summaryrefslogtreecommitdiffstats
path: root/env/src/main/java/org/onap/aaf/inno/env/util/Pool.java
diff options
context:
space:
mode:
Diffstat (limited to 'env/src/main/java/org/onap/aaf/inno/env/util/Pool.java')
-rw-r--r--env/src/main/java/org/onap/aaf/inno/env/util/Pool.java395
1 files changed, 395 insertions, 0 deletions
diff --git a/env/src/main/java/org/onap/aaf/inno/env/util/Pool.java b/env/src/main/java/org/onap/aaf/inno/env/util/Pool.java
new file mode 100644
index 0000000..204d51a
--- /dev/null
+++ b/env/src/main/java/org/onap/aaf/inno/env/util/Pool.java
@@ -0,0 +1,395 @@
+/*******************************************************************************
+ * ============LICENSE_START====================================================
+ * * org.onap.aaf
+ * * ===========================================================================
+ * * Copyright © 2017 AT&T Intellectual Property. 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.
+ * * ============LICENSE_END====================================================
+ * *
+ * * ECOMP is a trademark and service mark of AT&T Intellectual Property.
+ * *
+ ******************************************************************************/
+/*
+ * Pool
+ *
+ * 5/27/2011
+ */
+package org.onap.aaf.inno.env.util;
+
+import java.util.LinkedList;
+
+import org.onap.aaf.inno.env.APIException;
+import org.onap.aaf.inno.env.LogTarget;
+
+/**
+ * This Class pools on an As-Needed-Basis any particular kind of class, which is
+ * quite suitable for expensive operations.
+ *
+ * The user calls "get" on a Pool, and if a waiting resource (T) is available,
+ * it will be returned. Otherwise, one will be created with the "Creator" class
+ * (must be defined for (T)).
+ *
+ * You can Prime the instances to avoid huge startup costs
+ *
+ * The returned "Pooled" object simply has to call "done()" and the object is
+ * returned to the pool. If the developer does not return the object, a memory
+ * leak does not occur. There are no references to the object once "get" is
+ * called. However, the developer who does not return the object when done
+ * obviates the point of the pool, as new Objects are created in place of the
+ * Object not returned when another call to "get" is made.
+ *
+ * There is a cushion of extra objects, currently defaulted to MAX_RANGE. If the
+ * items returned become higher than the MAX_RANGE, the object is allowed to go
+ * out of scope, and be cleaned up. the default can be changed on a per-pool
+ * basis.
+ *
+ *
+ *
+ * @param <T>
+ */
+public class Pool<T> {
+ /**
+ * This is a constant which specified the default maximum number of unused
+ * objects to be held at any given time.
+ */
+ private static final int MAX_RANGE = 6; // safety
+
+ /**
+ * only Simple List needed.
+ *
+ * NOTE TO MAINTAINERS: THIS OBJECT DOES IT'S OWN SYNCHRONIZATION. All
+ * changes that touch list must account for correctly synchronizing list.
+ */
+ private LinkedList<Pooled<T>> list;
+
+ /**
+ * keep track of how many elements exist, to avoid asking list.
+ */
+ private int count;
+
+ /**
+ * Spares are those Object that are primed and ready to go.
+ */
+ private int spares;
+
+ /**
+ * Actual MAX number of spares allowed to hang around. Can be set to
+ * something besides the default MAX_RANGE.
+ */
+ private int max_range = MAX_RANGE;
+
+ /**
+ * The Creator for this particular pool. It must work for type T.
+ */
+ private Creator<T> creator;
+
+ /**
+ * Create a new Pool, given the implementation of Creator<T>, which must be
+ * able to create/destroy T objects at will.
