java/common/lib/backport-util-concurrent/README.html - qpid - Git at Google

 <html>
   <head>
     <title>backport-util-concurrent - Distributed Computing Laboratory</title>
     <meta content="java.util.concurrent backport">
   </head>
   <body>

 <h1>backport-util-concurrent</h1>
 <p>Backport ot JSR 166 (java.util.concurrent) to Java 1.4</p>
 <p><a href="http://dcl.mathcs.emory.edu/util/backport-util-concurrent/">http://dcl.mathcs.emory.edu/util/backport-util-concurrent/</a></p>
 <hr class="hide">

 <h2>Overview</h2>

 <p>
 This package is the backport of
 <a href="http://gee.cs.oswego.edu/dl/concurrency-interest/">java.util.concurrent</a> API,
 introduced in
 <a href="http://java.sun.com/j2se/1.5.0/docs/guide/concurrency/overview.html">Java 5.0</a>,
 to Java 1.4. The backport is based on public-domain sources from the
 <a href="http://gee.cs.oswego.edu/cgi-bin/viewcvs.cgi/jsr166/src/main/java/">
 JSR 166 CVS repository</a>, the
 <a href="http://gee.cs.oswego.edu/dl/classes/EDU/oswego/cs/dl/util/concurrent/intro.html">
 dl.util.concurrent</a> package, and the Doug Lea's
 <a href="http://gee.oswego.edu/dl/classes/collections/">
 collections</a> package.

 The backport is close to complete; unsupported functionality is
 limited to: 1) classes that rely on explicit
 JVM support and cannot be emulated at a satisfactory
 performance level, 2) some of the functions
 described in the original javadoc as "designed primarily for use in
 monitoring in system state, not for synchronization control", or 3)
 functionality
 that would require development of substantial amount of new code.
 </p>
 <p>
 The purpose of this library is to enable gradual migration from Java 1.4
 to 5.0: the library allows to develop concurrent applications for Java 1.4
 that should work with Java 5.0 simply by changing package names.
 </p>

 <h3>License</h3>

 <p>
 This software is released to the
 <a href="http://creativecommons.org/licenses/publicdomain">public domain</a>,
 in the spirit of the original code written by Doug Lea.
 The code can be used for any purpose, modified, and redistributed
 without acknowledgment. No warranty is provided, either express or implied.
 </p>
 <p>
 <em>Note:</em> versions prior to 2.1 had dependencies on proprietary code.
 	Versions 2.1 and newer do not have such dependencies anymore.
 </p>

 <h3>Features</h3>

 <p>
 The following functionality of java.util.concurrent is supported in the backport:
 <ul>
 <li>All JSR 166 executors, utilities, and everything related (thread pools,
 FutureTask, scheduled tasks and executors, etc.)</li>
 <li>Locks: ReentrantLock, Semaphore, ReentrantReadWriteLock, Conditions</li>
 <li>Queues: synchronous, array, linked, delay, and priority queues</li>
 <li>Deques: array, and linked deques</li>
 <li>Atomics: everything except reflection-based updaters</li>
 <li>Other concurrency utils: CountDownLatch, CyclicBarrier, Exchanger</li>
 <li>Concurrent collections: ConcurrentHashMap, CopyOnWriteArrayList, CopyOnWriteArraySet,
     ConcurrentLinkedQueue,
     ConcurrentSkipList[Map,Set]</li>
 <li>Retrofitted standard collections</li>
 </ul>
 </p>

 <h2>Compatibility</h2>

 JSR 166 functionality is tied closely to the Java 5.0 Virtual Machine, and some aspects of it
 cannot be fully
 backported to Java 1.4. This section discusses these differences between the backport and
 JSR 166.

