org.hibernate.stat.internal.StatsNamedContainer Maven / Gradle / Ivy
/*
* Hibernate, Relational Persistence for Idiomatic Java
*
* License: GNU Lesser General Public License (LGPL), version 2.1 or later.
* See the lgpl.txt file in the root directory or .
*/
package org.hibernate.stat.internal;
import java.util.concurrent.ConcurrentHashMap;
import java.util.concurrent.ConcurrentMap;
import java.util.function.Function;
import org.hibernate.internal.util.collections.BoundedConcurrentHashMap;
/**
* Decorates a ConcurrentHashMap implementation to make sure the methods are being
* used correctly for the purpose of Hibernate's statistics.
* In particular, we do like the semantics of ConcurrentHashMap#computeIfAbsent
* but not how it performs.
* See also:
* - http://jsr166-concurrency.10961.n7.nabble.com/Re-ConcurrentHashMap-computeIfAbsent-td11687.html
* - https://hibernate.atlassian.net/browse/HHH-13527
*
* @author Sanne Grinovero
*/
final class StatsNamedContainer {
private final ConcurrentMap map;
private final static Object NULL_TOKEN = new Object();
/**
* Creates a bounded container - based on BoundedConcurrentHashMap
*/
StatsNamedContainer(int capacity, int concurrencyLevel) {
this.map = new BoundedConcurrentHashMap( capacity, concurrencyLevel, BoundedConcurrentHashMap.Eviction.LRU );
}
/**
* Creates an unbounded container - based on ConcurrentHashMap
*/
StatsNamedContainer() {
this.map = new ConcurrentHashMap<>( );
}
public void clear() {
map.clear();
}
/**
* This method is inherently racy and expensive. Only use on non-hot paths, and
* only to get a recent snapshot.
* @return all keys in string form.
*/
public String[] keysAsArray() {
return map.keySet().toArray( new String[0] );
}
/**
* Similar semantics as you'd get by invoking {@link java.util.concurrent.ConcurrentMap#computeIfAbsent(Object, Function)},
* but we check for the key existence first.
* This is technically a redundant check, but it has been shown to perform better when the key existing is very likely,
* as in our case.
* Most notably, the ConcurrentHashMap implementation might block other accesses for the sake of making
* sure the function is invoked at most once: we don't need this guarantee, and prefer to reduce risk of blocking.
*/
public V getOrCompute(final String key, final Function function) {
final Object v1 = map.get( key );
if ( v1 != null ) {
if ( v1 == NULL_TOKEN ) {
return null;
}
return (V) v1;
}
else {
final V v2 = function.apply( key );
if ( v2 == null ) {
map.put( key, NULL_TOKEN );
return null;
}
else {
final Object v3 = map.putIfAbsent( key, v2 );
if ( v3 == null ) {
return v2;
}
else {
return (V) v3;
}
}
}
}
public V get(final String key) {
final Object o = map.get( key );
if ( o == NULL_TOKEN) {
return null;
}
else {
return (V) o;
}
}
}
© 2015 - 2025 Weber Informatics LLC | Privacy Policy