SuppressWarnings("unchecked")
|
|
public K getKey() {
|
|
checkIndexForEntryUse();
|
|
return (K) unmaskNull(traversalTable[index]);
|
|
}
|
|
|
|
@SuppressWarnings("unchecked")
|
|
public V getValue() {
|
|
checkIndexForEntryUse();
|
|
return (V) traversalTable[index+1];
|
|
}
|
|
|
|
@SuppressWarnings("unchecked")
|
|
public V setValue(V value) {
|
|
checkIndexForEntryUse();
|
|
V oldValue = (V) traversalTable[index+1];
|
|
traversalTable[index+1] = value;
|
|
// if shadowing, force into main table
|
|
if (traversalTable != IdentityHashMap.this.table)
|
|
put((K) traversalTable[index], value);
|
|
return oldValue;
|
|
}
|
|
|
|
public boolean equals(Object o) {
|
|
if (index < 0)
|
|
return super.equals(o);
|
|
|
|
return o instanceof Map.Entry<?, ?> e
|
|
&& e.getKey() == unmaskNull(traversalTable[index])
|
|
&& e.getValue() == traversalTable[index+1];
|
|
}
|
|
|
|
public int hashCode() {
|
|
if (lastReturnedIndex < 0)
|
|
return super.hashCode();
|
|
|
|
return (System.identityHashCode(unmaskNull(traversalTable[index])) ^
|
|
System.identityHashCode(traversalTable[index+1]));
|
|
}
|
|
|
|
public String toString() {
|
|
if (index < 0)
|
|
return super.toString();
|
|
|
|
return (unmaskNull(traversalTable[index]) + "="
|
|
+ traversalTable[index+1]);
|
|
}
|
|
|
|
private void checkIndexForEntryUse() {
|
|
if (index < 0)
|
|
throw new IllegalStateException("Entry was removed");
|
|
}
|
|
}
|
|
}
|
|
|
|
// Views
|
|
|
|
/**
|
|
* This field is initialized to contain an instance of the entry set
|
|
* view the first time this view is requested. The view is stateless,
|
|
* so there's no reason to create more than one.
|
|
*/
|
|
private transient Set<Map.Entry<K,V>> entrySet;
|
|
|
|
/**
|
|
* Returns an identity-based 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. If the map is modified while an iteration
|
|
* over the set is in progress, the results of the iteration are
|
|
* undefined. 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} methods. It does not support the {@code add} or
|
|
* {@code addAll} methods.
|
|
*
|
|
* <p><b>While the object returned by this method implements the
|
|
* {@code Set} interface, it does <i>not</i> obey {@code Set's} general
|
|
* contract. Like its backing map, the set returned by this method
|
|
* defines element equality as reference-equality rather than
|
|
* object-equality. This affects the behavior of its {@code contains},
|
|
* {@code remove}, {@code containsAll}, {@code equals}, and
|
|
* {@code hashCode} methods.</b>
|
|
*
|
|
* <p><b>The {@code equals} method of the returned set returns {@code true}
|
|
* only if the specified object is a set containing exactly the same
|
|
* object references as the returned set. The symmetry and transitivity
|
|
* requirements of the {@code Object.equals} contract may be violated if
|
|
* the set returned by this method is compared to a normal set. However,
|
|
* the {@code Object.equals} contract is guaranteed to hold among sets
|
|
* returned by this method.</b>
|
|
*
|
|
* <p>The {@code hashCode} method of the returned set returns the sum of
|
|
* the <i>identity hashcodes</i> of the elements in the set, rather than
|
|
* the sum of their hashcodes. This is mandated by the change in the
|
|
* semantics of the {@code equals} method, in order to enforce the
|
|
* general contract of the {@code Object.hashCode} method among sets
|
|
* returned by this method.
