org.springframework.util.ConcurrentReferenceHashMap Maven / Gradle / Ivy
/*
* Copyright 2002-2019 the original author or authors.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* https://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.springframework.util;
import java.lang.ref.ReferenceQueue;
import java.lang.ref.SoftReference;
import java.lang.ref.WeakReference;
import java.lang.reflect.Array;
import java.util.AbstractMap;
import java.util.AbstractSet;
import java.util.Collections;
import java.util.EnumSet;
import java.util.HashSet;
import java.util.Iterator;
import java.util.Map;
import java.util.NoSuchElementException;
import java.util.Set;
import java.util.concurrent.ConcurrentHashMap;
import java.util.concurrent.ConcurrentMap;
import java.util.concurrent.atomic.AtomicInteger;
import java.util.concurrent.locks.ReentrantLock;
import org.springframework.lang.Nullable;
/**
* A {@link ConcurrentHashMap} that uses {@link ReferenceType#SOFT soft} or
* {@linkplain ReferenceType#WEAK weak} references for both {@code keys} and {@code values}.
*
* This class can be used as an alternative to
* {@code Collections.synchronizedMap(new WeakHashMap>())} in order to
* support better performance when accessed concurrently. This implementation follows the
* same design constraints as {@link ConcurrentHashMap} with the exception that
* {@code null} values and {@code null} keys are supported.
*
* NOTE: The use of references means that there is no guarantee that items
* placed into the map will be subsequently available. The garbage collector may discard
* references at any time, so it may appear that an unknown thread is silently removing
* entries.
*
*
If not explicitly specified, this implementation will use
* {@linkplain SoftReference soft entry references}.
*
* @author Phillip Webb
* @author Juergen Hoeller
* @since 3.2
* @param the key type
* @param the value type
*/
public class ConcurrentReferenceHashMap extends AbstractMap implements ConcurrentMap {
private static final int DEFAULT_INITIAL_CAPACITY = 16;
private static final float DEFAULT_LOAD_FACTOR = 0.75f;
private static final int DEFAULT_CONCURRENCY_LEVEL = 16;
private static final ReferenceType DEFAULT_REFERENCE_TYPE = ReferenceType.SOFT;
private static final int MAXIMUM_CONCURRENCY_LEVEL = 1 << 16;
private static final int MAXIMUM_SEGMENT_SIZE = 1 << 30;
/**
* Array of segments indexed using the high order bits from the hash.
*/
private final Segment[] segments;
/**
* When the average number of references per table exceeds this value resize will be attempted.
*/
private final float loadFactor;
/**
* The reference type: SOFT or WEAK.
*/
private final ReferenceType referenceType;
/**
* The shift value used to calculate the size of the segments array and an index from the hash.
*/
private final int shift;
/**
* Late binding entry set.
*/
@Nullable
private volatile Set> entrySet;
/**
* Create a new {@code ConcurrentReferenceHashMap} instance.
*/
public ConcurrentReferenceHashMap() {
this(DEFAULT_INITIAL_CAPACITY, DEFAULT_LOAD_FACTOR, DEFAULT_CONCURRENCY_LEVEL, DEFAULT_REFERENCE_TYPE);
}
/**
* Create a new {@code ConcurrentReferenceHashMap} instance.
* @param initialCapacity the initial capacity of the map
*/
public ConcurrentReferenceHashMap(int initialCapacity) {
this(initialCapacity, DEFAULT_LOAD_FACTOR, DEFAULT_CONCURRENCY_LEVEL, DEFAULT_REFERENCE_TYPE);
}
/**
* Create a new {@code ConcurrentReferenceHashMap} instance.
* @param initialCapacity the initial capacity of the map
* @param loadFactor the load factor. When the average number of references per table
* exceeds this value resize will be attempted
*/
public ConcurrentReferenceHashMap(int initialCapacity, float loadFactor) {
this(initialCapacity, loadFactor, DEFAULT_CONCURRENCY_LEVEL, DEFAULT_REFERENCE_TYPE);
}
/**
* Create a new {@code ConcurrentReferenceHashMap} instance.
