com.google.gwt.thirdparty.guava.common.collect.ComputingConcurrentHashMap Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of guava Show documentation
Show all versions of guava Show documentation
Guava is a suite of core and expanded libraries that include
utility classes, google's collections, io classes, and much
much more.
Guava has only one code dependency - javax.annotation,
per the JSR-305 spec.
The newest version!
/*
* Copyright (C) 2010 The Guava 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
*
* http://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 com.google.common.collect;
import static com.google.common.base.Preconditions.checkNotNull;
import static com.google.common.base.Preconditions.checkState;
import com.google.common.base.Equivalence;
import com.google.common.base.Function;
import com.google.common.collect.MapMaker.RemovalCause;
import com.google.common.collect.MapMaker.RemovalListener;
import java.io.IOException;
import java.io.ObjectInputStream;
import java.io.ObjectOutputStream;
import java.lang.ref.ReferenceQueue;
import java.util.concurrent.ConcurrentMap;
import java.util.concurrent.ExecutionException;
import java.util.concurrent.atomic.AtomicReferenceArray;
import javax.annotation.Nullable;
import javax.annotation.concurrent.GuardedBy;
/**
* Adds computing functionality to {@link MapMakerInternalMap}.
*
* @author Bob Lee
* @author Charles Fry
*/
class ComputingConcurrentHashMap extends MapMakerInternalMap {
final Function super K, ? extends V> computingFunction;
/**
* Creates a new, empty map with the specified strategy, initial capacity, load factor and
* concurrency level.
*/
ComputingConcurrentHashMap(MapMaker builder,
Function super K, ? extends V> computingFunction) {
super(builder);
this.computingFunction = checkNotNull(computingFunction);
}
@Override
Segment createSegment(int initialCapacity, int maxSegmentSize) {
return new ComputingSegment(this, initialCapacity, maxSegmentSize);
}
@Override
ComputingSegment segmentFor(int hash) {
return (ComputingSegment) super.segmentFor(hash);
}
V getOrCompute(K key) throws ExecutionException {
int hash = hash(checkNotNull(key));
return segmentFor(hash).getOrCompute(key, hash, computingFunction);
}
@SuppressWarnings("serial") // This class is never serialized.
static final class ComputingSegment extends Segment {
ComputingSegment(MapMakerInternalMap map, int initialCapacity, int maxSegmentSize) {
super(map, initialCapacity, maxSegmentSize);
}
V getOrCompute(K key, int hash, Function super K, ? extends V> computingFunction)
throws ExecutionException {
try {
outer: while (true) {
// don't call getLiveEntry, which would ignore computing values
ReferenceEntry e = getEntry(key, hash);
if (e != null) {
V value = getLiveValue(e);
if (value != null) {
recordRead(e);
return value;
}
}
// at this point e is either null, computing, or expired;
// avoid locking if it's already computing
if (e == null || !e.getValueReference().isComputingReference()) {
boolean createNewEntry = true;
ComputingValueReference computingValueReference = null;
lock();
try {
preWriteCleanup();
int newCount = this.count - 1;
AtomicReferenceArray> table = this.table;
int index = hash & (table.length() - 1);
ReferenceEntry first = table.get(index);
for (e = first; e != null; e = e.getNext()) {
K entryKey = e.getKey();
if (e.getHash() == hash && entryKey != null
&& map.keyEquivalence.equivalent(key, entryKey)) {
ValueReference valueReference = e.getValueReference();
if (valueReference.isComputingReference()) {
createNewEntry = false;
} else {
V value = e.getValueReference().get();
if (value == null) {
enqueueNotification(entryKey, hash, value, RemovalCause.COLLECTED);
} else if (map.expires() && map.isExpired(e)) {
// This is a duplicate check, as preWriteCleanup already purged expired
// entries, but let's accomodate an incorrect expiration queue.
enqueueNotification(entryKey, hash, value, RemovalCause.EXPIRED);
} else {
recordLockedRead(e);
return value;
}
// immediately reuse invalid entries
evictionQueue.remove(e);
expirationQueue.remove(e);
this.count = newCount; // write-volatile
}
break;
}
}
if (createNewEntry) {
computingValueReference = new ComputingValueReference(computingFunction);
if (e == null) {
e = newEntry(key, hash, first);
e.setValueReference(computingValueReference);
table.set(index, e);
} else {
e.setValueReference(computingValueReference);
}
}
} finally {
unlock();
postWriteCleanup();
}
if (createNewEntry) {
// This thread solely created the entry.
return compute(key, hash, e, computingValueReference);
}
}
// The entry already exists. Wait for the computation.