|
|
*
|
|
* @return an identity-based set view of the keys contained in this map
|
|
* @see Object#equals(Object)
|
|
* @see System#identityHashCode(Object)
|
|
*/
|
|
public Set<K> keySet() {
|
|
Set<K> ks = keySet;
|
|
if (ks == null) {
|
|
ks = new KeySet();
|
|
keySet = ks;
|
|
}
|
|
return ks;
|
|
}
|
|
|
|
private class KeySet extends AbstractSet<K> {
|
|
public Iterator<K> iterator() {
|
|
return new KeyIterator();
|
|
}
|
|
public int size() {
|
|
return size;
|
|
}
|
|
public boolean contains(Object o) {
|
|
return containsKey(o);
|
|
}
|
|
public boolean remove(Object o) {
|
|
int oldSize = size;
|
|
IdentityHashMap.this.remove(o);
|
|
return size != oldSize;
|
|
}
|
|
/*
|
|
* Must revert from AbstractSet's impl to AbstractCollection's, as
|
|
* the former contains an optimization that results in incorrect
|
|
* behavior when c is a smaller "normal" (non-identity-based) Set.
|
|
*/
|
|
public boolean removeAll(Collection<?> c) {
|
|
Objects.requireNonNull(c);
|
|
boolean modified = false;
|
|
for (Iterator<K> i = iterator(); i.hasNext(); ) {
|
|
if (c.contains(i.next())) {
|
|
i.remove();
|
|
modified = true;
|
|
}
|
|
}
|
|
return modified;
|
|
}
|
|
public void clear() {
|
|
IdentityHashMap.this.clear();
|
|
}
|
|
public int hashCode() {
|
|
int result = 0;
|
|
for (K key : this)
|
|
result += System.identityHashCode(key);
|
|
return result;
|
|
}
|
|
public Object[] toArray() {
|
|
return toArray(new Object[0]);
|
|
}
|
|
@SuppressWarnings("unchecked")
|
|
public <T> T[] toArray(T[] a) {
|
|
int expectedModCount = modCount;
|
|
int size = size();
|
|
if (a.length < size)
|
|
a = (T[]) Array.newInstance(a.getClass().getComponentType(), size);
|
|
Object[] tab = table;
|
|
int ti = 0;
|
|
for (int si = 0; si < tab.length; si += 2) {
|
|
Object key;
|
|
if ((key = tab[si]) != null) { // key present ?
|
|
// more elements than expected -> concurrent modification from other thread
|
|
if (ti >= size) {
|
|
throw new ConcurrentModificationException();
|
|
}
|
|
a[ti++] = (T) unmaskNull(key); // unmask key
|
|
}
|
|
}
|
|
// fewer elements than expected or concurrent modification from other thread detected
|
|
if (ti < size || expectedModCount != modCount) {
|
|
throw new ConcurrentModificationException();
|
|
}
|
|
// final null marker as per spec
|
|
if (ti < a.length) {
|
|
a[ti] = null;
|
|
}
|
|
return a;
|
|
}
|
|
|
|
public Spliterator<K> spliterator() {
|
|
return new KeySpliterator<>(IdentityHashMap.this, 0, -1, 0, 0);
|
|
}
|
|
}
|
|
|
|
/**
|
|
* 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. If the map is
|
|
* modified while an iteration over the collection is in progress,
|
|
* the results of the iteration are undefined. The collection
|
|
* supports element removal, which removes the corresponding
|
|
* mapping from the map, via the {@code Iterator.remove},
|
|
* {@code Collection.remove}, {@code removeAll},
|
|
* {@code retainAll} and {@code clear} methods. It does not
|
|
* support the {@code add} or {@code addAll} methods.