* @param initialCapacity the initial capacity of the map
* @param concurrencyLevel the expected number of threads that will concurrently
* write to the map
*/
public ConcurrentReferenceHashMap(int initialCapacity, int concurrencyLevel) {
this(initialCapacity, DEFAULT_LOAD_FACTOR, concurrencyLevel, DEFAULT_REFERENCE_TYPE);
}
/**
* Create a new {@code ConcurrentReferenceHashMap} instance.
* @param initialCapacity the initial capacity of the map
* @param referenceType the reference type used for entries (soft or weak)
*/
public ConcurrentReferenceHashMap(int initialCapacity, ReferenceType referenceType) {
this(initialCapacity, DEFAULT_LOAD_FACTOR, DEFAULT_CONCURRENCY_LEVEL, referenceType);
}
/**
* Create a new {@code ConcurrentReferenceHashMap} instance.
* @param initialCapacity the initial capacity of the map
* @param loadFactor the load factor. When the average number of references per
* table exceeds this value, resize will be attempted.
* @param concurrencyLevel the expected number of threads that will concurrently
* write to the map
*/
public ConcurrentReferenceHashMap(int initialCapacity, float loadFactor, int concurrencyLevel) {
this(initialCapacity, loadFactor, concurrencyLevel, DEFAULT_REFERENCE_TYPE);
}
/**
* Create a new {@code ConcurrentReferenceHashMap} instance.
* @param initialCapacity the initial capacity of the map
* @param loadFactor the load factor. When the average number of references per
* table exceeds this value, resize will be attempted.
* @param concurrencyLevel the expected number of threads that will concurrently
* write to the map
* @param referenceType the reference type used for entries (soft or weak)
*/
@SuppressWarnings("unchecked")
public ConcurrentReferenceHashMap(
int initialCapacity, float loadFactor, int concurrencyLevel, ReferenceType referenceType) {
Assert.isTrue(initialCapacity >= 0, "Initial capacity must not be negative");
Assert.isTrue(loadFactor > 0f, "Load factor must be positive");
Assert.isTrue(concurrencyLevel > 0, "Concurrency level must be positive");
Assert.notNull(referenceType, "Reference type must not be null");
this.loadFactor = loadFactor;
this.shift = calculateShift(concurrencyLevel, MAXIMUM_CONCURRENCY_LEVEL);
int size = 1 << this.shift;
this.referenceType = referenceType;
int roundedUpSegmentCapacity = (int) ((initialCapacity + size - 1L) / size);
int initialSize = 1 << calculateShift(roundedUpSegmentCapacity, MAXIMUM_SEGMENT_SIZE);
Segment[] segments = (Segment[]) Array.newInstance(Segment.class, size);
int resizeThreshold = (int) (initialSize * getLoadFactor());
for (int i = 0; i < segments.length; i++) {
segments[i] = new Segment(initialSize, resizeThreshold);
}
this.segments = segments;
}
protected final float getLoadFactor() {
return this.loadFactor;
}
protected final int getSegmentsSize() {
return this.segments.length;
}
protected final Segment getSegment(int index) {
return this.segments[index];
}
/**
* Factory method that returns the {@link ReferenceManager}.
* This method will be called once for each {@link Segment}.
* @return a new reference manager
*/
protected ReferenceManager createReferenceManager() {
return new ReferenceManager();
}
/**
* Get the hash for a given object, apply an additional hash function to reduce
* collisions. This implementation uses the same Wang/Jenkins algorithm as
* {@link ConcurrentHashMap}. Subclasses can override to provide alternative hashing.