 <h3>Package Names</h3>

 <p>Since user libraries cannot define classes in
 java.* packages, all the original package names have been prefixed with
 <code>edu.emory.mathcs.backport</code>. For instance, <code>java.util.concurrent.Future</code> becomes <code>edu.emory.mathcs.backport.java.util.concurrent.Future</code>.
 </p>

 <h3>Differences between versions</h3>

 <p>
 The backport, version 1.1 and above, includes
 functionality of JSR 166 that has been added in Java 6.0.
 Pay attention to the "since" javadoc tags if conformance with specific
 Java platform versions is desired. Examples of "since 1.6" functionality include:
 deques, navigable maps and sets (including ConcurrentSkipList[Map,Set]),
 "newTaskFor" in AbstractExecutorService,
 "lazySet" in atomics, RunnableFuture and RunnableScheduledFuture,
 "allowCoreThreadTimeout" in ThreadPoolExecutor,
 "decorateTask" in ScheduledThreadPoolExecutor, MINUTES, HOURS, and DAYS in TimeUnit,
 and appropriate retrofits in collection classes. As of backport version 2.2, these
 features are based on beta versions of Java 6.0 APIs, which may still change in the future.
 </p>
 <p>Backport is developed carefully to retain link-time compatibility, i.e. it is generally
    safe to replace an old library JAR with a new one (with a possible exception of APIs
    based on beta releases, e.g. current "since 1.6" classes and methods). Serial compatibility
    (i.e. ability of
    one version to deserialize objects that has been serialized using a different version)
    is maintained on a  best-effort basis, and not always guaranteed.
    Please see details below. (Note that concurrency tools are usually not intended for
    persistent storage anyway). Compile-time compatibility: applications using
    wildcard imports (e.g. java.util.* and edu.emory.mathcs.backport.java.util.*) may
    cease to compile with updated backport versions (containing new classes)
    due to import ambiguities. In such cases, you must dis-ambiguate
    imports (i.e. use explicit imports as appropriate) and recompile.
 </p>
 <p>Notes for version 2.2:
    Link-time and serial compatibility is fully preserved for "since 1.5" APIs. For
    "since 1.6" APIs, link-time and serial compatibility is preserved except for
    navigable maps and sets, which API has recently changed slightly in
    Java 6.0 beta.
 </p>
 <p>Notes for version 2.1:
    Link-time compatibility is preserved fully.
    Serial compatibility is preserved except for the class ReentrantReadWriteLock.
 </p>
 <p>Notes for version 2.0:
 <ul>
 <li>the "concurrent.Concurrent" class has been removed as of backport 2.0.</li>
 <li>Serialized representations
     of several lock classes have changed between versions 1.1_01 and 2.0,
     as a result of certain bug fixes and performance improvements (see changelog).
     This means that locks and collections serialized with 1.1_01 will not be
     deserialized properly by 2.0.
 </li>
 </ul>
 </p>

 <h3>Unsupported functionality</h3>

 <p>
 Detailed listing of functionality that has not been backported
 is presented below.
 </p>

 <h4>Java 5.0 Syntax</h4>
 <p>
 Package java.util.concurrent exploits new language features
 introduced in Java 5.0. In particular, most API classes are
 <a href="http://java.sun.com/j2se/1.5.0/docs/guide/language/generics.html">generic types</a>.
 In the backport, they have been flattened to standard, non-generic
 classes. Still, programs linked against the backport should compile
 with Java 5.0 (after changing package names). Nevertheless, you may
 want to consider gradually switching to using generics once you make
 the transition to Java 5.0, since it gives better compile-time
 type checking.
 </p>

 <h4>In Condition</h4>
 <p>
 Method long awaitNanos(long nanosTimeout) is not supported, since the
 emulation cannot reliably report remaining times with nanosecond
 precision. Thus, it probably would be too dangerous to leave the
 emulated method in the Condition interface. However, the method is
 still available, for those who know what they are doing,
 in the <a href="doc/api/edu/emory/mathcs/backport/java/util/concurrent/helpers/Utils.html">util.concurrent.helpers.Utils</a> class.
 </p>

 <h4>In ReentrantLock</h4>
 <p>
 the following monitoring methods are supported only for fair locks:
 boolean hasQueuedThreads(), int getQueueLength(), Collection
 getQueuedThreads(), boolean isQueued(), boolean hasWaiters(Condition),
 int getWaitQueueLength(Condition),
 Collection getWaitingThreads(Condition).
 </p>