+ *
+ * @param creator
+ */
+ public Pool(Creator<T> creator) {
+ count = spares = 0;
+ this.creator = creator;
+ list = new LinkedList<Pooled<T>>();
+ }
+
+ /**
+ * Preallocate a certain number of T Objects. Useful for services so that
+ * the first transactions don't get hit with all the Object creation costs
+ *
+ * @param lt
+ * @param prime
+ * @throws APIException
+ */
+ public void prime(LogTarget lt, int prime) throws APIException {
+ for (int i = 0; i < prime; ++i) {
+ Pooled<T> pt = new Pooled<T>(creator.create(), this, lt);
+ synchronized (list) {
+ list.addFirst(pt);
+ ++count;
+ }
+ }
+
+ }
+
+ /**
+ * Destroy and remove all remaining objects. This is valuable for closing
+ * down all Allocated objects cleanly for exiting. It is also a good method
+ * for removing objects when, for instance, all Objects are invalid because
+ * of broken connections, etc.
+ */
+ public void drain() {
+ synchronized (list) {
+ for (int i = 0; i < list.size(); ++i) {
+ Pooled<T> pt = list.remove();
+ creator.destroy(pt.content);
+ pt.logTarget.log("Pool drained ", creator.toString());
+ }
+ count = spares = 0;
+ }
+
+ }
+
+ /**
+ * This is the essential function for Pool. Get an Object "T" inside a
+ * "Pooled<T>" object. If there is a spare Object, then use it. If not, then
+ * create and pass back.
+ *
+ * This one uses a Null LogTarget
+ *
+ * IMPORTANT: When the use of this object is done (and the object is still
+ * in a valid state), then "done()" should be called immediately to allow
+ * the object to be reused. That is the point of the Pool...
+ *
+ * If the Object is in an invalid state, then "toss()" should be used so the
+ * Pool doesn't pass on invalid objects to others.
+ *
+ * @param lt
+ * @return
+ * @throws APIException
+ */
+ public Pooled<T> get() throws APIException {
+ Pooled<T> pt;
+ synchronized (list) {
+ if (list.isEmpty()) {
+ pt = null;
+ } else {
+ pt = list.removeLast();
+ --count;
+ creator.reuse(pt.content);
+ }
+ }
+ if (pt == null) {
+ if (spares < max_range)
+ ++spares;
+ pt = new Pooled<T>(creator.create(), this, LogTarget.NULL);
+ } else {
+ if (spares > 1)
+ --spares;
+ }
+ return pt;
+ }
+
+ /**
+ * This is the essential function for Pool. Get an Object "T" inside a
+ * "Pooled<T>" object. If there is a spare Object, then use it. If not, then
+ * create and pass back.
+ *
+ * If you don't have access to a LogTarget from Env, use LogTarget.NULL
+ *
+ * IMPORTANT: When the use of this object is done (and the object is still
+ * in a valid state), then "done()" should be called immediately to allow
+ * the object to be reused. That is the point of the Pool...
+ *
+ * If the Object is in an invalid state, then "toss()" should be used so the
+ * Pool doesn't pass on invalid objects to others.
+ *
+ * @param lt
+ * @return
+ * @throws APIException
+ */
+ public Pooled<T> get(LogTarget lt) throws APIException {
+ Pooled<T> pt;
+ synchronized (list) {
+ if (list.isEmpty()) {
+ pt = null;
+ } else {
+ pt = list.remove();
+ --count;
+ creator.reuse(pt.content);
+ }
+ }
+ if (pt == null) {
+ if (spares < max_range)
+ ++spares;
+ pt = new Pooled<T>(creator.create(), this, lt);
+ lt.log("Pool created ", creator.toString());
+ } else {
+ if (spares > 1)
+ --spares;
+ }
+ return pt;
+ }
+
+ /**
+ * This function will validate whether the Objects are still in a usable
+ * state. If not, they are tossed from the Pool. This is valuable to have
+ * when Remote Connections go down, and there is a question on whether the
+ * Pooled Objects are still functional.
+ *
+ * @return
+ */
+ public boolean validate() {
+ boolean rv = true;
+ synchronized (list) {
+ for (Pooled<T> t : list) {
+ if (!creator.isValid(t.content)) {
+ rv = false;
+ t.toss();
+ list.remove(t);
+ }
+ }
+ }
+ return rv;
+ }
+
+ /**
+ * This is an internal method, used only by the Internal Pooled<T> class.