* @param o the object to hash (may be null)
* @return the resulting hash code
*/
protected int getHash(@Nullable Object o) {
int hash = (o != null ? o.hashCode() : 0);
hash += (hash << 15) ^ 0xffffcd7d;
hash ^= (hash >>> 10);
hash += (hash << 3);
hash ^= (hash >>> 6);
hash += (hash << 2) + (hash << 14);
hash ^= (hash >>> 16);
return hash;
}
@Override
@Nullable
public V get(@Nullable Object key) {
Reference ref = getReference(key, Restructure.WHEN_NECESSARY);
Entry entry = (ref != null ? ref.get() : null);
return (entry != null ? entry.getValue() : null);
}
@Override
@Nullable
public V getOrDefault(@Nullable Object key, @Nullable V defaultValue) {
Reference ref = getReference(key, Restructure.WHEN_NECESSARY);
Entry entry = (ref != null ? ref.get() : null);
return (entry != null ? entry.getValue() : defaultValue);
}
@Override
public boolean containsKey(@Nullable Object key) {
Reference ref = getReference(key, Restructure.WHEN_NECESSARY);
Entry entry = (ref != null ? ref.get() : null);
return (entry != null && ObjectUtils.nullSafeEquals(entry.getKey(), key));
}
/**
* Return a {@link Reference} to the {@link Entry} for the specified {@code key},
* or {@code null} if not found.
* @param key the key (can be {@code null})
* @param restructure types of restructure allowed during this call
* @return the reference, or {@code null} if not found
*/
@Nullable
protected final Reference getReference(@Nullable Object key, Restructure restructure) {
int hash = getHash(key);
return getSegmentForHash(hash).getReference(key, hash, restructure);
}
@Override
@Nullable
public V put(@Nullable K key, @Nullable V value) {
return put(key, value, true);
}
@Override
@Nullable
public V putIfAbsent(@Nullable K key, @Nullable V value) {
return put(key, value, false);
}
@Nullable
private V put(@Nullable final K key, @Nullable final V value, final boolean overwriteExisting) {
return doTask(key, new Task(TaskOption.RESTRUCTURE_BEFORE, TaskOption.RESIZE) {
@Override
@Nullable
protected V execute(@Nullable Reference ref, @Nullable Entry entry, @Nullable Entries entries) {
if (entry != null) {
V oldValue = entry.getValue();
if (overwriteExisting) {
entry.setValue(value);
}
return oldValue;
}
Assert.state(entries != null, "No entries segment");
entries.add(value);
return null;
}
});
}
@Override
@Nullable
public V remove(Object key) {
return doTask(key, new Task(TaskOption.RESTRUCTURE_AFTER, TaskOption.SKIP_IF_EMPTY) {
@Override
@Nullable
protected V execute(@Nullable Reference ref, @Nullable Entry entry) {
if (entry != null) {
if (ref != null) {
ref.release();
}
return entry.value;
}
return null;
}
});
}
@Override
public boolean remove(Object key, final Object value) {
Boolean result = doTask(key, new Task(TaskOption.RESTRUCTURE_AFTER, TaskOption.SKIP_IF_EMPTY) {
@Override
protected Boolean execute(@Nullable Reference ref, @Nullable Entry entry) {
if (entry != null && ObjectUtils.nullSafeEquals(entry.getValue(), value)) {
if (ref != null) {
ref.release();
}
return true;
}
return false;
}
});
return (result == Boolean.TRUE);
}
@Override
public boolean replace(K key, final V oldValue, final V newValue) {
Boolean result = doTask(key, new Task(TaskOption.RESTRUCTURE_BEFORE, TaskOption.SKIP_IF_EMPTY) {
@Override
protected Boolean execute(@Nullable Reference ref, @Nullable Entry entry) {
if (entry != null && ObjectUtils.nullSafeEquals(entry.getValue(), oldValue)) {
entry.setValue(newValue);
return true;
}
return false;
}
});
return (result == Boolean.TRUE);
}
@Override
@Nullable
public V replace(K key, final V value) {
return doTask(key, new Task(TaskOption.RESTRUCTURE_BEFORE, TaskOption.SKIP_IF_EMPTY) {
@Override
@Nullable
protected V execute(@Nullable Reference ref, @Nullable Entry entry) {
if (entry != null) {
V oldValue = entry.getValue();
entry.setValue(value);
return oldValue;
}
return null;
}
});
}
@Override
public void clear() {
for (Segment segment : this.segments) {
segment.clear();
}
}
/**
* Remove any entries that have been garbage collected and are no longer referenced.
