use ConcurrentLinkedHashMap as backing for LRUMap

pom.xml

@@ -95,6 +95,12 @@
       <version>${version.powermock}</version>
       <scope>test</scope>
     </dependency>
+    <dependency>
+      <groupId>com.google.guava</groupId>
+      <artifactId>guava-testlib</artifactId>
+      <version>31.1-jre</version>
+      <scope>test</scope>
+    </dependency>
     <!-- For testing TestNoClassDefFoundDeserializer -->
     <dependency>
       <groupId>javax.measure</groupId>

src/main/java/com/fasterxml/jackson/databind/util/LRUMap.java

@@ -1,7 +1,6 @@
 package com.fasterxml.jackson.databind.util;
 
-import java.io.*;
-import java.util.concurrent.ConcurrentHashMap;
+import com.fasterxml.jackson.databind.util.internal.PrivateMaxEntriesMap;
 
 /**
  * Helper for simple bounded maps used for reusing lookup values.
@@ -10,44 +9,36 @@
  * on assumption that all use cases are for caching where persistence
  * does not make sense. The only thing serialized is the cache size of Map.
  *<p>
- * NOTE: since version 2.4.2, this is <b>NOT</b> an LRU-based at all; reason
- * being that it is not possible to use JDK components that do LRU _AND_ perform
- * well wrt synchronization on multi-core systems. So we choose efficient synchronization
- * over potentially more efficient handling of entries.
+ * NOTE: since Jackson 2.14, the implementation evicts the least recently used
+ * entry when max size is reached.
  *<p>
- * And yes, there are efficient LRU implementations such as
- * <a href="https://code.google.com/p/concurrentlinkedhashmap/">concurrentlinkedhashmap</a>;
- * but at this point we really try to keep external deps to minimum.
- * Plan from Jackson 2.12 is to focus more on pluggability as {@link LookupCache} and
- * let users, frameworks, provide their own cache implementations.
+ * Since Jackson 2.12, there has been pluggable {@link LookupCache} interface which
+ * allows users, frameworks, provide their own cache implementations.
  */
 public class LRUMap<K,V>
     implements LookupCache<K,V>, // since 2.12
         java.io.Serializable
 {
-    private static final long serialVersionUID = 1L;
+    private static final long serialVersionUID = 2L;
 
-    protected final transient int _maxEntries;
+    protected final int _initialEntries;
+    protected final int _maxEntries;
+    protected final transient PrivateMaxEntriesMap<K,V> _map;
 
-    protected final transient ConcurrentHashMap<K,V> _map;
-
     public LRUMap(int initialEntries, int maxEntries)
     {
-        // We'll use concurrency level of 4, seems reasonable
-        _map = new ConcurrentHashMap<K,V>(initialEntries, 0.8f, 4);
+        _initialEntries = initialEntries;
         _maxEntries = maxEntries;
+        // We'll use concurrency level of 4, seems reasonable
+        _map = new PrivateMaxEntriesMap.Builder<K, V>()
+                .initialCapacity(initialEntries)
+                .maximumCapacity(maxEntries)
+                .concurrencyLevel(4)
+                .build();
     }
 
     @Override
     public V put(K key, V value) {
-        if (_map.size() >= _maxEntries) {
-            // double-locking, yes, but safe here; trying to avoid "clear storms"
-            synchronized (this) {
-                if (_map.size() >= _maxEntries) {
-                    clear();
-                }
-            }
-        }
         return _map.put(key, value);
     }
 
@@ -56,21 +47,12 @@ public V put(K key, V value) {
      */
     @Override
     public V putIfAbsent(K key, V value) {
-        // not 100% optimal semantically, but better from correctness (never exceeds
-        // defined maximum) and close enough all in all:
-        if (_map.size() >= _maxEntries) {
-            synchronized (this) {
-                if (_map.size() >= _maxEntries) {
-                    clear();
-                }
-            }
-        }
         return _map.putIfAbsent(key, value);
     }
 
     // NOTE: key is of type Object only to retain binary backwards-compatibility
     @Override
-    public V get(Object key) {  return _map.get(key); }
+    public V get(Object key) { return _map.get(key); }
 
     @Override
     public void clear() { _map.clear(); }
@@ -84,23 +66,7 @@ public V putIfAbsent(K key, V value) {
     /**********************************************************
      */
 
-    /**
-     * Ugly hack, to work through the requirement that _value is indeed final,
-     * and that JDK serialization won't call ctor(s) if Serializable is implemented.
-     * 
-     * @since 2.1
-     */
-    protected transient int _jdkSerializeMaxEntries;
-
-    private void readObject(ObjectInputStream in) throws IOException {
-        _jdkSerializeMaxEntries = in.readInt();
-    }
-
-    private void writeObject(ObjectOutputStream out) throws IOException {
-        out.writeInt(_jdkSerializeMaxEntries);
-    }
-
     protected Object readResolve() {
-        return new LRUMap<Object,Object>(_jdkSerializeMaxEntries, _jdkSerializeMaxEntries);
+        return new LRUMap<K,V>(_initialEntries, _maxEntries);
     }
 }

src/main/java/com/fasterxml/jackson/databind/util/internal/EvictionListener.java

@@ -0,0 +1,45 @@
+/*
+ * Copyright 2010 Google Inc. All Rights Reserved.
+ *
+ * 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.fasterxml.jackson.databind.util.internal;
+
+/**
+ * A listener registered for notification when an entry is evicted. An instance
+ * may be called concurrently by multiple threads to process entries. An
+ * implementation should avoid performing blocking calls or synchronizing on
+ * shared resources.
+ * <p>
+ * The listener is invoked by {@link PrivateMaxEntriesMap} on a caller's
+ * thread and will not block other threads from operating on the map. An
+ * implementation should be aware that the caller's thread will not expect
+ * long execution times or failures as a side effect of the listener being
+ * notified. Execution safety and a fast turn around time can be achieved by
+ * performing the operation asynchronously, such as by submitting a task to an
+ * {@link java.util.concurrent.ExecutorService}.
+ *
+ * @author ben.manes@gmail.com (Ben Manes)
+ * @see <a href="http://code.google.com/p/concurrentlinkedhashmap/">
+ *      http://code.google.com/p/concurrentlinkedhashmap/</a>
+ */
+interface EvictionListener<K, V> {
+
+    /**
+     * A call-back notification that the entry was evicted.
+     *
+     * @param key the entry's key
+     * @param value the entry's value
+     */
+    void onEviction(K key, V value);
+}