|
|
*
|
|
* <p><b>While the object returned by this method implements the
|
|
* {@code Collection} interface, it does <i>not</i> obey
|
|
* {@code Collection's} general contract. Like its backing map,
|
|
* the collection returned by this method defines element equality as
|
|
* reference-equality rather than object-equality. This affects the
|
|
* behavior of its {@code contains}, {@code remove} and
|
|
* {@code containsAll} methods.</b>
|
|
*/
|
|
public Collection<V> values() {
|
|
Collection<V> vs = values;
|
|
if (vs == null) {
|
|
vs = new Values();
|
|
values = vs;
|
|
}
|
|
return vs;
|
|
}
|
|
|
|
private class Values extends AbstractCollection<V> {
|
|
public Iterator<V> iterator() {
|
|
return new ValueIterator();
|
|
}
|
|
public int size() {
|
|
return size;
|
|
}
|
|
public boolean contains(Object o) {
|
|
return containsValue(o);
|
|
}
|
|
public boolean remove(Object o) {
|
|
for (Iterator<V> i = iterator(); i.hasNext(); ) {
|
|
if (i.next() == o) {
|
|
i.remove();
|
|
return true;
|
|
}
|
|
}
|
|
return false;
|
|
}
|
|
public void clear() {
|
|
IdentityHashMap.this.clear();
|
|
}
|
|
public Object[] toArray() {
|
|
return toArray(new Object[0]);
|
|
}
|
|
@SuppressWarnings("unchecked")
|
|
public <T> T[] toArray(T[] a) {
|
|
int expectedModCount = modCount;
|
|
int size = size();
|
|
if (a.length < size)
|
|
a = (T[]) Array.newInstance(a.getClass().getComponentType(), size);
|
|
Object[] tab = table;
|
|
int ti = 0;
|
|
for (int si = 0; si < tab.length; si += 2) {
|
|
if (tab[si] != null) { // key present ?
|
|
// more elements than expected -> concurrent modification from other thread
|
|
if (ti >= size) {
|
|
throw new ConcurrentModificationException();
|
|
}
|
|
a[ti++] = (T) tab[si+1]; // copy value
|
|
}
|
|
}
|
|
// fewer elements than expected or concurrent modification from other thread detected
|
|
if (ti < size || expectedModCount != modCount) {
|
|
throw new ConcurrentModificationException();
|
|
}
|
|
// final null marker as per spec
|
|
if (ti < a.length) {
|
|
a[ti] = null;
|
|
}
|
|
return a;
|
|
}
|
|
|
|
public Spliterator<V> spliterator() {
|
|
return new ValueSpliterator<>(IdentityHashMap.this, 0, -1, 0, 0);
|
|
}
|
|
}
|
|
|
|
/**
|
|
* Returns a {@link Set} view of the mappings contained in this map.
|
|
* Each element in the returned set is a reference-equality-based
|
|
* {@code Map.Entry}. The set is backed by the map, so changes
|
|
* to the map are reflected in the set, and vice-versa. If the
|
|
* map is modified while an iteration over the set is in progress,
|
|
* the results of the iteration are undefined. 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}
|
|
* methods. It does not support the {@code add} or
|
|
* {@code addAll} methods.
|
|
*
|
|
* <p>Like the backing map, the {@code Map.Entry} objects in the set
|
|
* returned by this method define key and value equality as
|
|
* reference-equality rather than object-equality. This affects the
|
|
* behavior of the {@code equals} and {@code hashCode} methods of these
|
|
* {@code Map.Entry} objects. A reference-equality based {@code Map.Entry
|
|
* e} is equal to an object {@code o} if and only if {@code o} is a
|
|
* {@code Map.Entry} and {@code e.getKey()==o.getKey() &&
|
|
* e.getValue()==o.getValue()}. To accommodate these equals
|
|
* semantics, the {@code hashCode} method returns
|
|
* {@code System.identityHashCode(e.getKey()) ^
|
|
* System.identityHashCode(e.getValue())}.
|
|
*
|
|
* <p><b>Owing to the reference-equality-based semantics of the
|
|
* {@code Map.Entry} instances in the set returned by this method,
|
|
* it is possible that the symmetry and transitivity requirements of
|
|
* the {@link Object#equals(Object)} contract may be violated if any of
|
|
* the entries in the set is compared to a normal map entry, or if
|
|
* the set returned by this method is compared to a set of normal map
|
|
* entries (such as would be returned by a call to this method on a normal
|
|
* map). However, the {@code Object.equals} contract is guaranteed to
|
|
* hold among identity-based map entries, and among sets of such entries.