* Under normal circumstances garbage collected entries are automatically purged as
* items are added or removed from the Map. This method can be used to force a purge,
* and is useful when the Map is read frequently but updated less often.
*/
public void purgeUnreferencedEntries() {
for (Segment segment : this.segments) {
segment.restructureIfNecessary(false);
}
}
@Override
public int size() {
int size = 0;
for (Segment segment : this.segments) {
size += segment.getCount();
}
return size;
}
@Override
public boolean isEmpty() {
for (Segment segment : this.segments) {
if (segment.getCount() > 0) {
return false;
}
}
return true;
}
@Override
public Set> entrySet() {
Set> entrySet = this.entrySet;
if (entrySet == null) {
entrySet = new EntrySet();
this.entrySet = entrySet;
}
return entrySet;
}
@Nullable
private T doTask(@Nullable Object key, Task task) {
int hash = getHash(key);
return getSegmentForHash(hash).doTask(hash, key, task);
}
private Segment getSegmentForHash(int hash) {
return this.segments[(hash >>> (32 - this.shift)) & (this.segments.length - 1)];
}
/**
* Calculate a shift value that can be used to create a power-of-two value between
* the specified maximum and minimum values.
* @param minimumValue the minimum value
* @param maximumValue the maximum value
* @return the calculated shift (use {@code 1 << shift} to obtain a value)
*/
protected static int calculateShift(int minimumValue, int maximumValue) {
int shift = 0;
int value = 1;
while (value < minimumValue && value < maximumValue) {
value <<= 1;
shift++;
}
return shift;
}
/**
* Various reference types supported by this map.
*/
public enum ReferenceType {
/** Use {@link SoftReference SoftReferences}. */
SOFT,
/** Use {@link WeakReference WeakReferences}. */
WEAK
}
/**
* A single segment used to divide the map to allow better concurrent performance.
*/
@SuppressWarnings("serial")
protected final class Segment extends ReentrantLock {
private final ReferenceManager referenceManager;
private final int initialSize;
/**
* Array of references indexed using the low order bits from the hash.
* This property should only be set along with {@code resizeThreshold}.
*/
private volatile Reference[] references;
/**
* The total number of references contained in this segment. This includes chained
* references and references that have been garbage collected but not purged.
*/
private final AtomicInteger count = new AtomicInteger(0);
/**
* The threshold when resizing of the references should occur. When {@code count}
* exceeds this value references will be resized.
*/
private int resizeThreshold;
public Segment(int initialSize, int resizeThreshold) {
this.referenceManager = createReferenceManager();
this.initialSize = initialSize;
this.references = createReferenceArray(initialSize);
this.resizeThreshold = resizeThreshold;
}
@Nullable
public Reference getReference(@Nullable Object key, int hash, Restructure restructure) {
if (restructure == Restructure.WHEN_NECESSARY) {
restructureIfNecessary(false);
}
if (this.count.get() == 0) {
return null;
}
// Use a local copy to protect against other threads writing
Reference[] references = this.references;
int index = getIndex(hash, references);
Reference head = references[index];
return findInChain(head, key, hash);
}
/**
* Apply an update operation to this segment.
* The segment will be locked during the update.
* @param hash the hash of the key
* @param key the key
* @param task the update operation
* @return the result of the operation
*/
@Nullable
public T doTask(final int hash, @Nullable final Object key, final Task task) {
boolean resize = task.hasOption(TaskOption.RESIZE);
if (task.hasOption(TaskOption.RESTRUCTURE_BEFORE)) {
restructureIfNecessary(resize);
}
if (task.hasOption(TaskOption.SKIP_IF_EMPTY) && this.count.get() == 0) {
return task.execute(null, null, null);
}
lock();
try {
final int index = getIndex(hash, this.references);
final Reference head = this.references[index];
Reference ref = findInChain(head, key, hash);
Entry entry = (ref != null ? ref.get() : null);
Entries entries = value -> {
@SuppressWarnings("unchecked")
Entry newEntry = new Entry<>((K) key, value);
Reference newReference = Segment.this.referenceManager.createReference(newEntry, hash, head);
Segment.this.references[index] = newReference;
Segment.this.count.incrementAndGet();
};
return task.execute(ref, entry, entries);
}
finally {
unlock();
if (task.hasOption(TaskOption.RESTRUCTURE_AFTER)) {
restructureIfNecessary(resize);
}
}
}
/**
* Clear all items from this segment.