 <h4>In ReentrantReadWriteLock</h4>
 <p>
 The current backport implementation is based on dl.util.concurrent class
 ReentrantWriterPreferenceReadWriteLock, and thus slightly departs
 from java.util.concurrent that does not specify acquisition order but
 allows to enable/disable fairness. The backport implementation does not
 have a single-parameter constructor allowing to specify fairness policy;
 it always behaves like writer-preference lock with no fairness guarantees.
 Because of these characteristics, this class is compliant with JSR 166
 specification of non-fair reentrant read-write locks, while the exact
 semantics of fair locks are not supported (and the appropriate
 constructor is thus not provided).
 </p>
 <p>
 Also, the following instrumentation and status methods are not
 supported: Collection getQueuedWriterThreads(), Collection
 getQueuedReaderThreads(), boolean hasQueuedThreads(), boolean
 hasQueuedThread(Thread), Collection getQueuedThreads(), boolean
 hasWaiters(Condition), int getWaitQueueLength(Condition), Collection
 getWaitingThreads(Condition).
 </p>

 <h4>In Semaphore</h4>
 <p>
 Blocking atomic multi-acquires: acquire(int permits)
 and
 tryAcquire(int permits, long timeout, TimeUnit unit)
 are supported only for FAIR semaphores.
 </p>

 <h4>Platform-level functionality</h4>
 <p>
 To emulate System.nanoTime(), the method
 <a href="doc/api/edu/emory/mathcs/backport/java/util/concurrent/helpers/Utils.html#nanoTime()">nanoTime()</a>
 is provided in the class
 dl.util.concurrent.helpers.Utils. On Java 1.4.2, it attempts to use
 high-precision timer via sun.misc.Perf (thanks to Craig Mattocks
 for suggesting this). On older Java platforms, or when sun.misc.Perf
 is not supported, it falls back to System.currentTimeMillis().
 </p>
 <p>
 Class <a href="doc/api/edu/emory/mathcs/backport/java/util/concurrent/helpers/ThreadHelpers.html">ThreadHelpers</a>
 (added in 1.0_01) is provided to emulate certain aspects of Thread.UncaughtExceptionHandler.
 </p>

 <h4>Note on nanosecond precision timing</h4>
 <p>
 The backport strives to honor nanosecond timeouts, if such are requested,
 by using two-parameter variant of Object.wait(). Note, however, that most
 Java platforms before 5.0 will round up the timeout to full milliseconds
 anyway.
 </p>

 <h4>Low-level concurrency classes</h4>
 <p>
 The following classes are not supported:
 LockSupport, AbstractQueuedSynchronizer, AbstractQueuedLongSynchronizer.
 </p><p><i>Rationale: </i> on Java 5.0, these classes depend on explicit
 JVM support, delegating to low-level OS concurrency primitives. There seems
 to be no simple way of emulating them without introducing prohibitive
 performance overhead. (If you think they should be present in the backport
 anyway, let me know).
 </p>

 <h4>Atomic utilities</h4>
 <p>
 The following "atomic" utilities are not supported:
 Atomic[Integer,Long,Reference]FieldUpdater.

 </p>

 <h2>Robustness</h2>

 <p>
 Backport-util-concurrent is based in large part on source code from JSR 166
 and dl.util.concurrent, both very
 well tested. Whenever possible, the JSR 166 code was used. In cases when
 it was infeasible (e.g. for performance reasons),
 the dl.util.concurrent code was adapted. The new
 code was introduced only when absolutely necessary, e.g. to make
 dl.util.concurrent code conforming to JSR 166 interfaces and semantics. This
 partially explains why so few bugs have been reported again the backport,
 despite over 10,000 downloads and many deployments in commercial and
 open-source projects.
 </p>