|
|
* </b>
|
|
*
|
|
* @return a set view of the identity-mappings contained in this map
|
|
*/
|
|
public Set<Map.Entry<K,V>> entrySet() {
|
|
Set<Map.Entry<K,V>> es = entrySet;
|
|
if (es != null)
|
|
return es;
|
|
else
|
|
return entrySet = new EntrySet();
|
|
}
|
|
|
|
private class EntrySet extends AbstractSet<Map.Entry<K,V>> {
|
|
public Iterator<Map.Entry<K,V>> iterator() {
|
|
return new EntryIterator();
|
|
}
|
|
public boolean contains(Object o) {
|
|
return o instanceof Entry<?, ?> entry
|
|
&& containsMapping(entry.getKey(), entry.getValue());
|
|
}
|
|
public boolean remove(Object o) {
|
|
return o instanceof Entry<?, ?> entry
|
|
&& removeMapping(entry.getKey(), entry.getValue());
|
|
}
|
|
public int size() {
|
|
return size;
|
|
}
|
|
public void clear() {
|
|
IdentityHashMap.this.clear();
|
|
}
|
|
/*
|
|
* Must revert from AbstractSet's impl to AbstractCollection's, as
|
|
* the former contains an optimization that results in incorrect
|
|
* behavior when c is a smaller "normal" (non-identity-based) Set.
|
|
*/
|
|
public boolean removeAll(Collection<?> c) {
|
|
Objects.requireNonNull(c);
|
|
boolean modified = false;
|
|
for (Iterator<Map.Entry<K,V>> i = iterator(); i.hasNext(); ) {
|
|
if (c.contains(i.next())) {
|
|
i.remove();
|
|
modified = true;
|
|
}
|
|
}
|
|
return modified;
|
|
}
|
|
|
|
public Object[] toArray() {
|
|
return toArray(new Object[0]);
|
|
}
|
|
|
|
@SuppressWarnings("unchecked")
|
|
public <T> T[] toArray(T[] a) {
|
|
int expectedModCount = modCount;
|
|
int size = size();
|
|
if (a.length < size)
|
|
a = (T[]) Array.newInstance(a.getClass().getComponentType(), size);
|
|
Object[] tab = table;
|
|
int ti = 0;
|
|
for (int si = 0; si < tab.length; si += 2) {
|
|
Object key;
|
|
if ((key = tab[si]) != null) { // key present ?
|
|
// more elements than expected -> concurrent modification from other thread
|
|
if (ti >= size) {
|
|
throw new ConcurrentModificationException();
|
|
}
|
|
a[ti++] = (T) new AbstractMap.SimpleEntry<>(unmaskNull(key), tab[si + 1]);
|
|
}
|
|
}
|
|
// fewer elements than expected or concurrent modification from other thread detected
|
|
if (ti < size || expectedModCount != modCount) {
|
|
throw new ConcurrentModificationException();
|
|
}
|
|
// final null marker as per spec
|
|
if (ti < a.length) {
|
|
a[ti] = null;
|
|
}
|
|
return a;
|
|
}
|
|
|
|
public Spliterator<Map.Entry<K,V>> spliterator() {
|
|
return new EntrySpliterator<>(IdentityHashMap.this, 0, -1, 0, 0);
|
|
}
|
|
}
|
|
|
|
@java.io.Serial
|
|
private static final long serialVersionUID = 8188218128353913216L;
|
|
|
|
/**
|
|
* Saves the state of the {@code IdentityHashMap} instance to a stream
|
|
* (i.e., serializes it).
|
|
*
|
|
* @serialData The <i>size</i> of the HashMap (the number of key-value
|
|
* mappings) ({@code int}), followed by the key (Object) and
|
|
* value (Object) for each key-value mapping represented by the
|
|
* IdentityHashMap. The key-value mappings are emitted in no
|
|
* particular order.
|
|
*/
|
|
@java.io.Serial
|
|
private void writeObject(ObjectOutputStream s)
|
|
throws java.io.IOException {
|
|
// Write out size (number of mappings) and any hidden stuff
|
|
s.defaultWriteObject();
|
|
|
|
// Write out size again (maintained for backward compatibility)
|
|
s.writeInt(size);
|
|
|
|
// Write out keys and values (alternating)
|
|
Object[] tab = table;
|
|
for (int i = 0; i < tab.length; i += 2) {
|
|
Object key = tab[i];
|
|
if (key != null) {
|
|
s.writeObject(unmaskNull(key));
|
|
s.writeObject(tab[i + 1]);
|
|
}
|
|
}
|
|
}
|
|
|
|
/**
|
|
* Reconstitutes the {@code IdentityHashMap} instance from a stream (i.e.,
|
|
* deserializes it).