checkState(!Thread.holdsLock(e), "Recursive computation");
// don't consider expiration as we're concurrent with computation
V value = e.getValueReference().waitForValue();
if (value != null) {
recordRead(e);
return value;
}
// else computing thread will clearValue
continue outer;
}
} finally {
postReadCleanup();
}
}
V compute(K key, int hash, ReferenceEntry e,
ComputingValueReference computingValueReference)
throws ExecutionException {
V value = null;
long start = System.nanoTime();
long end = 0;
try {
// Synchronizes on the entry to allow failing fast when a recursive computation is
// detected. This is not fool-proof since the entry may be copied when the segment
// is written to.
synchronized (e) {
value = computingValueReference.compute(key, hash);
end = System.nanoTime();
}
if (value != null) {
// putIfAbsent
V oldValue = put(key, hash, value, true);
if (oldValue != null) {
// the computed value was already clobbered
enqueueNotification(key, hash, value, RemovalCause.REPLACED);
}
}
return value;
} finally {
if (end == 0) {
end = System.nanoTime();
}
if (value == null) {
clearValue(key, hash, computingValueReference);
}
}
}
}
/**
* Used to provide computation exceptions to other threads.
*/
private static final class ComputationExceptionReference implements ValueReference {
final Throwable t;
ComputationExceptionReference(Throwable t) {
this.t = t;
}
@Override
public V get() {
return null;
}
@Override
public ReferenceEntry getEntry() {
return null;
}
@Override
public ValueReference copyFor(
ReferenceQueue queue, V value, ReferenceEntry entry) {
return this;
}
@Override
public boolean isComputingReference() {
return false;
}
@Override
public V waitForValue() throws ExecutionException {
throw new ExecutionException(t);
}
@Override
public void clear(ValueReference newValue) {}
}
/**
* Used to provide computation result to other threads.
*/
private static final class ComputedReference implements ValueReference {
final V value;
ComputedReference(@Nullable V value) {
this.value = value;
}
@Override
public V get() {
return value;
}
@Override
public ReferenceEntry getEntry() {
return null;
}
@Override
public ValueReference copyFor(
ReferenceQueue queue, V value, ReferenceEntry entry) {
return this;
}
@Override
public boolean isComputingReference() {
return false;
}
@Override
public V waitForValue() {
return get();
}
@Override
public void clear(ValueReference newValue) {}
}
private static final class ComputingValueReference implements ValueReference {
final Function super K, ? extends V> computingFunction;
@GuardedBy("ComputingValueReference.this") // writes
volatile ValueReference computedReference = unset();
public ComputingValueReference(Function super K, ? extends V> computingFunction) {
this.computingFunction = computingFunction;
}
@Override
public V get() {
// All computation lookups go through waitForValue. This method thus is
// only used by put, to whom we always want to appear absent.
return null;
}
@Override
public ReferenceEntry getEntry() {
return null;
}
@Override
public ValueReference copyFor(
ReferenceQueue queue, @Nullable V value, ReferenceEntry entry) {
return this;
}
@Override
public boolean isComputingReference() {
return true;
}
/**
* Waits for a computation to complete. Returns the result of the computation.
*/
@Override
public V waitForValue() throws ExecutionException {
if (computedReference == UNSET) {
boolean interrupted = false;
try {
synchronized (this) {
while (computedReference == UNSET) {
try {
wait();
} catch (InterruptedException ie) {
interrupted = true;
}
}
}
} finally {
if (interrupted) {
Thread.currentThread().interrupt();
}
}
}
return computedReference.waitForValue();
}
@Override
public void clear(ValueReference newValue) {
// The pending computation was clobbered by a manual write. Unblock all
// pending gets, and have them return the new value.
setValueReference(newValue);
// TODO(fry): could also cancel computation if we had a thread handle
}
V compute(K key, int hash) throws ExecutionException {
V value;
try {
value = computingFunction.apply(key);
} catch (Throwable t) {
setValueReference(new ComputationExceptionReference(t));
throw new ExecutionException(t);
}
setValueReference(new ComputedReference(value));
return value;
}
void setValueReference(ValueReference valueReference) {
synchronized (this) {
if (computedReference == UNSET) {
computedReference = valueReference;
notifyAll();
}
}
}
}
// Serialization Support
private static final long serialVersionUID = 4;
@Override
Object writeReplace() {
return new ComputingSerializationProxy(keyStrength, valueStrength, keyEquivalence,
valueEquivalence, expireAfterWriteNanos, expireAfterAccessNanos, maximumSize,
concurrencyLevel, removalListener, this, computingFunction);
}
static final class ComputingSerializationProxy extends AbstractSerializationProxy {
final Function super K, ? extends V> computingFunction;
ComputingSerializationProxy(Strength keyStrength, Strength valueStrength,
Equivalence