lrucache.java
来自「一个不错的cache」· Java 代码 · 共 233 行
JAVA
233 行
/* * Copyright (c) 2002-2003 by OpenSymphony * All rights reserved. */package com.opensymphony.oscache.base.algorithm;import com.opensymphony.oscache.util.ClassLoaderUtil;import org.apache.commons.collections.SequencedHashMap;import org.apache.commons.logging.Log;import org.apache.commons.logging.LogFactory;import java.util.*;/** * <p>LRU (Least Recently Used) algorithm for the cache.</p> * * <p>This class tries to provide the best possible performance by first * attempting to use the JDK 1.4.x <code>LinkedHashSet</code> class, * followed by the Jakarta commons-collections <code>SequencedHashMap</code> * class, and finally resorting to the <code>LinkedList</code> class if * neither of the above classes are available. If this class has to revert * to using a <code>LinkedList</code> a warning is logged since the performance * penalty can be severe.</p> * * <p>No synchronization is required in this class since the * <code>AbstractConcurrentReadCache</code> already takes care of any * synchronization requirements.</p> * * @version $Revision: 1.5 $ * @author <a href="mailto:salaman@teknos.com">Victor Salaman</a> * @author <a href="mailto:fbeauregard@pyxis-tech.com">Francois Beauregard</a> * @author <a href="mailto:abergevin@pyxis-tech.com">Alain Bergevin</a> * @author <a href="mailto:chris@swebtec.com">Chris Miller</a> */public class LRUCache extends AbstractConcurrentReadCache { private static final Log log = LogFactory.getLog(LRUCache.class); /** * Cache queue containing all cache keys. */ private Collection list; /** * Jakarta commons collections unfortunately doesn't provide us with an ordered * hash set, so we have to use their ordered map instead. */ private Map map; /** * A flag indicating if we are using a List for the key collection. This happens * when we're running under JDK 1.3 or lower and there is no commons-collections * in the classpath. */ private boolean isList = false; /** * A flag indicating if we are using a Map for the key collection. This happens * when we're running under JDK 1.3 and commons-collections is available. */ private boolean isMap = false; /** * A flag indicating if we are using a Set for the key collection. This happens * when we're running under JDK 1.4 and is the best case scenario. */ private boolean isSet = false; /** * A flag indicating whether there is a removal operation in progress. */ private volatile boolean removeInProgress = false; /** * Constructs an LRU Cache. */ public LRUCache() { super(); // Decide if we're running under JRE 1.4+. If so we can use a LinkedHashSet // instead of a LinkedList for a big performance boost when removing elements. try { ClassLoaderUtil.loadClass("java.util.LinkedHashSet", this.getClass()); list = new LinkedHashSet(); isSet = true; } catch (ClassNotFoundException e) { // There's no LinkedHashSet available so we'll try for the jakarta-collections // SequencedHashMap instead [CACHE-47] try { ClassLoaderUtil.loadClass("org.apache.commons.collections.SequencedHashMap", this.getClass()); map = new SequencedHashMap(); isMap = true; } catch (ClassNotFoundException e1) { // OK, time to get all inefficient and resort to a LinkedList. We log this // as a warning since it potentially can have a big impact. log.warn("When using the LRUCache under JRE 1.3.x, commons-collections.jar should be added to your classpath to increase OSCache's performance."); list = new LinkedList(); isList = true; } } } /** * Constructors a LRU Cache of the specified capacity. * * @param capacity The maximum cache capacity. */ public LRUCache(int capacity) { this(); maxEntries = capacity; } /** * An item was retrieved from the list. The LRU implementation moves * the retrieved item's key to the front of the list. * * @param key The cache key of the item that was retrieved. */ protected void itemRetrieved(Object key) { // Prevent list operations during remove while (removeInProgress) { try { Thread.sleep(5); } catch (InterruptedException ie) { } } // We need to synchronize here because AbstractConcurrentReadCache // doesn't prevent multiple threads from calling this method simultaneously. if (isMap) { synchronized (map) { map.remove(key); map.put(key, Boolean.TRUE); } } else { synchronized (list) { list.remove(key); list.add(key); } } } /** * An object was put in the cache. This implementation adds/moves the * key to the end of the list. * * @param key The cache key of the item that was put. */ protected void itemPut(Object key) { // Since this entry was just accessed, move it to the back of the list. if (isMap) { synchronized (map) { // A further fix for CACHE-44 map.remove(key); map.put(key, Boolean.TRUE); } } else { synchronized (list) { // A further fix for CACHE-44 list.remove(key); list.add(key); } } } /** * An item needs to be removed from the cache. The LRU implementation * removes the first element in the list (ie, the item that was least-recently * accessed). * * @return The key of whichever item was removed. */ protected Object removeItem() { removeInProgress = true; Object toRemove; try { toRemove = removeFirst(); } catch (Exception e) { // List is empty. // this is theorically possible if we have more than the size concurrent // thread in getItem. Remove completed but add not done yet. // We simply wait for add to complete. do { try { Thread.sleep(5); } catch (InterruptedException ie) { } } while (isMap ? (map.size() == 0) : (list.size() == 0)); toRemove = removeFirst(); } removeInProgress = false; return toRemove; } /** * Remove specified key since that object has been removed from the cache. * * @param key The cache key of the item that was removed. */ protected void itemRemoved(Object key) { if (isMap) { map.remove(key); } else { list.remove(key); } } /** * Removes the first object from the list of keys. * * @return the object that was removed */ private Object removeFirst() { Object toRemove; if (isSet) { Iterator it = list.iterator(); toRemove = it.next(); it.remove(); } else if (isMap) { toRemove = ((SequencedHashMap) map).getFirstKey(); map.remove(toRemove); } else { toRemove = ((List) list).remove(0); } return toRemove; }}
⌨️ 快捷键说明
复制代码Ctrl + C
搜索代码Ctrl + F
全屏模式F11
增大字号Ctrl + =
减小字号Ctrl + -
显示快捷键?