|
|
*/
|
|
@java.io.Serial
|
|
private void readObject(ObjectInputStream s)
|
|
throws java.io.IOException, ClassNotFoundException {
|
|
// Size (number of mappings) is written to the stream twice
|
|
// Read first size value and ignore it
|
|
s.readFields();
|
|
|
|
// Read second size value, validate and assign to size field
|
|
int size = s.readInt();
|
|
if (size < 0)
|
|
throw new java.io.StreamCorruptedException
|
|
("Illegal mappings count: " + size);
|
|
int cap = capacity(size);
|
|
SharedSecrets.getJavaObjectInputStreamAccess().checkArray(s, Object[].class, cap*2);
|
|
this.size = size;
|
|
init(cap);
|
|
|
|
// Read the keys and values, and put the mappings in the table
|
|
for (int i=0; i<size; i++) {
|
|
@SuppressWarnings("unchecked")
|
|
K key = (K) s.readObject();
|
|
@SuppressWarnings("unchecked")
|
|
V value = (V) s.readObject();
|
|
putForCreate(key, value);
|
|
}
|
|
}
|
|
|
|
/**
|
|
* The put method for readObject. It does not resize the table,
|
|
* update modCount, etc.
|
|
*/
|
|
private void putForCreate(K key, V value)
|
|
throws java.io.StreamCorruptedException
|
|
{
|
|
Object k = maskNull(key);
|
|
Object[] tab = table;
|
|
int len = tab.length;
|
|
int i = hash(k, len);
|
|
|
|
Object item;
|
|
while ( (item = tab[i]) != null) {
|
|
if (item == k)
|
|
throw new java.io.StreamCorruptedException();
|
|
i = nextKeyIndex(i, len);
|
|
}
|
|
tab[i] = k;
|
|
tab[i + 1] = value;
|
|
}
|
|
|
|
@SuppressWarnings("unchecked")
|
|
@Override
|
|
public void forEach(BiConsumer<? super K, ? super V> action) {
|
|
Objects.requireNonNull(action);
|
|
int expectedModCount = modCount;
|
|
|
|
Object[] t = table;
|
|
for (int index = 0; index < t.length; index += 2) {
|
|
Object k = t[index];
|
|
if (k != null) {
|
|
action.accept((K) unmaskNull(k), (V) t[index + 1]);
|
|
}
|
|
|
|
if (modCount != expectedModCount) {
|
|
throw new ConcurrentModificationException();
|
|
}
|
|
}
|
|
}
|
|
|
|
@SuppressWarnings("unchecked")
|
|
@Override
|
|
public void replaceAll(BiFunction<? super K, ? super V, ? extends V> function) {
|
|
Objects.requireNonNull(function);
|
|
int expectedModCount = modCount;
|
|
|
|
Object[] t = table;
|
|
for (int index = 0; index < t.length; index += 2) {
|
|
Object k = t[index];
|
|
if (k != null) {
|
|
t[index + 1] = function.apply((K) unmaskNull(k), (V) t[index + 1]);
|
|
}
|
|
|
|
if (modCount != expectedModCount) {
|
|
throw new ConcurrentModificationException();
|
|
}
|
|
}
|
|
}
|
|
|
|
/**
|
|
* Similar form as array-based Spliterators, but skips blank elements,
|
|
* and guestimates size as decreasing by half per split.