 <h3>Unit tests</h3>
 <p>
 Version 2.1 of the library passes all the relevant 1859 unit tests from
 <a href="http://gee.cs.oswego.edu/cgi-bin/viewcvs.cgi/jsr166/src/test/tck/">TCK test package</a>
 designed for java.util.concurrent (the tests of unsupported functionality
 were skipped).
 </p>
 <p>
 The following unit tests have been completed (listed in the alphabetical order):
 AbstractExecutorServiceTest,
 AbstractQueueTest,
 ArrayBlockingQueueTest,
 ArrayDequeTest,
 AtomicBooleanTest,
 AtomicIntegerArrayTest,
 AtomicIntegerTest,
 AtomicLongArrayTest,
 AtomicLongTest,
 AtomicMarkableReferenceTest,
 AtomicReferenceArrayTest,
 AtomicReferenceTest,
 AtomicStampedReferenceTest,
 ConcurrentHashMapTest,
 ConcurrentLinkedQueueTest,
 ConcurrentSkipListMapTest,
 ConcurrentSkipListSubMapTest,
 ConcurrentSkipListSetTest,
 ConcurrentSkipListSubSetTest,
 CopyOnWriteArrayListTest,
 CopyOnWriteArraySetTest,
 CountDownLatchTest,
 CyclicBarrierTest,
 DelayQueueTest,
 EntryTest,
 ExchangerTest,
 ExecutorsTest,
 ExecutorCompletionServiceTest,
 FutureTaskTest,
 LinkedBlockingDequeTest,
 LinkedBlockingQueueTest,
 LinkedListTest,
 PriorityBlockingQueueTest,
 PriorityQueueTest,
 ReentrantLockTest,
 ReentrantReadWriteLockTest,
 ScheduledExecutorTest,
 ScheduledExecutorSubclassTest,
 SemaphoreTest,
 SynchronousQueueTest,
 SystemTest (testing Utils.nanoTime()),
 ThreadLocalTest,
 ThreadPoolExecutorTest,
 ThreadPoolExecutorSubclassTest,
 TimeUnitTest,
 TreeMapTest,
 TreeSubMapTest,
 TreeSetTest,
 TreeSubSetTest.
 </p>

 <h3>Stress tests</h3>
 <p>
 Starting from version 1.1_01, the
 backport is being stress-tested using the "loops" tests from JSR 166
 (courtesy of Doug Lea and the JSR 166 Expert Group). The tests, included
 in the development source
 bundle, thoroughly evaluate behavior and performance of various types
 of locks, queues, maps, futures, and other API classes, under various
 conditions (contention etc.) and circumstances, including cancellation.
 </p>

 <p>
 Despite exhaustive testing, as any software, this library may still contain
 bugs. If you find one, please report it, or better yet, contribute a fix.
 </p>

 <h3>Known problems</h3>

 <p>
 <strong>Note: </strong>A bug has been reported against Sun 1.4.2_04 JVMs, and fixed in
 1.4.2_06 (see ID 4917709) that makes those JVMs to occassionally crash with SIGSEGV
 during backport stress tests,
 particularly MapLoops and MapChecks. It is therefore recommended to use JVM versions
 1.4.2_06 or newer when using the backport (although the crashes seem to not happen also
 on 1.4.2_03, and perhaps on earlier JVMs). Detected in version: 2.0.
 </p>
 <p>
 <strong>Note: </strong>due to what is apparently a bug in SUN JVM implementations
 for Solaris, observed on 1.4.2_03 and 1.4.2_06,
 the 'ExecutorsTest.testPrivilegedThreadFactory()'
 unit test fails with ClassNotFoundException when launched from a class path
 that has backport classes stored as individual files in the "classes" directory. The
 problem disappears when the classes are put in a JAR file. The bug is most likely
 related to handling context class loaders. It is therefore advised to use JAR
 files instead of class files when running code that explicitly or implicitly
 modifies context class loaders, as does privileged thread factory.
 Detected in version: 2.0.
 </p>
 <p>
 <strong>Note: </strong>missed signals have been observed on Linux 2.6.3-7 kernel
 for SMP w/64GB support under contention and in the presence of frequent timeouts.
 (The bug was captured during TimeoutProducerConsumerLoops on SynchronousQueue).
 Apparently, this is caused by a kernel bug. The problem have been observed on
 several different JVMs. It does not occur on newer kernels.
 Detected in version: 2.0.
 </p>
 <p>
 As evident from the above, IT IS CRUCIAL THAT YOU RUN THE STRESS TESTS on the
 target configuration before using the backport in a production environment.
 Concurrency issues are tricky, and possible bugs in JVMs, operating systems,
 and this software, usually won't show up until under heavy loads. Stress tests
 included with this distribution test this software under extreme conditions,
 so if they are consistently passing, there's a very good chance that everything
 works fine.
 </p>