*/
public void clear() {
if (this.count.get() == 0) {
return;
}
lock();
try {
this.references = createReferenceArray(this.initialSize);
this.resizeThreshold = (int) (this.references.length * getLoadFactor());
this.count.set(0);
}
finally {
unlock();
}
}
/**
* Restructure the underlying data structure when it becomes necessary. This
* method can increase the size of the references table as well as purge any
* references that have been garbage collected.
* @param allowResize if resizing is permitted
*/
protected final void restructureIfNecessary(boolean allowResize) {
int currCount = this.count.get();
boolean needsResize = allowResize && (currCount > 0 && currCount >= this.resizeThreshold);
Reference ref = this.referenceManager.pollForPurge();
if (ref != null || (needsResize)) {
restructure(allowResize, ref);
}
}
private void restructure(boolean allowResize, @Nullable Reference ref) {
boolean needsResize;
lock();
try {
int countAfterRestructure = this.count.get();
Set> toPurge = Collections.emptySet();
if (ref != null) {
toPurge = new HashSet<>();
while (ref != null) {
toPurge.add(ref);
ref = this.referenceManager.pollForPurge();
}
}
countAfterRestructure -= toPurge.size();
// Recalculate taking into account count inside lock and items that
// will be purged
needsResize = (countAfterRestructure > 0 && countAfterRestructure >= this.resizeThreshold);
boolean resizing = false;
int restructureSize = this.references.length;
if (allowResize && needsResize && restructureSize < MAXIMUM_SEGMENT_SIZE) {
restructureSize <<= 1;
resizing = true;
}
// Either create a new table or reuse the existing one
Reference[] restructured =
(resizing ? createReferenceArray(restructureSize) : this.references);
// Restructure
for (int i = 0; i < this.references.length; i++) {
ref = this.references[i];
if (!resizing) {
restructured[i] = null;
}
while (ref != null) {
if (!toPurge.contains(ref)) {
Entry entry = ref.get();
if (entry != null) {
int index = getIndex(ref.getHash(), restructured);
restructured[index] = this.referenceManager.createReference(
entry, ref.getHash(), restructured[index]);
}
}
ref = ref.getNext();
}
}
// Replace volatile members
if (resizing) {
this.references = restructured;
this.resizeThreshold = (int) (this.references.length * getLoadFactor());
}
this.count.set(Math.max(countAfterRestructure, 0));
}
finally {
unlock();
}
}
@Nullable
private Reference findInChain(Reference ref, @Nullable Object key, int hash) {
Reference currRef = ref;
while (currRef != null) {
if (currRef.getHash() == hash) {
Entry entry = currRef.get();
if (entry != null) {
K entryKey = entry.getKey();
if (ObjectUtils.nullSafeEquals(entryKey, key)) {
return currRef;
}
}
}
currRef = currRef.getNext();
}
return null;
}
@SuppressWarnings({"rawtypes", "unchecked"})
private Reference[] createReferenceArray(int size) {
return new Reference[size];
}
private int getIndex(int hash, Reference[] references) {
return (hash & (references.length - 1));
}
/**
* Return the size of the current references array.
*/
public final int getSize() {
return this.references.length;
}
/**
* Return the total number of references in this segment.
*/
public final int getCount() {
return this.count.get();
}
}
/**
* A reference to an {@link Entry} contained in the map. Implementations are usually
* wrappers around specific Java reference implementations (e.g., {@link SoftReference}).
* @param the key type
* @param the value type
*/
protected interface Reference {
/**
* Return the referenced entry, or {@code null} if the entry is no longer available.
*/
@Nullable
Entry get();
/**
* Return the hash for the reference.