|
|
*/
|
|
static class IdentityHashMapSpliterator<K,V> {
|
|
final IdentityHashMap<K,V> map;
|
|
int index; // current index, modified on advance/split
|
|
int fence; // -1 until first use; then one past last index
|
|
int est; // size estimate
|
|
int expectedModCount; // initialized when fence set
|
|
|
|
IdentityHashMapSpliterator(IdentityHashMap<K,V> map, int origin,
|
|
int fence, int est, int expectedModCount) {
|
|
this.map = map;
|
|
this.index = origin;
|
|
this.fence = fence;
|
|
this.est = est;
|
|
this.expectedModCount = expectedModCount;
|
|
}
|
|
|
|
final int getFence() { // initialize fence and size on first use
|
|
int hi;
|
|
if ((hi = fence) < 0) {
|
|
est = map.size;
|
|
expectedModCount = map.modCount;
|
|
hi = fence = map.table.length;
|
|
}
|
|
return hi;
|
|
}
|
|
|
|
public final long estimateSize() {
|
|
getFence(); // force init
|
|
return (long) est;
|
|
}
|
|
}
|
|
|
|
static final class KeySpliterator<K,V>
|
|
extends IdentityHashMapSpliterator<K,V>
|
|
implements Spliterator<K> {
|
|
KeySpliterator(IdentityHashMap<K,V> map, int origin, int fence, int est,
|
|
int expectedModCount) {
|
|
super(map, origin, fence, est, expectedModCount);
|
|
}
|
|
|
|
public KeySpliterator<K,V> trySplit() {
|
|
int hi = getFence(), lo = index, mid = ((lo + hi) >>> 1) & ~1;
|
|
return (lo >= mid) ? null :
|
|
new KeySpliterator<>(map, lo, index = mid, est >>>= 1,
|
|
expectedModCount);
|
|
}
|
|
|
|
@SuppressWarnings("unchecked")
|
|
public void forEachRemaining(Consumer<? super K> action) {
|
|
if (action == null)
|
|
throw new NullPointerException();
|
|
int i, hi, mc; Object key;
|
|
IdentityHashMap<K,V> m; Object[] a;
|
|
if ((m = map) != null && (a = m.table) != null &&
|
|
(i = index) >= 0 && (index = hi = getFence()) <= a.length) {
|
|
for (; i < hi; i += 2) {
|
|
if ((key = a[i]) != null)
|
|
action.accept((K)unmaskNull(key));
|
|
}
|
|
if (m.modCount == expectedModCount)
|
|
return;
|
|
}
|
|
throw new ConcurrentModificationException();
|
|
}
|
|
|
|
@SuppressWarnings("unchecked")
|
|
public boolean tryAdvance(Consumer<? super K> action) {
|
|
if (action == null)
|
|
throw new NullPointerException();
|
|
Object[] a = map.table;
|
|
int hi = getFence();
|
|
while (index < hi) {
|
|
Object key = a[index];
|
|
index += 2;
|
|
if (key != null) {
|
|
action.accept((K)unmaskNull(key));
|
|
if (map.modCount != expectedModCount)
|
|
throw new ConcurrentModificationException();
|
|
return true;
|
|
}
|
|
}
|
|
return false;
|
|
}
|
|
|
|
public int characteristics() {
|
|
return (fence < 0 || est == map.size ? SIZED : 0) | Spliterator.DISTINCT;
|
|
}
|
|
}
|
|
|
|
static final class ValueSpliterator<K,V>
|
|
extends IdentityHashMapSpliterator<K,V>
|
|
implements Spliterator<V> {
|
|
ValueSpliterator(IdentityHashMap<K,V> m, int origin, int fence, int est,
|
|
int expectedModCount) {
|
|
super(m, origin, fence, est, expectedModCount);
|
|
}
|
|
|
|
public ValueSpliterator<K,V> trySplit() {
|
|
int hi = getFence(), lo = index, mid = ((lo + hi) >>> 1) & ~1;
|
|
return (lo >= mid) ? null :
|
|
new ValueSpliterator<>(map, lo, index = mid, est >>>= 1,
|
|
expectedModCount);
|
|
}
|
|
|
|