 <h2>Changelog</h2>

 <p>

 Version 2.2 (Jun 4, 2006) [<a href="http://www.mathcs.emory.edu/dcl/util/backport-util-concurrent/dist/backport-util-concurrent-2.2/backport-util-concurrent-2.2-changelog.html">CVS log</a>]

 <ul>
 <li>New features</li>
   <ul>
     <li>The backport now compiles under Java 5.0.</li>
     <li>Enhancements in the Navigable[Map,Set] interfaces.</li>
     <li>Blocking atomic multi-acquires in fair semaphores.</li>
     <li>Javadoc enhancements (reconciled with recent java.util.concurrent).</li>
     <li>Shutdown upon finalization for factory-created executors.</li>
   </ul>
 <li>Bug fixes</li>
   <ul>
     <li>broken type-checked map in Collections. Thanks for Taras Puchko for finding this bug
         and submitting the fix.</li>
     <li>Collections.reverseComparator(Comparator) not working properly when null passed as
         the argument.</li>
   </ul>
 <li>Tests</li>
   <ul>
     <li>Updated and reconciled with java.util.concurrent tests.</li>
   </ul>
 </ul>


 Version 2.1 (Jan 28, 2006) [<a href="http://www.mathcs.emory.edu/dcl/util/backport-util-concurrent/dist/backport-util-concurrent-2.1/backport-util-concurrent-2.1-changelog.html">CVS log</a>]
 <ul>
 <li>New features</li>
   <ul>
   <li>Descending iterators in deques</li>
   <li>Use newTaskFor() in ExecutionCompletionService.submit()</li>
   <li>toArray(Object[]) appends null at the end in LinkedBlockingQueue</li>
   <li>Overflow detection in ReentrantLock</li>
   <li>ReentrantReadWriteLock: better conformance with JSR 166 by adding public inner classes for ReadLock and WriteLock</li>
   <li>CopyOnWriteArraySet.equals() optimized towards small sets</li>
   <li>Snapshot iterators in queues</li>
   <li>Improved performance of toArray() in several collection classes</li>
   <li>More collection stuff ported, including new things in Arrays, and base collection classes with toArray() supporting
       concurrent collections</li>
   <li>Improved comparison of ScheduledFutureTasks in the ScheduledThreadPoolExecutor</li>
   </ul>
 <li>Licensing</li>
 <ul>
   <li>New, public domain implementations for CopyOnWriteArrayList, TreeMap, TreeSet, LinkedList, Collections, Arrays</li>
 </ul>
 <li>Bug fixes</li>
   <ul>
   <li>Methods equals() and hashCode() were broken in PriorityQueue. The fix allows PriorityQueues to be used as hash keys.</li>
   <li>ReentrantReadWriteLock.getWriteHoldCount could return a posititive value even if the write lock was not owned by the  inquiring thread</li>
   <li>Condition variables were not working properly with reentrant locks when the hold count was greater than 1. Await methods were releasing only a single hold, not all of them, as they should</li>
   <li>Handling of non-comparable entries (which is an erroneous condition) by priority queues has been made more
       deterministic. (This issue/fix does not affect correctly written programs)</li>
   <li>Fix of CR 6312056 (ConcurrentHashMap.entrySet().iterator() can return entry with never-existent value)</li>
   <li>Livelock in Exchanger if used by more than two threads</li>
   <li>Erroneous behavior of interrupted CyclicBarrier and locks on some (buggy) JVMs (thanks to Yew-Yap Goh for
       reporting this)</li>
   </ul>
 <li>Tests</li>
   <ul>
   <li>New and improved "loops" tests, including CollectionLoops, IteratorLoops, StringMapLoops,
       TSPExchangerTest, TimeoutExchangerLoops, UnboundedQueueFillEmptyLoops, EntryTest</li>
   <li>New "serial compatibility" test</li>
   </ul>
 </ul>