*/
int getHash();
/**
* Return the next reference in the chain, or {@code null} if none.
*/
@Nullable
Reference getNext();
/**
* Release this entry and ensure that it will be returned from
* {@code ReferenceManager#pollForPurge()}.
*/
void release();
}
/**
* A single map entry.
* @param the key type
* @param the value type
*/
protected static final class Entry implements Map.Entry {
@Nullable
private final K key;
@Nullable
private volatile V value;
public Entry(@Nullable K key, @Nullable V value) {
this.key = key;
this.value = value;
}
@Override
@Nullable
public K getKey() {
return this.key;
}
@Override
@Nullable
public V getValue() {
return this.value;
}
@Override
@Nullable
public V setValue(@Nullable V value) {
V previous = this.value;
this.value = value;
return previous;
}
@Override
public String toString() {
return (this.key + "=" + this.value);
}
@Override
@SuppressWarnings("rawtypes")
public final boolean equals(@Nullable Object other) {
if (this == other) {
return true;
}
if (!(other instanceof Map.Entry)) {
return false;
}
Map.Entry otherEntry = (Map.Entry) other;
return (ObjectUtils.nullSafeEquals(getKey(), otherEntry.getKey()) &&
ObjectUtils.nullSafeEquals(getValue(), otherEntry.getValue()));
}
@Override
public final int hashCode() {
return (ObjectUtils.nullSafeHashCode(this.key) ^ ObjectUtils.nullSafeHashCode(this.value));
}
}
/**
* A task that can be {@link Segment#doTask run} against a {@link Segment}.
*/
private abstract class Task {
private final EnumSet options;
public Task(TaskOption... options) {
this.options = (options.length == 0 ? EnumSet.noneOf(TaskOption.class) : EnumSet.of(options[0], options));
}
public boolean hasOption(TaskOption option) {
return this.options.contains(option);
}
/**
* Execute the task.
* @param ref the found reference (or {@code null})
* @param entry the found entry (or {@code null})
* @param entries access to the underlying entries
* @return the result of the task
* @see #execute(Reference, Entry)
*/
@Nullable
protected T execute(@Nullable Reference ref, @Nullable Entry entry, @Nullable Entries entries) {
return execute(ref, entry);
}
/**
* Convenience method that can be used for tasks that do not need access to {@link Entries}.
* @param ref the found reference (or {@code null})
* @param entry the found entry (or {@code null})
* @return the result of the task
* @see #execute(Reference, Entry, Entries)
*/
@Nullable
protected T execute(@Nullable Reference ref, @Nullable Entry entry) {
return null;
}
}
/**
* Various options supported by a {@code Task}.
*/
private enum TaskOption {
RESTRUCTURE_BEFORE, RESTRUCTURE_AFTER, SKIP_IF_EMPTY, RESIZE
}
/**
* Allows a task access to {@link ConcurrentReferenceHashMap.Segment} entries.
*/
private interface Entries {
/**
* Add a new entry with the specified value.
* @param value the value to add
*/
void add(@Nullable V value);
}
/**
* Internal entry-set implementation.
*/
private class EntrySet extends AbstractSet> {
@Override
public Iterator> iterator() {
return new EntryIterator();
}
@Override
public boolean contains(@Nullable Object o) {
if (o instanceof Map.Entry) {
Map.Entry entry = (Map.Entry) o;
Reference ref = ConcurrentReferenceHashMap.this.getReference(entry.getKey(), Restructure.NEVER);
Entry otherEntry = (ref != null ? ref.get() : null);
if (otherEntry != null) {
return ObjectUtils.nullSafeEquals(otherEntry.getValue(), otherEntry.getValue());
}
}
return false;
}
@Override
public boolean remove(Object o) {
if (o instanceof Map.Entry) {
Map.Entry entry = (Map.Entry) o;
return ConcurrentReferenceHashMap.this.remove(entry.getKey(), entry.getValue());
}
return false;
}
@Override
public int size() {
return ConcurrentReferenceHashMap.this.size();
}
@Override
public void clear() {
ConcurrentReferenceHashMap.this.clear();
}
}
/**
* Internal entry iterator implementation.