+ *
+ * The Pooled<T> class "offers" it's Object back after use. It is an
+ * "offer", because Pool will simply destroy and remove the object if it has
+ * more than enough spares.
+ *
+ * @param lt
+ * @param used
+ * @return
+ */
+ // Used only by Pooled<T>
+ private boolean offer(LogTarget lt, Pooled<T> used) {
+ if (count < spares) {
+ synchronized (list) {
+ list.addFirst(used);
+ ++count;
+ }
+ lt.log("Pool recovered ", creator.toString());
+ } else {
+ lt.log("Pool destroyed ", creator.toString());
+ creator.destroy(used.content);
+ }
+ return false;
+ }
+
+ /**
+ * The Creator Interface give the Pool the ability to Create, Destroy and
+ * Validate the Objects it is maintaining. Thus, it is a specially written
+ * Implementation for each type.
+ *
+ *
+ * @param <T>
+ */
+ public interface Creator<T> {
+ public T create() throws APIException;
+
+ public void destroy(T t);
+
+ public boolean isValid(T t);
+
+ public void reuse(T t);
+ }
+
+ /**
+ * The "Pooled<T>" class is the transient class that wraps the actual Object
+ * T for API use/ It gives the ability to return ("done()", or "toss()") the
+ * Object to the Pool when processing is finished.
+ *
+ * For Safety, i.e. to avoid memory leaks and invalid Object States, there
+ * is a "finalize" method. It is strictly for when coder forgets to return
+ * the object, or perhaps hasn't covered the case during Exceptions or
+ * Runtime Exceptions with finally (preferred). This should not be
+ * considered normal procedure, as finalize() is called at an undetermined
+ * time during garbage collection, and is thus rather useless for a Pool.
+ * However, we don't want Coding Mistakes to put the whole program in an
+ * invalid state, so if something happened such that "done()" or "toss()"
+ * were not called, the resource is still cleaned up as well as possible.
+ *
+ *
+ * @param <T>
+ */
+ public static class Pooled<T> {
+ public final T content;
+ private Pool<T> pool;
+ protected LogTarget logTarget;
+
+ /**
+ * Create the Wrapping Object Pooled<T>.
+ *
+ * @param t
+ * @param pool
+ * @param logTarget
+ */
+ public Pooled(T t, Pool<T> pool, LogTarget logTarget) {
+ content = t;
+ this.pool = pool;
+ this.logTarget = logTarget;
+ }
+
+ /**
+ * This is the key API for the Pool, as calling "done()" offers this
+ * object back to the Pool for reuse.
+ *
+ * Do not use the Pooled<T> object again after calling "done()".
+ */
+ public void done() {
+ if (pool != null) {
+ pool.offer(logTarget, this);
+ }
+ }
+
+ /**
+ * The user of the Object may discover that the Object t is no longer in
+ * a valid state. Don't put Garbage back in the Refrigerator... Toss it,
+ * if it's no longer valid.
+ *
+ * toss() is also used for draining the Pool, etc.
+ *
+ * toss() will attempt to destroy the Object by using the Creator
+ * Interface.
+ *
+ */
+ public void toss() {
+ if (pool != null) {
+ pool.creator.destroy(content);
+ }
+ // Don't allow finalize to put it back in.
+ pool = null;
+ }
+
+ /**
+ * Just in case someone neglected to offer back object... Do not rely on
+ * this, as there is no specific time when finalize is called, which
+ * rather defeats the purpose of a Pool.
+ */
+ @Override
+ protected void finalize() throws Throwable {
+ if (pool != null) {
+ done();
+ pool = null;
+ }
+ }
+ }
+
+ /**
+ * Get the maximum number of spare objects allowed at any moment
+ *
+ * @return
+ */
+ public int getMaxRange() {
+ return max_range;
+ }
+
+ /**
+ * Set a Max Range for numbers of spare objects waiting to be used.
+ *
+ * No negative numbers are allowed
+ *
+ * @return
+ */
+ public void setMaxRange(int max_range) {
+ // Do not allow negative numbers
+ this.max_range = Math.max(0, max_range);
+ }
+
+}