 Version 2.0_01 (Aug 3, 2005) [<a href="http://www.mathcs.emory.edu/dcl/util/backport-util-concurrent/dist/backport-util-concurrent-2.0_01/backport-util-concurrent-2.0_01-changelog.html">CVS log</a>]
 <ul>
 <li>Compatibility fix: ConcurrentHashMap was no longer inheriting from java.util.AbstractMap, although it was in version 1.1_01. Now it does again.</li>
 <li>Licensing: new, public-domain implementation of PriorityQueue, and refactoring of backported AbstractMap so that it also contains only the public domain code.</li>
 </ul>
 Version 2.0 (Jul 6, 2005) [<a href="http://www.mathcs.emory.edu/dcl/util/backport-util-concurrent/dist/backport-util-concurrent-2.0/backport-util-concurrent-2.0-changelog.html">CVS log</a>]
 <ul>
 <li>New features</li>
 <ul>
 <li>Features and fixes resulting from reconcillation with JSR 166 as of Jul 4,
     such as:
     lazySet in atomics,
     <a href="doc/api/edu/emory/mathcs/backport/java/util/concurrent/AbstractExecutorService.html#newTaskFor(java.lang.Runnable, java.lang.Object)">AbstractExecutorService.newTaskFor()</a>,
     <a href="doc/api/edu/emory/mathcs/backport/java/util/concurrent/RunnableFuture.html">RunnableFuture</a>,
     <a href="doc/api/edu/emory/mathcs/backport/java/util/concurrent/RunnableScheduledFuture.html">RunnableScheduledFuture</a>,
     <a href="doc/api/edu/emory/mathcs/backport/java/util/concurrent/ScheduledThreadPoolExecutor.html#decorateTask(java.lang.Runnable, edu.emory.mathcs.backport.java.util.concurrent.RunnableScheduledFuture)">ScheduledThreadPoolExecutor.decorateTask()</a>,
     better interrupt detection in ThreadPoolExecutor, avoiding garbage retention
     with timeouts in SynchronousQueue, fixed reset in CyclicBarrier,
     remove(x,null) -> false in ConcurrentHashMap, changes in navigable maps,
     addAll fixed in CopyOnWriteArrayList, etc.