*/
private class EntryIterator implements Iterator> {
private int segmentIndex;
private int referenceIndex;
@Nullable
private Reference[] references;
@Nullable
private Reference reference;
@Nullable
private Entry next;
@Nullable
private Entry last;
public EntryIterator() {
moveToNextSegment();
}
@Override
public boolean hasNext() {
getNextIfNecessary();
return (this.next != null);
}
@Override
public Entry next() {
getNextIfNecessary();
if (this.next == null) {
throw new NoSuchElementException();
}
this.last = this.next;
this.next = null;
return this.last;
}
private void getNextIfNecessary() {
while (this.next == null) {
moveToNextReference();
if (this.reference == null) {
return;
}
this.next = this.reference.get();
}
}
private void moveToNextReference() {
if (this.reference != null) {
this.reference = this.reference.getNext();
}
while (this.reference == null && this.references != null) {
if (this.referenceIndex >= this.references.length) {
moveToNextSegment();
this.referenceIndex = 0;
}
else {
this.reference = this.references[this.referenceIndex];
this.referenceIndex++;
}
}
}
private void moveToNextSegment() {
this.reference = null;
this.references = null;
if (this.segmentIndex < ConcurrentReferenceHashMap.this.segments.length) {
this.references = ConcurrentReferenceHashMap.this.segments[this.segmentIndex].references;
this.segmentIndex++;
}
}
@Override
public void remove() {
Assert.state(this.last != null, "No element to remove");
ConcurrentReferenceHashMap.this.remove(this.last.getKey());
}
}
/**
* The types of restructuring that can be performed.
*/
protected enum Restructure {
WHEN_NECESSARY, NEVER
}
/**
* Strategy class used to manage {@link Reference References}.
* This class can be overridden if alternative reference types need to be supported.
*/
protected class ReferenceManager {
private final ReferenceQueue> queue = new ReferenceQueue<>();
/**
* Factory method used to create a new {@link Reference}.
* @param entry the entry contained in the reference
* @param hash the hash
* @param next the next reference in the chain, or {@code null} if none
* @return a new {@link Reference}
*/
public Reference createReference(Entry entry, int hash, @Nullable Reference next) {
if (ConcurrentReferenceHashMap.this.referenceType == ReferenceType.WEAK) {
return new WeakEntryReference<>(entry, hash, next, this.queue);
}
return new SoftEntryReference<>(entry, hash, next, this.queue);
}
/**
* Return any reference that has been garbage collected and can be purged from the
* underlying structure or {@code null} if no references need purging. This
* method must be thread safe and ideally should not block when returning
* {@code null}. References should be returned once and only once.
* @return a reference to purge or {@code null}
*/
@SuppressWarnings("unchecked")
@Nullable
public Reference pollForPurge() {
return (Reference) this.queue.poll();
}
}
/**
* Internal {@link Reference} implementation for {@link SoftReference SoftReferences}.
*/
private static final class SoftEntryReference extends SoftReference> implements Reference {
private final int hash;
@Nullable
private final Reference nextReference;
public SoftEntryReference(Entry entry, int hash, @Nullable Reference next,
ReferenceQueue> queue) {
super(entry, queue);
this.hash = hash;
this.nextReference = next;
}
@Override
public int getHash() {
return this.hash;
}
@Override
@Nullable
public Reference getNext() {
return this.nextReference;
}
@Override
public void release() {
enqueue();
clear();
}
}
/**
* Internal {@link Reference} implementation for {@link WeakReference WeakReferences}.
*/
private static final class WeakEntryReference extends WeakReference> implements Reference {
private final int hash;
@Nullable
private final Reference nextReference;
public WeakEntryReference(Entry entry, int hash, @Nullable Reference next,
ReferenceQueue> queue) {
super(entry, queue);
this.hash = hash;
this.nextReference = next;
}
@Override
public int getHash() {
return this.hash;
}
@Override
@Nullable
public Reference getNext() {
return this.nextReference;
}
@Override
public void release() {
enqueue();
clear();
}
}
}