public void forEachRemaining(Consumer<? super V> action) {
|
|
if (action == null)
|
|
throw new NullPointerException();
|
|
int i, hi, mc;
|
|
IdentityHashMap<K,V> m; Object[] a;
|
|
if ((m = map) != null && (a = m.table) != null &&
|
|
(i = index) >= 0 && (index = hi = getFence()) <= a.length) {
|
|
for (; i < hi; i += 2) {
|
|
if (a[i] != null) {
|
|
@SuppressWarnings("unchecked") V v = (V)a[i+1];
|
|
action.accept(v);
|
|
}
|
|
}
|
|
if (m.modCount == expectedModCount)
|
|
return;
|
|
}
|
|
throw new ConcurrentModificationException();
|
|
}
|
|
|
|
public boolean tryAdvance(Consumer<? super V> action) {
|
|
if (action == null)
|
|
throw new NullPointerException();
|
|
Object[] a = map.table;
|
|
int hi = getFence();
|
|
while (index < hi) {
|
|
Object key = a[index];
|
|
@SuppressWarnings("unchecked") V v = (V)a[index+1];
|
|
index += 2;
|
|
if (key != null) {
|
|
action.accept(v);
|
|
if (map.modCount != expectedModCount)
|
|
throw new ConcurrentModificationException();
|
|
return true;
|
|
}
|
|
}
|
|
return false;
|
|
}
|
|
|
|
public int characteristics() {
|
|
return (fence < 0 || est == map.size ? SIZED : 0);
|
|
}
|
|
|
|
}
|
|
|
|
static final class EntrySpliterator<K,V>
|
|
extends IdentityHashMapSpliterator<K,V>
|
|
implements Spliterator<Map.Entry<K,V>> {
|
|
EntrySpliterator(IdentityHashMap<K,V> m, int origin, int fence, int est,
|
|
int expectedModCount) {
|
|
super(m, origin, fence, est, expectedModCount);
|
|
}
|
|
|
|
public EntrySpliterator<K,V> trySplit() {
|
|
int hi = getFence(), lo = index, mid = ((lo + hi) >>> 1) & ~1;
|
|
return (lo >= mid) ? null :
|
|
new EntrySpliterator<>(map, lo, index = mid, est >>>= 1,
|
|
expectedModCount);
|
|
}
|
|
|
|
public void forEachRemaining(Consumer<? super Map.Entry<K, V>> action) {
|
|
if (action == null)
|
|
throw new NullPointerException();
|
|
int i, hi, mc;
|
|
IdentityHashMap<K,V> m; Object[] a;
|
|
if ((m = map) != null && (a = m.table) != null &&
|
|
(i = index) >= 0 && (index = hi = getFence()) <= a.length) {
|
|
for (; i < hi; i += 2) {
|
|
Object key = a[i];
|
|
if (key != null) {
|
|
@SuppressWarnings("unchecked") K k =
|
|
(K)unmaskNull(key);
|
|
@SuppressWarnings("unchecked") V v = (V)a[i+1];
|
|
action.accept
|
|
(new AbstractMap.SimpleImmutableEntry<>(k, v));
|
|
|
|
}
|
|
}
|
|
if (m.modCount == expectedModCount)
|
|
return;
|
|
}
|
|
throw new ConcurrentModificationException();
|
|
}
|
|
|
|
public boolean tryAdvance(Consumer<? super Map.Entry<K,V>> action) {
|
|
if (action == null)
|
|
throw new NullPointerException();
|
|
Object[] a = map.table;
|
|
int hi = getFence();
|
|
while (index < hi) {
|
|
Object key = a[index];
|
|
@SuppressWarnings("unchecked") V v = (V)a[index+1];
|
|
index += 2;
|
|
if (key != null) {
|
|
@SuppressWarnings("unchecked") K k =
|
|
(K)unmaskNull(key);
|
|
action.accept
|
|
(new AbstractMap.SimpleImmutableEntry<>(k, v));
|
|
if (map.modCount != expectedModCount)
|
|
throw new ConcurrentModificationException();
|
|
return true;
|
|
}
|
|
}
|
|
return false;
|
|
}
|
|
|
|
public int characteristics() {
|
|
return (fence < 0 || est == map.size ? SIZED : 0) | Spliterator.DISTINCT;
|
|
}
|
|
}
|
|
|
|
}
|