 </li>
 <li>New backported classes:
     <a href="doc/api/edu/emory/mathcs/backport/java/util/concurrent/ConcurrentLinkedQueue.html">ConcurrentLinkedQueue</a>,
     ConcurrentSkipList[<a href="doc/api/edu/emory/mathcs/backport/java/util/concurrent/ConcurrentSkipListMap.html">Map</a>,<a href="doc/api/edu/emory/mathcs/backport/java/util/concurrent/ConcurrentSkipListSet.html">Set</a>]</li>
 <li>Optimizations (replacement of ReentrantLock by synchronized) in:
     CyclicBarrier, DelayQueue, Exchanger, ThreadPoolExecutor</li>
 <li>Optimizations of atomic variables (simple reads are now volatile rather than
     synchronized)</li>
 <li>New backported methods in the fair implementation of the ReentrantLock:
     <a href="doc/api/edu/emory/mathcs/backport/java/util/concurrent/locks/ReentrantLock.html#hasWaiters(edu.emory.mathcs.backport.java.util.concurrent.locks.Condition)">hasWaiters(Condition)</a>,
     <a href="doc/api/edu/emory/mathcs/backport/java/util/concurrent/locks/ReentrantLock.html#getWaitQueueLength(edu.emory.mathcs.backport.java.util.concurrent.locks.Condition)">getWaitQueueLength(Condition)</a>,
     <a href="doc/api/edu/emory/mathcs/backport/java/util/concurrent/locks/ReentrantLock.html#getWaitingThreads(edu.emory.mathcs.backport.java.util.concurrent.locks.Condition)">getWaitingThreads(Condition)</a></li>
 <li>Retrofitted collection classes:
     <a href="doc/api/edu/emory/mathcs/backport/java/util/AbstractMap.html">AbstractMap</a>,
     <a href="doc/api/edu/emory/mathcs/backport/java/util/Collections.html">Collections</a>,
     <a href="doc/api/edu/emory/mathcs/backport/java/util/LinkedList.html">LinkedList</a>,
     Tree[<a href="doc/api/edu/emory/mathcs/backport/java/util/TreeMap.html">Map</a>,<a href="doc/api/edu/emory/mathcs/backport/java/util/TreeSet.html">Set</a>]</li>
 <li>Numerous javadoc clarifications and fixes</li>
 </ul>
 <li>Bug fixes</li>
 <ul>
 <li>Upon deserialization, ReentrantLock, ReentrantReadWriteLock, and Semaphore
     were potentially in a locked (or even illegal) state, contrary to the javadoc</li>
 <li>In the fair implementation of ReentrantLock, wait queues of <em>condition variables</em>
     were not actually fair - they are now</li>
 <li>LinkedBlockingQueue had potential deadlocks in remove() and toArray(). It has
     now been replaced by a completely new implementation, based on java.u.c (rather
     than dl.u.c)</li>
 <li>Race condition in Condition.awaitUninterruptibly() could cause signals to be
     missed if they were coinciding with interrupt attempts</li>
 </ul>

 <li>Tests</li>
 <ul>
 <li>Updated unit tests for atomics, AbstractQueuedSynchonizer, ConcurrentHashMap,
     CyclicBarrier, ExecutorCompletionService, LinkedBlockingQueue, ReentrantLock,
     ReentrantReadWriteLock, ScheduledExecutor, ThreadPoolExecutor</li>
 <li>New unit tests for ConcurrentLinkedQueue, ConcurrentSkipList[Map,Set],
     Utils.nanoTime(), LinkedList, Tree[Map,Set]</li>
 <li>Updated numerous stress tests, and new ones added: CachedThreadPoolLoops,
     [Collection,Map]WordLoops, CASLoops, and more</li>
 </ul>
 </ul>
 Version 1.1_01 (Feb 7, 2005) [<a href="http://www.mathcs.emory.edu/dcl/util/backport-util-concurrent/dist/backport-util-concurrent-1.1_01/backport-util-concurrent-1.1_01-changelog.html">CVS log</a>]
 <ul>
 <li>Bugfix: race condition in the fair implementation of ReentrantLock
 caused it to occassionally cause IllegalMonitorState exceptions. Non-fair
 implementation was not affected, however, classes that depend on fair reentrant locks,
 namely: fair ArrayBlockingQueue, fair SynchronousQueue, and PriorityQueue, were
 affected.
 Thanks to Ramesh Nethi for reporting this bug and helping to track it down.</li>

 <li>Testing: backport has been stress-tested using the "loops" tests
 (courtesy of Doug Lea and the JSR 166 Expert Group). The tests
 are included in the development source bundle.</li>
 </ul>

 Version 1.1 (Jan 21, 2005) [<a href="http://www.mathcs.emory.edu/dcl/util/backport-util-concurrent/dist/backport-util-concurrent-1.1/backport-util-concurrent-1.1-changelog.html">CVS log</a>]
 <ul>
 <li>Bugfix: on Windows platforms with Java 1.4.2, the library
 were sometimes behaving as if timeouts were ignored or misinterpreted,
 typically resulting in indefinite waits. This resulted from an internal
 timer overflow that occurred every several hours, and was also manifested
 as a discontinuity in System.nanoTime() values. The problem would happen
 if the overflow occurred during blocked timed wait, if additionally
 a spurious wakeup occurred after the overflow but before timeout
 in the underlying Object.wait().
 This has now been fixed; 1.0_01 users are urged to upgrade to version 1.1.
 Thanks to Ramesh Nethi for reporting this bug and greatly contributing
 to tracking it down.</li>

 <li>Feature: backport has been reconciled with JSR 166 CVS repository
     on Jan 14, 2005. This results in a handful of new things:
     <ul>
     <li>New
         <a href="doc/api/edu/emory/mathcs/backport/java/util/concurrent/TimeUnit.html">time units</a>:
         MINUTES, HOURS, and DAYS.</li>
     <li><a href="doc/api/edu/emory/mathcs/backport/java/util/concurrent/ThreadPoolExecutor.html#allowCoreThreadTimeOut(boolean)">allowCoreThreadTimeOut</a>
         in ThreadPoolExecutor, which enables <em>bounded</em>
         pools that kills threads if they are idle for too long.</li>
     <li>ThreadPoolExecutor now handles excessive interruption requests more gracefully
         (previously, it was reported to be able to crash older JVMs).</li>
     <li><a href="doc/api/edu/emory/mathcs/backport/java/util/Deque.html">Deques</a>.</li>
     <li>Javadoc improvements.</li>
     </ul>
 </li>
 </ul>

 Version 1.0_01 (Dec 28, 2004) [<a href="http://www.mathcs.emory.edu/dcl/util/backport-util-concurrent/dist/backport-util-concurrent-1.0_01/backport-util-concurrent-1.0_01-changelog.html">CVS log</a>]
 <ul>
 <li>Feature: development source bundle with ant scripts allowing to build and
     test the distribution is now available for download.</li>

 <li>Feature: emulation of UncaughtExceptionHandler, in class
     <a href="doc/api/edu/emory/mathcs/backport/java/util/concurrent/helpers/ThreadHelpers.html">ThreadHelpers</a>.</li>

 <li>Documentation: improved, more consistent and accurate javadoc.</li>

 <li>Bugfix: NoClassDefFoundError when using nanoTime() on Java prior to 1.4.2.
     Thanks to Gabriel Wolosin for reporting this bug.</li>

 <li>Bugfix: deadlocks in ConcurrentLinkedQueue when drainTo() or clear() was
     invoked when there was blocked put(). Thanks to Jean Morissette for
     reporting this bug.</li>

 <li>Bugfix: minor glitch in Utils.nanoTime() would cause timer to lose
     accuracy, about 1ns every 11 days, if JVM was running continuously.
     (Note: as it turned out, the fix itself had a bug; see the log for
     version 1.1)</li>
 </ul>

 Version 1.0 (Dec 1, 2004)
 <ul>
 <li>Initial revision</li>
 </ul>

 <h2>Documentation and Support</h2>

 <p>
 For more information:
 <ul>
 <LI><a href="doc/api/">Browse Javadoc</a></LI>
 <LI>Consult the original
     <a href="http://gee.cs.oswego.edu/dl/concurrency-interest/">
     java.util.concurrent</a> documentation and Java 5.0
     <a href="http://java.sun.com/j2se/1.5.0/docs/guide/concurrency/overview.html">Concurrency Utilities Overview</a></LI>

 <li>Check the <a href="http://dcl.mathcs.emory.edu/util/backport-util-concurrent/">project Web page</a> for updates.</li>

 <li>For questions, comments, and discussion, use the
 <a href="http://altair.cs.oswego.edu/mailman/listinfo/concurrency-interest">Concurrency-Interest
 mailing list</a> (courtesy of Doug Lea and the JSR 166 Expert Group). Please clearly indicate
 that your message regards the backport rather than the original JSR 166 API, by prefixing
 the subject line with "backport: " and including appropriate annotation in the message body.
 You may also send
 e-mail directly to <a href="mailto:dawidk@mathcs.emory.edu">Dawid Kurzyniec</a>.</li>
 </ul>
 </p>

 <hr>
 Copyright (C) 2004-2006 <a href="http://dcl.mathcs.emory.edu/">Distributed Computing Laboratory</a>, Emory University<br>

 </body>
 </html>